Change out the developer-centric administration menu with one that meets the needs of your site admins, editors and authors with the Admin Menu Swap module.

The second webinar in the Bringing Drupal AI Into Your DNA series focused on rolling up your sleeves and getting Drupal’s AI module working on real sites. Hosted by Jamie Abrahams, co-founder of FreelyGive, and Marcus Johansson, lead developer on the Drupal AI initiative, this session balanced developer-focused walkthroughs with honest takes on what’s working and what still needs your help.

This session is part of a free, public training series backed by the European Commission in partnership with the Drupal community. The goal: to equip builders, developers, and digital agencies with ethical, open-source tools and knowledge to bring AI into Drupal sites responsibly.

Watch the webinar recording now.

The first session in the series — “Bringing Drupal AI Into Your DNA: Introduction” — featured Jamie alongside Dries Buytaert and introduced the strategic initiative, the long-term vision, and the open contribution model being built around AI in Drupal.

This second session moves from vision to implementation: it’s about getting the AI module installed, configured, and ready to use.

Keep Calm and Install AI

Right from the top, Jamie made the audience feel comfortable diving in, regardless of where they’re starting from:

This one is going to be slightly more developer focused, but you’re going to know how to get started.

Jamie Abrahams

The session covered:

• Installing the AI module via Composer and verifying dependencies.
• Connecting the module to OpenAI or your LLM of choice, with examples for local dev and production.
• Managing permissions and API keys securely.
• Running a basic content generation test to confirm that the setup works end-to-end.

Community Matters

Jamie’s role as co-founder of FreelyGive and maintainer of the AI module keeps him right at the centre of this fast-growing ecosystem. He shared that:

I’ve been focusing on native Drupal CRM and been particularly obsessed with AI recently 

Jamie Abrahams

He made it clear this is all a work in progress. That’s exactly why the Drupal community’s input is so critical. Expect a bit of trial and error, but know that the maintainers genuinely want your feedback.

Marcus’ Hands on Demos

Marcus, who leads parts of the Drupal AI initiative, gave the developer audience concrete examples of how to work with AI functionality, from voice output to local installations. These live demos showed exactly how flexible the AI module can be and where it’s heading next. Expect more video walkthroughs, local dev examples, and client modules that link to other providers.

Join In!

If you’re interested in Drupal + AI, there’s no better time to get involved. Your contributions, whether that’s testing new releases, sharing feedback, or writing documentation, matter.

Next steps:

• Watch the webinar replay if you missed it
• Check out the AI Module on Drupal.org
• Join the #ai channel on Drupal Slack to connect with Jamie, Marcus, and the rest of the community

This webinar is exactly what more open source projects need: practical, transparent, and focused on real implementation, not hype. Install the AI module. Test it. Break it. Improve it. And share what you learn,  because that’s how Drupal AI moves forward.

Upcoming Sessions

In the coming months, the webinar series continues with targeted training sessions that build upon each other. All sessions are free, recorded, and open to the public.

• 2 September: AI Search
• 23 September: AI Agents (No-Code Creation)
• 7 October: Advanced: Build AI Agents with Code
   

Automating all the things - 24 ideas for things you can automate with your website Jul 29, 2025 By John Locke 0

Last December I wrote up 24 specific things we've automated on Drupal, and published them as an Advent Calender -- one little nugget each day.

Sustainable/Open Business Read More

The Drupal Association is excited to announce the winner of 2025 Community At-Large Board Elections and additional board members joining this year.

We extend a sincere thank you to Fei Lauren, Lynne Capozzi and Rosa Ordinana for their service and dedication, not only to Drupal, but to the Drupal community. Your time spent on the board made such a difference to the future of the Drupal project, and we thank you all for participating with grace, thoughtfulness, and insightful contributions.

We would like to congratulate and welcome Dominique De Cooman and Glenn Hilton as our newest board members.

An additional congratulations to Maya Schaeffer for winning the community-elected seat during our 2025 At-Large Board Elections.

We cannot wait to see all the amazing things the new board members will accomplish while on the Drupal Association Board.

Detailed Voting Results

There were 7 candidates in this year’s At-Large board member election. 528 voters cast their ballots out of a pool of 1459 eligible voters.

Under Approval Voting, each voter can give a vote to one or more candidates. 

The final total of votes were as follows:

Candidate

Votes

Alexander Varwijk

114

Carlos Mario Ospina Anzola

140

Matthew Saunders

158

Matt Glaman

264

Maya Schaeffer

269

Vladimir Roudakov

106

Will Huggins

120

On behalf of all the staff and board of the Drupal Association, a heartfelt Drupal Thanks to all of you who stood for the elections this year. It truly is a big commitment to contribution, the Drupal Association, and the community, and we are so grateful for all of your voices. Thank you for your willingness to serve, and we hope you’ll consider participating again in 2026!

Proper Robots.txt file configuration is a cornerstone of SEO optimization for every Drupal website. In Drupal, RobotsTxt can serve as an excellent tool for controlling web crawlers’ access, as it affects how your site appears in search engines. Let's explore how to efficiently create a robots.txt file in Drupal using the RobotsTxt module.

Want to set up a Drupal website quick and easy? Read this article to learn about Lando and how you can leverage it to set up your local Drupal website.

Today we are talking about Working from home, heading back to the office, and the current state of remote work with guest Kaleem Clarkson. We’ll also cover Microsoft 365 Connector as our module of the week.

For show notes visit: https://www.talkingDrupal.com/513

Topics

• Exploring Remote Work with Kaleem Clarkson
• Trust Issues in Management
• Employee Red Flags and Data-Driven Decisions
• Managerial Concerns with Return to Office Policies
• Respectful Implementation of Return to Office
• Challenges of Enforcing Office Mandates
• Benefits of In-Person Work
• Hybrid Work Models and Their Challenges
• Variations in Hybrid Work Policies
• Impact of Seniority on Office Policies
• Cutting DEI Initiatives: Fear and Legal Risks
• Employer Brand and Social Contracts

Resources

Blend Me Inc

Guests

Kaleem Clarkson - kclarkson

Hosts

Nic Laflin - nLighteneddevelopment.com nicxvan John Picozzi - epam.com johnpicozzi Rich Lawson - richlawson.co rklawson

