The AI-Driven DXP: New horizons for marketers, Martin Anderson-Clutz | EvolveDigital Toronto 2026
Summary
Martin Anderson-Clutz, Senior Product Marketing Manager for Drupal at Acquia, delivered a presentation at EvolveDigital Toronto 2026 arguing that the rise of AI-first internet fundamentally changes what a Digital Experience Platform (DXP) must do — and that simply grafting AI onto legacy CMS workflows is insufficient. He opened by framing the current landscape: customers increasingly begin their buying journeys inside LLMs rather than on websites, bot traffic already exceeds human traffic (with bad bot traffic projected to surpass human traffic by 2030), and AI crawlers are destabilizing websites not designed for such load. These shifts mean the website's role is no longer to educate visitors but to act as a conversion and brand validation layer for people who have largely made their decisions elsewhere — or to serve agents that may complete purchases on behalf of customers who never visit the site at all.
Anderson-Clutz proposed three core capabilities every forward-looking DXP must develop. First, an agile content engine: AI should automate research (identifying subtopics and competitive content strategies), assist with drafting and editorial tasks (tone normalization, grammar), and enable rapid post-publication optimization through automated audience segmentation and A/B testing. He cautioned that while AI can draft full blog posts, human oversight remains essential — citing an example where an AI chatbot issued 80–100% discounts that the brand had to honor. Second, an experience layer: websites in 2026 will primarily serve visitors who want to validate an existing purchase intent, so frictionless social proof (targeted testimonials, case studies, technical specs) matters more than broad education. He also highlighted inside-out layout generation (describing a layout inside the CMS and having AI build it) versus outside-in approaches (tools like Cursor or Lovable generating frontends that are then imported into the CMS), stressing the importance of brand-context management in both workflows. He advised sticking with segmented personalization for now rather than true one-to-one, as current models are not yet trustworthy enough for unsupervised content generation. Third, agent-friendly architecture: websites must treat APIs as the new UI, serving content in structured formats like JSON or Markdown rather than HTML, implementing well-documented agentic skills, and supporting agent-to-agent protocols such as MCP so that downstream AI systems can hand off tasks seamlessly.
For platform selection, he recommended choosing a DXP that is LLM-agnostic (avoiding lock-in as model rankings shift rapidly), natively agentic, composable, and orchestration-friendly (integrating with platforms like n8n). He made an explicit case for Drupal: its open-source, collaboration-driven module ecosystem allows faster iteration than paid plugin markets; its AI initiative has secured substantial funding and full-time contributors; and its 2026 roadmap focuses on production-readiness rather than proofs of concept. He noted Drupal had a working OpenAI API integration within three weeks of that API's public launch. Q&A touched on agentic content modeling workflows, the trade-offs between structured and unstructured content for AI ingestion, and editor preview challenges in headless architectures — with Anderson-Clutz pointing to Drupal Canvas as a promising tool for syncing component libraries between the CMS and headless frontends.
Key Takeaways
- •Redesign your website's purpose: it is now a brand validation and conversion layer for visitors who have already decided to buy, not an educational funnel — optimize accordingly.
- •Serve content in AI-consumable formats (JSON, Markdown) alongside HTML, and expose well-documented REST/GraphQL/JSON API endpoints so LLM crawlers and agents can ingest your content cleanly.
- •Implement agent-to-agent protocols (e.g., MCP) and define reusable agentic skills so your DXP can participate in broader AI-driven martech orchestration stacks.
- •Choose an LLM-agnostic DXP to avoid being locked to a single AI provider as model rankings shift rapidly — flexibility to swap models per task unlocks compounding progress.
- •Keep humans in the loop for any AI-generated content that reaches customers; unsupervised one-to-one personalization carries real financial and brand risk (e.g., chatbots granting 80–100% discounts).
- •Use AI to accelerate the full content lifecycle — topic research, drafting, editorial normalization, and post-publication A/B testing — not just one stage.
- •Prioritize segmented personalization in 2026; true real-time one-to-one personalization requires model maturity and governance infrastructure most teams do not yet have.
- •Treat bot traffic as a primary audience: AI crawlers already exceed human traffic and the gap will widen, so your infrastructure and content delivery must be hardened and optimized for them.
- •Open-source, collaboration-driven CMS platforms (e.g., Drupal) can iterate faster on AI integration than commercial plugin ecosystems because top contributors collaborate rather than compete.
- •Ensure your CMS enforces brand context when generating layouts with AI tools — whether inside-out (describe layout in CMS) or outside-in (import from Cursor/Lovable) — to maintain coherent brand experience.
Transcript
Okay. Hi everyone. Hope you're all caffeinated, smacked out. Lunch is soon. So, um you know, uh looking forward to to the gathering of everyone at lunch, but we've got a couple more sessions. Very interesting. Um it seems like AI is in the title of every talk I've been to today. Um and you know, actually good content. So, looking forward to uh hearing from Martin Anderson. He's with Aqua as and just as a new title, he's taken on senior product marketing manager for Drupal with Aqua. So Martin is...