MOTW Correspondent

Martin Anderson-Clutz - mandclu.com mandclu

• Brief description:
  • Have you ever wanted your Drupal site to integrate with Microsoft 365, so users can log in with their Azure AD credentials, and then have direct access to shared files, see recent emails, and more? There’s a module for that.
• Module name/project name:
  • Microsoft 365 Connector
• Brief history
  • How old: created in July 2019 by immoreel, though the most recent release is by Boris Doesborg (batigolix), both of Finalist, a Dutch Drupal shop
  • Versions available: 5.0.22 and 5.1.0-beta1, the latter of which supports Drupal 9.4, 10, and 11
• Maintainership
  • Actively maintained
  • Security coverage
  • Test coverage
  • Two documentation guide available
  • Number of open issues: 18 open issues, 1 of which is a bug, though it is postponed waiting for more info
• Usage stats:
  • 365 sites
• Module features and usage
  • This module integrates your Drupal site with the Microsoft Graph API, a unified API that provides a single endpoint for accessing data and intelligence from Microsoft 365 services, including Exchange, SharePoint, OneDrive, and more
  • Microsoft 365 Connector includes more than a dozen submodules, each of which provide specific capabilities like Single Sign-On, syncing data to Drupal user accounts, sending Teams messages from within Drupal, and more
  • You can also use this module to do things like automatically add an event node to your Outlook calendar, and invite other people at the same time
  • It’s worth noting that in the documentation guide the submodules are named “Office 365”, which is probably what the module was named until around 5 years ago when Microsoft retired the Office 365 name
  • Finally, setting up this module requires registering an app in Azure AD, so it’s not for the casual user. But if you're working on an intranet or similar collaboration platform for an organization that is heavily invested in the Microsoft 365 suite, this could make for a compelling integration

Drupal has evolved significantly from it's initial developer-focused content management system. In 2025, it's entering one of its most notable moments yet, characterized by low-code content creation, AI-driven personalization, out-of-the-box website templates, and enterprise scalability. This shift reflects Drupal's commitment to accessibility, innovation, and performance as it moves away from legacy systems like Drupal 7 and marches toward the future. Here are 10 key things to watch for in 2025: 1. Drupal CMS 2.0 Drupal CMS 2.0 is planned for Q3 2025, October, as a major release. It succeeds

In the coming weeks, we are going to see some big changes in how (and where) credits are given when working on issues. The new system can do everything that the current system can, but it will enable us to progress further in the GitLab issue migration.

User and organisation profiles won’t change, and they will still have the same credit listings, only reading them from the new system. Likewise, marketplace ranking won’t change either.

Opening the door for GitLab issues

The next significant milestone for us (Drupal Association and the Drupal community) is to migrate Drupal.org issues to GitLab issues, as this will provide everyone with the advantages of using a proven and dedicated system for managing project issues. This is part of the GitLab acceleration initiative that we started years ago.

However, an important part of working with Drupal is the credit recognition, for both individuals and organisations. Note that GitLab issues, while having numerous advantages when dealing with issues, queues, labels… lack the ability to attribute credit to individuals and organisations out of the box. You could attribute commits, but remember that many contributions do not require any code (and therefore commits), so we cannot rely on that.

Due to this, we will “decouple” the Contribution Records system from the issue management. We will connect them, but they will be two separate systems. This also opens the door to different ways of recording contributions for the future, such as providing credit for translations for Drupal from localize.drupal.org, for example.
 

The new Contribution Records system

The new system is built on the modern new.drupal.org site, and this will be the first part of the new site that the community will interact with. 

It will look like the screenshot below, with a full page dedicated to the Contribution Record. We can see how the source issue is linked to it. All the fields will be pre-populated, including the contributors and their participation in the issue, so you won’t need to re-enter them again. There will be a link on every drupal.org issue that will take you directly to the Contribution Record page.

Maintainers will then, with the information given, choose whether or not to grant credit to each contributor. This is the case already with the current system.

As maintainer, you wiill still have a section to copy/paste the commit message, using the new commit message format. 

And as a contributor, you will still be able to attribute your work as you do now.

Endpoints to query this information

Our initial integration makes the www.drupal.org read the credit information from new.drupal.org via API calls. These are:

• https://new.drupal.org/contribution-records-by-user
• https://new.drupal.org/contribution-records-by-organization
• https://new.drupal.org/contribution-records-by-organization-by-user

You should add filters to those endpoints, to limit the data to be retrieved. The options are documented in the README file of the underlying module.

What’s next?

The first step is to adopt the usage of the new system, to check that all the data and marketplace calculations are unaffected, and to fix any possible issues that the new system could have.

We will have Drupal.org issues connected to the new Contribution Records, so we will all need to get used to going “to the new place” to record credit information. Remember that when we use GitLab issues, the workflow will be exactly the same.

This leads to what many of us have been waiting for a long time: the move to GitLab issues! We will open an opt-in period for projects to move their issues to GitLab and help us iron out the integration. After that opt-in period, we will progressively move issues for the rest of the projects. There will be a time when some projects will manage their issues via Drupal.org and some others via GitLab, but one thing will remain the same: the new Contribution Records system.

If you want to report bugs or give feedback you can do it via the issue queue for the Contribution Records module.

The uncomfortable truth is finally being spoken aloud in Drupal circles: the boutique agency model might be dying. John Faber, Managing Partner at Chapter Three, deserves credit for finally asking the question that many have been avoiding: can pure-play Drupal agencies survive the current market? His recent LinkedIn post wasn't just provocative, it was overdue. For too long, we've watched agencies struggle with razor-thin margins, scramble for projects, and pretend that "craft over scale" is a sustainable business strategy when AI tools can now deliver functional websites faster than a junior developer can set up their local environment.

What makes this moment particularly telling isn't the economic pressure alone that's been building for years. It's the structural shift underneath. The value of a "developer hour" has fundamentally changed when clients can get 80% of what they need from AI-powered platforms for a fraction of the cost. The agencies caught in the middle (too small to pivot to enterprise consulting, too expensive to compete with automated solutions) are finding themselves in an increasingly uncomfortable squeeze. The responses to Faber's post revealed a community grappling with whether to double down on specialization or admit that the market has moved beyond their traditional offerings.

The irony is rich: Drupal, built on the principle of democratizing web publishing, may now be too complex for the very market segment that sustained its agency ecosystem. While enterprise clients still need sophisticated digital architecture, the small-to-medium projects that kept boutique shops profitable are evaporating. The question isn't whether agencies need to evolve, it's whether they have the capital and courage to transform before the window closes entirely.

INTERVIEW

• The Drupaler-Therapist Who Diagnoses More Than Code

DISCOVER DRUPAL

• Drupal Leaflet Choropleth Module: Create Interactive Geographic Data Maps Without Custom Code
• Can a Pure-Play Drupal Agency Survive? The Case for Reinvention
• Drupal Marketplace Pilot Will Launch Without Listing Fees, Focuses on Accessibility and On-Site Sales Options
• Proposal Seeks to Simplify Drupal Theme and Render Systems for Design System Support

EVENT

• DrupalCamp Colorado 2025 Returns with AI Topics, Accessibility Commitments, and Free Events
• Drupal and PHP Communities Partnering for Longhorn PHP Conference 2025
• Splash Awards Germany & Austria 2025: Submissions Now Open
• Drupal Uruguay Meetup 2025: Connect, Learn, and Collaborate
• DrupalCon Chicago 2026 Call for Speakers Now Open

ORGANIZATION NEWS

• Morpht Delivers NOCS Website Redesign in Three Weeks for National Child Safety Campaign
• Vardot Offers Free Enterprise Drupal Website Audit
• Drupal Career Online Next Semester Begins September 1
   

We acknowledge that there are more stories to share. However, due to selection constraints, we must pause further exploration for now.

To get timely updates, follow us on LinkedIn, Twitter, Bluesky, and Facebook. You can also join us on Drupal Slack at #thedroptimes.

Thank you, 
Sincerely 
Alka Elizabeth 
Sub-editor, The DropTimes.

Can AI actually help with real Drupal development? I wanted to find out.

This morning, I fired up Claude Code and pointed it at my personal Drupal site. In a 30-minute session, I asked it to help me build new features and modernize some of my code. I expected mixed results but was surprised by how useful it proved to be.

I didn't touch my IDE once or write a single line of code myself. Claude handled everything from creating a custom Drush command to refactoring constructors and converting Drupal annotations to PHP attributes.

If you're curious what AI-assisted Drupal development actually feels like, this video captures the experience.

Since the deprecation of Universal Analytics in 2023, Google Analytics 4 (GA4) has been the standard for web analysis. Becoming proficient in event-based tracking is the secret to achieving the maximum potential of GA4 on conversion and user insight. To be able to understand how your users interact with your site is to know how to optimize performance and convert. GA4 introduces a powerful event-based tracking model that is more developed than traditional URL path or file type filtering and offers a more precise, adaptive, and actionable way to measure user behavior. This guide discusses why

Monitoring tools for Kubernetes: Prometheus vs Grafana vs Datadog Editor Fri, 07/25/2025 - 13:01

Kubernetes operates without pause; that's why it needs ongoing monitoring. Otherwise, small issues will turn into big losses: from a decrease in system performance to the loss of users. This is particularly true for Drupal projects, where a mere lag of a few seconds might cause users to leave. Honestly, for us, monitoring isn't just "nice to have"—it's the bedrock of staying competitive online. It's how we catch problems before they blow up, figure out where our resources are really going, and constantly boost efficiency. The main contenders here are Datadog, Grafana, and Prometheus, and they each tackle metric collection and processing in their own unique way. So, how do these tools actually compare? Let's get into their unique features.

Dive into the Drupal AI Initiative's progress in June 2025. Discover how 1xINTERNET is helping to build a foundation for AI-powered content management through key contributions in workflow, collaboration, and outreach.

From translating Drupal into Ukrainian nearly solo to architecting one of the platform’s largest distributions, Andrii Podanenko’s journey is anything but ordinary. In this conversation with The DropTimes, he and Zhanna Khoma of ITCare talk about building AI-integrated Drupal projects, mentoring without ego, and why empathy, not just code, drives lasting impact.

Events are more than just dates on a calendar. They’re opportunities to build trust and visibility for your brand, educate and inspire your audience, and spark meaningful actions like registrations, donations, or follow-ups.

> "To misuse a woodworking metaphor, I think we're experiencing a shift from hand tools to power tools. You still need someone who understands the basics to get good results from the tools, but they're not chiseling fine furniture by hand anymore. They're throwing heaps of wood through the tablesaw instead. More productive, but more likely to lose a finger if you're not careful." > – mrmincent, [Hacker News comment on Claude Code](https://news.ycombinator.com/item?id=44431095), via [Simon Willison](https://simonwillison.net/2025/Jul/1/mrmincent/) If, like me, you work in web development, design, or digital strategy, this quote might hit close to home. But it may not go far enough. We are not just moving from chisels to table saws. We are about to hand out warehouse-sized CNC machines and robotic arms. This is not just an upgrade in tools. The Industrial Revolution didn't just replace handcraft with machines. It upended entire industries. History does not repeat itself, but it often rhymes. For over two centuries, new tools have changed not just how work gets done, but what we can accomplish. AI is changing how websites are built and how people find information online. Individual developers are already using AI tools today, but broader organizational adoption will unfold over the years ahead. It's clear that AI will have a deep impact on the web industry. Over time, this shift will affect those of us who have built our careers in web development, marketing, design, and digital strategy. I am one of them. Most of [my career](https://dri.es/about) has been rooted in [Drupal](https://www.drupal.org), which makes this both personal and difficult to write. But this shift is bigger than any one person or platform. There are more than 175,000 digital agencies worldwide, employing millions of people. Behind those numbers are teams, individuals, and livelihoods built over decades. Our foundation is shifting. It touches all of us, and we all need to adapt. If you are feeling uncertain about where this is heading, you are not alone. ### Why I am writing this I am _not_ writing this to be an alarmist. I actually feel a mix of emotions. I am excited about the possibilities AI offers, but also concerned about the risks and uneasy about the speed and scale of change. As the project lead of [Drupal](https://www.drupal.org/), I ask myself: "How can I best guide our community of contributors, agencies, and end users through these changes?". Like many of you, I am trying to understand what the rise of AI means for our users, teams, partners, contributors, products, and values. I want to help however I can. I don't claim to have all the answers, but I hope this post sparks discussion, encourages deeper thinking, and helps us move forward together. This is not a roadmap, just a reflection of where my thinking is today. I do feel confident that we need to keep moving forward, stay open-minded, and engage with the changes AI brings head-on. Even with all that uncertainty, I feel energized. Some of the hardest challenges the Drupal community has faced, such as improving usability or maintaining documentation, may finally become more manageable. I see ways AI can support Drupal's mission, lower barriers to online publishing, make Drupal more accessible, and help build a stronger, more inclusive Open Web. The future is both exciting and uncertain. But this post isn't just for the Drupal community. It's for anyone working in or around a digital agency who is asking: "What does AI mean for my team, my clients, and my future?". I will focus more directly on Drupal in my next blog post, so feel free to [subscribe](https://dri.es/subscribe). If you are thinking about how AI is affecting your work, whether in the Drupal ecosystem or elsewhere, I would love to hear from you. The more we share ideas, concerns, and experiments, the better prepared we will all be to shape this next chapter together. ### The current downturn is real, but will pass Before diving into AI, I'd be remiss not to acknowledge the current economic situation. Agencies across all platforms, not just those working with Drupal, are experiencing challenging market conditions, especially in the US and parts of Europe. While much of the industry is focused on AI, the immediate pain many agencies are feeling is not caused by it. High interest rates, inflation, and global instability have made client organizations more cautious with spending. Budgets are tighter, sales cycles are longer, competition is fiercer, and more work is staying in-house. As difficult as this is, it is not new. Economic cycles and political uncertainty have always come and gone. What makes this moment different is not the current downturn, but what comes next. ### AI will transform the industry at an accelerating pace AI has not yet reshaped agency work in a meaningful way, but that change is knocking at the door. At the current pace of progress, web development and digital agency work are on the verge of the most significant disruption since the rise of the internet. One of the most visible areas of change has been content creation. AI generates drafts blog posts, landing pages, social media posts, email campaigns, and more. This speeds up production but also changes the workflow. Human input shifts toward editing, strategy, and brand alignment rather than starting from a blank page. Code generation tools are also handling more implementation tasks. Senior developers can move faster, while junior developers are taking on responsibilities that once required more experience. As a result, developers are spending more time reviewing and refining AI-generated code than writing everything from scratch. Traditional user interfaces are becoming less important as AI shifts user interactions toward natural language, voice, and more predictive or adaptive experiences. These still require thoughtful design, but the nature of UI work is changing. AI can now turn visual mockups into functional components and, in some cases, generate complete interfaces with minimal or no human input. These shifts also challenge the way agencies bill for their work. When AI can do in minutes what once took hours or days, hourly billing becomes harder to justify. If an agency charges $150 an hour for something clients know AI can do faster, those clients will look elsewhere. To stay competitive, agencies will need to focus less on time spent and more on outcomes, expertise, and impact. AI is also changing how people find and interact with information online. As users turn to AI assistants for answers, the role of the website as a central destination is being disrupted. This shift changes how clients think about content, traffic, and performance, which are core areas of agency work. Traditional strategies like SEO become less effective when users get what they need without ever visiting a site. Through all of this, human expertise will remain essential. People are needed to set direction, guide priorities, review AI output, and take responsibility for quality and business outcomes. We still rely on individuals who know what to build, why it matters, and how to ensure the results are accurate, reliable, and aligned with real-world needs. When AI gets it wrong, it is still people who are accountable. Someone must own the decisions and stand behind the results. But taken together, these changes will reshape how agencies operate and compete. To stay viable, agencies need to evolve their service offerings and rethink how they create and deliver value. That shift will also require changes to team structures, pricing models, and delivery methods. This is not just about adopting new tools. It is about reimagining what an agency does and how it works. The hardest part may not be the technology. It is the human cost. Some people will see their roles change faster than they can adapt. Others may lose their jobs or face pressure to use tools that conflict with their values or standards. Adding to the challenge, adopting AI requires investment at a moment when many agencies around the world are focused on survival. For teams already stretched thin, transformation may feel out of reach. The good news is that AI's full impact will take years to unfold, giving agencies time to adapt. Still, moments like this can create major opportunities. In past downturns, technology shifts made room for new players and helped established firms reinvent themselves. The key is recognizing that this is not just about learning new tools. It is about positioning yourself where human judgment, relationships, and accountability for outcomes remain essential, even as AI takes on more of the execution. ### The diminishing value of platform expertise alone For years, CMS-focused agencies have built their businesses on deep platform expertise. Clients relied on them for custom development, performance tuning, security, and infrastructure. This specialized knowledge commanded a premium. In effect, AI increases the supply of skilled work without a matching rise in demand. By automating tasks that once required significant expertise, it makes technical expertise abundant and much cheaper to produce. And according to the principles of supply and demand, when supply rises and demand stays the same, prices fall. This is not a new pattern. SaaS website builders already commoditized basic site building, reducing the perceived value of simple implementations and pushing agencies toward more complex, higher-value projects. Now, AI is _accelerating_ that shift. It is extending the same kind of disruption into complex and enterprise-level work, bringing speed and automation to tasks that once required expensive and experienced teams. In other words, AI erodes the commercial value of platform expertise by making many technical tasks less scarce. Agencies responded to earlier waves of commoditization by moving up the stack, toward work that was more strategic, more customized, and harder to automate. AI is raising the bar again. Once more, agencies need to move further up the stack. And they need to do it faster than before. ### The pattern of professional survival This is not the first time professionals have faced a major shift. Throughout history, every significant technological change has required people to adapt. Today, skilled radiologists interpret complex scans with help from AI systems. Financial analysts use algorithmic tools to process data while focusing on high-level strategy. The professionals who understand their domain deeply find ways to work with new technology instead of competing against it. Still, not every role survives. Elevator operators disappeared when elevators became automatic. Switchboard operators faded as direct dialing became standard. At the same time, these shifts unlocked growth. The number of elevators increased, making tall buildings more practical. The telephone became a household staple. As routine work was automated away, new industries and careers emerged. The same will happen with AI. Some roles will go away. Others will change. Entirely new opportunities will emerge, many in areas we have not yet imagined. I have lived through multiple waves of technological change. I witnessed the rise of the web, which created entirely new industries and upended existing ones. I experienced the shift from hand-coding to content management systems, which helped build today's thriving agency ecosystem. I saw mobile reshape how people access information, opening up new business models. Each transition brought real uncertainty. In the moment, the risks felt immediate and the disruption felt personal. But over time, these shifts consistently led to new forms of prosperity, new kinds of work, and new ways to create value. ### The great agency unbundling AI can help agencies streamline how they work today, but when major technology shifts happen, success rarely comes from becoming more efficient at yesterday's model. The bigger opportunity lies in recognizing when the entire system is being restructured. The real question is not just "How do we use AI to become a more efficient agency?" but "How will the concept of an agency be redefined?". Most agencies today _bundle_ together strategy, design, development, project management, and ongoing maintenance. This bundle made economic sense when coordination was costly and technical skills were scarce enough to command premium rates. AI is now _unbundling_ that model. It separates work based on what can be automated, what clients can bring in-house, and what still requires deep expertise. At the same time, it is _rebundling_ services around different principles, such as speed, specialization, measurable outcomes, accountability, and the value of human judgment. ### The accountability gap As AI automates routine tasks, execution becomes commoditized. But human expertise takes on new dimensions. Strategic vision, domain expertise, and cross-industry insights remain difficult to automate. More critically, trust and accountability stay fundamentally human. When AI hallucinates or produces unexpected results, organizations need people who can take responsibility and navigate the consequences. We see this pattern everywhere: airline pilots remain responsible for their passengers despite autopilot handling most of the journey, insurance companies use advanced software to generate quotes but remain liable for the policies they issue, and drivers are accountable for accidents even when following GPS directions. The tools may be automated, but responsibility for mistakes and results remains human. For agencies, this means that while AI can generate campaigns, write code, and design interfaces, clients still need someone accountable for strategy, quality, and outcomes. This accountability gap between what AI can produce and what organizations will accept liability for creates lasting space for human expertise. ### The rise of orchestration platforms Beyond human judgment, a new architectural pattern is emerging. Traditional Digital Experience Platforms (DXPs) excel at managing complex content, workflows, and integrations within a unified system. But achieving sophisticated automation often requires significant custom development, long implementation cycles, and deep platform expertise. Now, visual workflow builders, API orchestration platforms, and the Model Context Protocol are enabling a different approach. Instead of building custom integrations or waiting for platform vendors to add features, teams can wire together AI models, automation tools, CRMs, content systems, and analytics platforms through drag-and-drop interfaces. What once required months of development can often be prototyped in days. But moving from prototype to production requires deep expertise. It involves architecting event-driven systems, managing state across distributed workflows, implementing proper error handling for AI failures, and ensuring compliance across automated decisions. The tools may be visual, but making them work reliably at scale, maintaining security, ensuring governance, and building systems that can evolve with changing business needs demands sophisticated technical knowledge. This orchestration capability represents a new technical high ground. Agencies that master this expanded stack can deliver solutions faster while maintaining the reliability and scalability that enterprises require. ### Six strategies for how agencies could evolve Agencies need two types of strategies: ways to compete better in today's model and ways to position for the restructured system that's emerging. The strategies that follow are not mutually exclusive. Many agencies will combine elements from several based on their strengths, clients, and markets. #### Competing in today's market **1. Become AI-augmented, not AI-resistant.** To stay competitive, agencies should explore how AI can improve efficiency across their entire operation. Developers should experiment with code assistants, project managers should use AI to draft updates and reports, and sales teams should apply it to lead qualification or proposal writing. The goal is not to replace people, but to become more effective at handling fast-paced, low-cost work while creating more space for strategic, value-added thinking. **2. Focus on outcomes, not effort.** As AI reduces delivery time, billing for hours makes less sense. Agencies can shift toward pricing based on value created rather than time spent. Instead of selling a redesign, offer to improve conversion rates. This approach aligns better with client goals and helps justify pricing even as technical work becomes faster. **3. Sell through consultation, not execution.** As technology changes faster than most clients can keep up with, agencies have a chance to step into a more consultative role. Instead of just delivering projects, they can help clients understand their problems and shape the right solutions. Agencies that combine technical know-how with business insight can become trusted partners, especially as clients look for clarity and results. #### Positioning for what comes next **4. Become the layer between AI and clients.** Don't just use AI tools to build websites faster. Position yourself as the essential layer that connects AI capabilities with real client needs. This means building quality control systems that review AI-generated code before deployment and becoming the trusted partner that translates AI possibilities into measurable results. Train your team to become "AI translators" who can explain technical capabilities in business terms and help clients understand what's worth automating versus what requires human judgment. **5. Package repeatable solutions.** When custom work becomes commoditized, agencies need ways to stand out. Turn internal knowledge into named, repeatable offerings. This might look like a "membership toolkit for nonprofits" or a "lead gen system for B2B SaaS". These templated solutions are easier to explain, sell, and scale. AI lowers the cost of building and maintaining them, making this model more realistic than it was in the past. This gives agencies a way to differentiate based on expertise and value, not just technical execution. **6. Build systems that manage complex digital workflows.** Stop thinking in terms of one-off websites. Start building systems that manage complex, ongoing digital workflows. Agencies should focus on orchestrating tools, data, and AI agents in real time to solve business problems and drive automation. For example, a website might automatically generate social media posts from new blog content, update landing pages based on campaign performance, or adjust calls to action during a product launch. All of this can happen with minimal human involvement, but these systems are still non-trivial to build and require oversight and accountability. This opportunity feels significant. As marketing stacks grow more complex and AI capabilities expand, someone needs to coordinate how these systems work together in a structured and intelligent way. This is not just about connecting APIs. It is about designing responsive, event-driven systems using low-code orchestration tools, automation platforms, and AI agents. ### Open Source needs agencies, proprietary platforms don't Every AI feature a technology platform adds potentially takes work off the agency's plate. Whether the platform is open source or proprietary, each new capability reduces the need for custom development. But open source and proprietary platforms are driven by very different incentives. Proprietary platforms sell directly to end clients. For them, replacing agency services is a growth strategy. The more they automate, the more revenue they keep. This is already happening. Squarespace builds entire websites from prompts. Shopify Magic writes product descriptions and designs storefronts. Open source platforms are adding AI features as well, but operate under different incentives. Drupal doesn't monetize end users. Drupal's success depends on a healthy ecosystem where agencies contribute improvements that keep the platform competitive. Replacing agencies doesn't help Drupal; it weakens the very ecosystem that sustains it. As the Project Lead of Drupal, I think constantly about how Drupal the product and its ecosystem of digital agencies can evolve _together_. They need to move in step to navigate change and help shape what comes next. This creates a fundamental difference in how platforms may evolve. Proprietary platforms are incentivized to automate and sell directly. Open source platforms thrive by leaving meaningful work for agencies, who in turn strengthen the platform through contributions and market presence. For digital agencies, one key question stands out: do you want to work with platforms that grow by replacing you, or with platforms that grow by supporting you? ### Looking ahead Digital agencies face a challenging but exciting transition. While some platform expertise is becoming commoditized, entirely new categories of value are emerging. The long-term opportunity isn't just about getting better at being an agency using AI tools. It's about positioning yourself to capture value as digital experiences evolve around intelligent systems. Agencies that wait for perfect tools, continue billing by the hour for custom development, try to serve all industries, or rely on platform knowledge will be fighting yesterday's battles. They're likely to struggle. But agencies that move early, experiment with purpose, and position themselves as the essential layer between AI capabilities and real client needs are building tomorrow's competitive advantages. Success comes from recognizing that this transition creates the biggest opportunity for differentiation that agencies have seen in years. For those working with Drupal, the open source foundation creates a fundamental advantage. Unlike agencies dependent on proprietary platforms that might eventually compete with them, Drupal agencies can help shape the platform's AI evolution to support their success rather than replace them. We are shifting from hand tools to power tools. The craft remains, but both how we work and what we work on are changing. We are not just upgrading our tools; we are entering a world of CNC machines and robotic arms that automate tasks once done by hand. Those who learn to use these new capabilities, combining the efficiency of automation with human judgment, will create things that were not possible before. In the next post, I'll share why I believe Drupal is especially well positioned to lead in this new era of AI-powered digital experience. *I've rewritten this blog post at least three times. Throughout the process, I received valuable feedback from several Drupal agency leaders and contributors, whose insights helped shape the final version. In alphabetical order by last name: [Jamie Abrahams](https://www.drupal.org/u/yautja_cetanu), [Christoph Breidert](https://www.drupal.org/u/breidert), [Dominique De Cooman](https://www.drupal.org/u/domidc), [George DeMet](https://www.drupal.org/u/gdemet), [Alex Dergachev](https://www.drupal.org/u/dergachev), [Justin Emond](https://www.drupal.org/u/jemond), [John Faber](https://www.drupal.org/u/flavor), [Seth Brown](https://www.drupal.org/u/sethlbrown), [Seth Gregory](https://www.drupal.org/u/sgregory), and [Michael Meyers](https://www.drupal.org/u/michaelemeyers).*

This is an unusual post for my blog, but I'm sharing it to start a broader conversation about an idea we're exploring: a marketplace for Drupal Site Templates. Both the Drupal CMS Leadership Team and the Drupal Association have discussed this concept, but no decision has been made. I'm posting to share our current thinking and invite feedback as we shape this together.

This post will also be cross-posted to Drupal.org, where comments are open. You're also welcome to join the conversation in the #drupal-cms-marketplace channel on Drupal Slack.

In my DrupalCon Atlanta keynote, I introduced the concept of Site Templates for Drupal. If you haven't seen my keynote yet, I recommend watching it first. It provides helpful context for the rest of this post.

Site Templates provide pre-configured website starting points that combine Drupal recipes, themes, and default content. While Site Templates will help users launch websites faster, I also posed a bigger question: should we create a marketplace where users can discover and download or install these templates? And if so, should that marketplace offer only open source Site Templates, or should we also allow commercial templates?

What are Site Templates?

Site Templates combine Drupal recipes, a theme, design elements, and default content to give users a fully functional website right from the start. They solve one of Drupal's biggest challenges: the time it takes to build a site from scratch.

Unlike a bare Drupal installation, a Site Template provides all the components needed for a specific use case. A restaurant template might include menu management, reservation systems, and food photography. A nonprofit template could feature donation processing, event management, and impact reporting tools.

Why consider a marketplace?

A Drupal marketplace for Site Templates would:

1. Help new users launch a professional-looking site instantly
2. Showcase Drupal's full potential through high-quality starting points
3. Generate new revenue opportunities for Drupal agencies and developers
4. Support Drupal's long-term sustainability through a revenue-sharing model with the Drupal Association

Should we support both open source and commercial Site Templates?

Fully open source Site Templates align naturally with Drupal's values. They could function much like community-contributed modules and themes, and we hope that many contributors will take this approach.

A marketplace requires ongoing investment. The Drupal Association would need to maintain the platform, review submissions, provide support, and ensure templates meet high standards. Without dedicated resources, quality and sustainability would suffer.

This is why supporting both open source and commercial templates makes sense. Paid templates can create a sustainable revenue stream to fund infrastructure, quality control, and support.

Commercial incentives also give creators a reason to invest in polished, well-documented, and well-supported templates.

How can a template be commercial while respecting Drupal's open source values?

First, rest assured: Drupal modules will always be open source.

Drupal is licensed under the GNU General Public License, or GPL. We've always taken a conservative approach to interpreting the GPL. In practice, this means we treat any code that builds on or interacts closely with Drupal as subject to the GPL. This includes PHP, Twig templates, etc. If it relies on Drupal's APIs or is executed by Drupal, it must be GPL-licensed.

Some parts of a site template fall into a gray area. JavaScript is an example. If JavaScript code is integrated with Drupal, we treat it as GPL-covered. If JavaScript code is standalone, such as a self-contained React component, it may not be considered a derivative work. The same may apply to CSS or configuration files not tightly coupled with Drupal APIs. These cases aren't always clear, but our stance has been to treat all code that ships with and interacts with Drupal as GPL. This keeps things simple.

Other parts of a Site Template are likely not subject to the GPL. Assets like images, fonts and icons are not code and are not derived from Drupal. The same applies to demo content, such as placeholder text or sample articles. These elements are not integrated with Drupal in a technical sense and can use other licenses, including commercial ones.

So when we talk about a commercial Site Template, we mean one that combines open source code with separately licensed assets or is sold alongside value-added services like documentation, support, or updates.

What would people actually be paying for in a commercial template?

While the legal distinction clarifies which parts of a Site Template can be licensed commercially, it's only part of the picture. The real question is the value proposition: what are users actually paying for when they choose a commercial template?

When purchasing a commercial template, users wouldn't just be paying for code. They're potentially paying for:

• Professional design assets and media
• Time saved in configuration and setup
• Documentation and support
• Ongoing updates and maintenance

This approach aligns with the Free Software Foundation's stance (the organization that created the GPL), which has always supported commercial distribution of free software. Creating a commercial template means balancing open source code with separately licensed assets. However, the real commercial value often extends beyond just the files you can license differently.

A sustainable commercial strategy combines proper licensing with controlled distribution channels and value-added services, like support. This approach ensures the value of a site template isn't limited to easily copied assets, but includes expertise that can't be simply downloaded. This is how a template can be commercial while staying true to Drupal's open source values.

How would we maintain quality in the marketplace?

A marketplace filled with low-quality or abandoned templates would damage Drupal's reputation. To ensure quality we probably need:

• Technical reviews of templates for security and performance
• Standards for documentation and support
• Processes to handle outdated or abandoned templates
• Community ratings and reviews
• Processes for resolving disputes

These quality assurance measures require ongoing time, effort, and funding. This is why including a commercial component makes sense. A revenue-sharing model with template creators could help fund platform maintenance, reviews, support, and other efforts needed to keep the marketplace high quality and trustworthy.

What pricing models might be available?

We don't know yet, but we've heard many good suggestions from the community.

Ideas include one-time purchases for unlimited use, annual subscriptions with ongoing updates, and a marketplace membership model for template creators.

The marketplace could support multiple approaches, giving creators flexibility to choose what works best for their offerings.

Is it fair for template creators to profit while module contributors aren't paid?

When a site template is sold commercially, it raises an important question. What about the maintainers of the modules included in the template? The template builder receives payment. The Drupal Association may collect a revenue share. But the individual contributors who created the modules or core functionality do not receive direct compensation, even though their work is essential to the Site Template.

This may feel frustrating or unfair. Contributors often donate their time to improve Drupal for everyone. Seeing others earn money by building on that work without recognition can be disheartening, and could even discourage future contributions. It's an important concern, and one we plan to take seriously as we evaluate the marketplace model.

At the same time, this dynamic is not new. Agencies and developers already build paid Drupal sites using contributed modules without directly compensating the people who made the underlying code possible. This is both legal, expected, and common in open source.

A marketplace would not create this reality, but it would make it more visible. That visibility gives us a chance to confront a long-standing tension in open source: the gap between those who contribute foundational work and those who profit from it. As I wrote in Makers and Takers, sustaining open source requires a better balance between contribution and benefit. A marketplace could give us a way to explore new approaches to recognize, support, and sustain the people who make Drupal possible. Transparency alone won't solve the issue, but it opens the door to progress and experimentation.

When commercial activity happens off Drupal.org, there is no way to recognize the contributors who made it possible. When it happens on Drupal.org, we have an opportunity to do better. We can explore models for financial support, community recognition, and long-term sustainability.

Others could build marketplaces for Drupal templates, but these would likely focus on profit rather than community support. An official Drupal Association marketplace allows us to reinvest in the project and the people behind it. It keeps value within our ecosystem, and gives us a platform to explore more equitable ways to sustain open source contribution.

Would this hurt digital agencies?

Many organizations pay thousands of dollars to digital agencies as part of a custom Drupal build. If Site Templates are available at a much lower cost, will that undercut agencies?

I don't believe it will.

Organizations investing in a Drupal website are not paying for a theme alone. Agencies provide strategy, consulting, design, customization, user testing, performance optimization, and long-term support. A template offers a starting point, but doesn't replace tailored work or a trusted partnership.

Could templates help agencies grow?

A template marketplace could expand the Drupal ecosystem by lowering the barrier to entry, making Drupal accessible to smaller organizations. Some of these will grow and require custom solutions, creating more opportunities for agencies in the long run.

Templates can also serve as powerful demonstrations of an agency's capabilities, attracting clients who want similar but customized solutions. For agencies, templates become both a product and a marketing tool.

What revenue opportunities would digital agencies have?

A template marketplace offers two revenue streams for Drupal agencies and freelancers.

First, agencies would earn direct income from template sales through revenue-sharing with the Drupal Association. While this income might be modest, it could provide recurring revenue as the marketplace grows.

Second, templates could serve as a foundation for more comprehensive service packages, including hosting, maintenance, and customization services.

How would templates connect agencies with new clients?

A marketplace could connect end users directly with service providers. With proper consent, contact details from template purchases could be shared with creators, opening the door to professional service opportunities. Template listings could also include a built-in contact form, making it easy for users to request customization or support.

This lead generation benefits both sides. Users access trusted professionals who understand their implementation, while agencies connect with qualified prospects who have already invested in their work. A marketplace becomes a matchmaking platform connecting those who need Drupal expertise with those who provide it.

Why is now the right time for this initiative?

With Drupal CMS, we're focused on growth. To succeed, we need to address two long-standing challenges: the lack of ready-made themes and a steep learning curve. Some of our new tools (Recipes, Experience Builder, and Site Templates) allow us to address these longstanding issues.

The result? We can take Drupal's flexibility and make it more available across different markets and use cases.

What was the initial reaction at DrupalCon?

The day after my keynote, we organized a Birds of a Feather (BoF) session to discuss the marketplace concept. Approximately 75 people attended, representing a cross-section of our community.

The discussion was lively and constructive. Participants raised thoughtful concerns about quality control, licensing, and impact on module contributors. They also offered suggestions for implementation, pricing, and sustainability models.

At the session's conclusion, we informally polled the audience. We asked people to raise their hand showing 1 finger if they thought a marketplace was a terrible idea, and 5 if they considered it a very impactful idea. Most responses were 4, with some 5s. Very few people indicated less than 3.

This initial reaction is encouraging, though we recognize that much work remains to address the questions raised during the session.

We also opened the #drupal-cms-marketplace channel in Drupal Slack to continue the conversation with the wider community.

What are the next steps?

The Drupal CMS Leadership Team and the Drupal Association Innovation Working Group have been exploring this idea the past month.

We believe it could be one of our strongest opportunities to grow Drupal adoption, support our Maker community, and strengthen the Drupal Association. (As a disclaimer: I serve on both the Drupal CMS Leadership Team and the Drupal Association Board of Directors.)

To be clear, no decision has been made. We recognize this initiative would have a substantial impact on our community and ecosystem. Before moving forward, we need to assess:

• Feasibility: Can we build and operate a marketplace efficiently?
• Sustainability: How will we support ongoing operations?
• Ecosystem impact: How would this affect contributors, agencies, and users?
• Funding: How do we bootstrap this initiative when we don't have spare resources?
• Values alignment: Does this approach honor Drupal's open source principles?
• Governance: Who makes decisions about the marketplace and how?

We cannot and should not make these assessments in isolation. We need the Drupal community's involvement through:

• Research into similar marketplaces and their impact
• User experience design for the marketplace interface
• Technical prototyping of the marketplace infrastructure
• Financial analysis of various revenue models
• Legal research on open source licensing considerations
• Community input on governance structures

Our goal is to make a decision by DrupalCon Vienna, 6 months from now, or sooner if clarity emerges. We want that decision to reflect input from the CMS Leadership Team, the Drupal Association Board, Certified Drupal Partners, and the wider Drupal community.

We're chartering a Marketplace Working Group with stakeholders from across the Drupal ecosystem. I'm pleased to announce that Tiffany Farriss (Drupal Association Board Member) has agreed to lead this effort. Please join the #drupal-cms-marketplace channel on Drupal Slack to share your thoughts and follow the conversation.

Drupal's greatest strength has always been its community and adaptability. I believe that by thoughtfully exploring new ideas together, we can make Drupal more accessible and widely adopted while staying true to our core values.

Thank you to everyone on the Drupal Association Innovation Working Group and the Drupal CMS Leadership Team who took the time to review this post and share thoughtful feedback. I really appreciate your input.

Three months ago, we launched Drupal CMS 1.0, our biggest step forward in years. Our goal is ambitious: to reimagine Drupal as both radically easier to use and a platform for faster innovation.

In my DrupalCon Atlanta keynote last week, I reflected on the journey so far, but mostly talked about the work ahead. If you missed the keynote, you can watch the video below, or download my slides (56 MB).

If you want to try Drupal CMS, you can explore the trial experience, use the new desktop launcher, or install it with DDEV. If you're curious about what we're working on next, keep reading.

1. Experience Builder

Some of the most common requests from Drupal users and digital agencies is a better page-building experience, simpler theming, and high-quality themes out of the box.

At DrupalCon Atlanta, I shared our progress on Experience Builder. The keynote recording includes two demos: one highlighting new site building features, and another showing how to create and design components directly in the browser.

I also demonstrated how Drupal's AI agents can generate Experience Builder components. While this was an early design experiment, it offered a glimpse into how AI could make site building faster and more intuitive. You can watch that demo in the keynote video as well.

We still have work to do, but we're aiming to release Experience Builder 1.0, the first stable version, by DrupalCon Vienna. In the meantime, try our demo release.

2. Drupal Site Templates

One of the biggest opportunities for Drupal CMS is making it faster and easier to launch a complete website. The introduction of Recipes was a big step forward. I covered Recipes in detail in my DrupalCon Barcelona 2024 keynote. But there is still more we can do.

Imagine quickly creating a campaign or fundraising site for a nonprofit, a departmental website for a university, a portfolio site for a creative agency, or even a travel-focused ecommerce site selling tours, like the one Sarah needed in the DrupalCon Barcelona demo.

This is why we are introducing Site Templates: ready-made starting points for common use cases. They help users go from a fresh install to a fully functional site with minimal setup or configuration.

Site Templates are made possible by Recipes and Experience Builder. Recipes provide higher-level building blocks, while Experience Builder introduces a new way to design and create themes. Site Templates will bring everything together into more complete, ready-to-use solutions.

If successful, Site Templates could replace Drupal distributions, a concept that has been part of Drupal for nearly 20 years. The key advantage is that Site Templates are much easier to build and maintain.

3. A marketplace discussion

The first Site Templates may be included directly in Drupal CMS 2.0 itself. Over time, we hope to offer hundreds of site templates through a marketplace on Drupal.org.

At DrupalCon Atlanta, I announced that we'll be exploring a marketplace for Site Templates, including the option for Commercial Site Templates. We believe it's an idea worth evaluating because it could bring several benefits to the Drupal project:

1. Help new users launch a professional-looking site instantly
2. Showcase Drupal's full potential through high-quality examples
3. Generate new revenue opportunities for Drupal agencies and developers
4. Support Drupal's sustainability through a revenue-sharing model with the Drupal Association

You can watch the keynote recording to learn more. I've also published a detailed blog post that dives deeper into the marketplace idea.

Looking ahead

Drupal CMS has brought a lot of fresh momentum to the Drupal project, but we're not done yet! The rest of this year, we'll keep building on this foundation with a clear set of priorities:

• Launching Experience Builder 1.0
• Releasing our first Site Templates
• Expanding our marketing efforts
• Exploring the launch of a Site Template marketplace
• Building out our AI framework and AI agents

If you have time and interest, please consider getting involved. Every contribution makes a difference. Not sure where to begin? Join us on Drupal Slack. We're always happy to welcome new faces. Key channels include #drupal-cms-development, #ai, #experience-builder, #drupal-cms-templates, and #drupal-cms-marketplace.

As I said in the keynote: We have all the pieces, now we just need to bring them together!

Imagine waking up to discover that overnight, AI agents rewrote 500 product descriptions, reorganized 300 pages for SEO, and updated 9,000 alt-text descriptions on your website.

As you review the changes over coffee, you find three product descriptions featuring nonexistent features. If published, customers will order based on false expectations. Then you notice another problem: AI rewrote hundreds of alt-text descriptions, erasing the ones your team crafted for accessibility.

AI-driven content management isn't a distant scenario. Soon, Content Management Systems (CMS) may deploy hundreds of AI agents making bulk edits across thousands of pages.

The challenge? Traditional CMS workflows weren't designed for AI-powered editing at scale. What features should an AI-first CMS include? What safeguards would prevent errors? What workflows would balance efficiency with quality control? I'm outlining some rough ideas to start a conversation and inspire Drupal contributors to help build this future.

1. Smart review queues: scaling human oversight

AI-generated content needs different quality checks than human work. Current editorial workflows aren't optimized to handle its output volume.

I envision "AI review queues" with specialized tools like:

• Spot-checking: Instead of manually reviewing everything, editors can sample AI content strategically. They focus on key areas, like top-selling products or pages flagged by anomaly detection. Reviewing just 5% of the changes could provide confidence; good samples suggest the broader set works well. If issues are found, it signals the need for deeper review.
• Rolled-up approvals: Instead of approving AI edits one by one, CMS platforms could summarize large-scale AI changes into a single reviewable batch.

2. Git-like content versioning: selective control over AI changes

Say an AI translated your site into Spanish with mixed results. Meanwhile, editors updated the English content. Without sophisticated versioning, you face a tough choice: keep poor translations or roll everything back, losing days of human work.

CMS platforms need Git-like branch-based versioning for content. AI contributions should exist in separate branches that teams can merge, modify, or reject independently.

3. Configuration versioning: keeping AI from breaking your CMS

AI isn't just generating content. It is also modifying site configurations, permissions, content models and more. Many CMS platforms don't handle "configuration versioning" well. Changes to settings and site structures are often harder to track and undo.

CMS platforms also need Git-like versioning for configuration changes, allowing humans to track, review, and roll back AI-driven modifications just as easily as content edits. This ensures AI can assist with complex site management tasks without introducing silent, irreversible changes.

4. Enhanced audit trails: understanding AI decisions

Standard CMS audit logs track who made changes and when, but AI operations demand deeper insights. When multiple AI agents modify your site, we need to know which agent made each change, why it acted, and what data influenced its decision. Without these explanations, tracking down and fixing AI errors becomes nearly impossible.

AI audit trails should record confidence scores showing how certain an agent was about its changes (60% vs 95% certainty makes a difference). They need to document reasoning paths explaining how each agent reached its conclusion, track which model versions and parameters were used, and preserve the prompt contexts that guided the AI's decisions. This comprehensive tracking creates accountability in multi-agent environments where dozens of specialized AIs might collaborate on content.

This transparency also supports compliance requirements, ensuring organizations can demonstrate responsible AI oversight.

5. AI guardrails: enforcing governance and quality control

AI needs a governance layer to ensure reliability and compliance. Imagine a healthcare system where AI-generated medical claims must reference approved clinical studies, or a financial institution where AI cannot make investment recommendations without regulatory review.

Without these guardrails, AI could generate misleading or non-compliant content, leading to legal risks, financial penalties, or loss of trust.

Instead of just blocking AI from certain tasks, AI-generated content should be checked for missing citations, regulatory violations, and factual inconsistencies before publication.

Implementing these safeguards likely requires a "rules engine" that intercepts and reviews AI outputs. This could involve pattern matching to detect incorrect content, as well as fact verification against approved databases and trusted sources. For example, a healthcare CMS could automatically verify AI-generated medical claims against clinical research databases. A financial platform might flag investment advice containing unapproved claims for compliance review.

Strategic priorities for modern CMS platforms

I can't predict exactly how these ideas will take shape, but I believe their core principles address real needs in AI-integrated content management. As AI takes on a bigger role in how we manage content, building the right foundation now will pay off regardless of specific implementations. Two key investment areas stand out:

1. Improved version control – AI and human editors will increasingly work in parallel, requiring more sophisticated versioning for both content and configuration. Traditional CMS platforms must evolve to support Git-like branching, precise rollback controls, and configuration tracking, ensuring both content stability and site integrity.
2. AI oversight infrastructure – As AI generates and modifies content at scale, CMS platforms will need structured oversight systems. This includes specialized review queues, audit logs, and governance frameworks.