The DXP Catalyst Update - Apr 2, 2025

INTRO
Welcome to This Week’s DXP Catalyst Update

I recently attended the Optimizely Roadshow event in NYC, where I caught the keynote and main stage session. It was especially interesting to attend this during the same week as Adobe Summit - there were clear overlaps in themes, particularly around AI agents, multi-agent orchestration, and the broader evolution of agentic AI. Both vendors are staking a leadership position when it comes to embedding these capabilities across their product ecosystems.

In this week’s DXP Catalyst Update, I’ll share highlights from the main keynote. I won’t dive into the breakout sessions on CMS and commerce roadmaps or the cross-functional discussions on personalization and the future of experimentation. For the purposes of this newsletter, I’m focusing on Optimizely’s perspective on key trends, their broader vision, and a high-level look at what’s ahead over the next year.

EVENTS
Optimizely Roadshow NYC 2025

This event served as both a product showcase and a strategic vision session, anchored around the belief that we’re entering a new era in digital experience -powered by brand-aware AI agents, warehouse-native analytics, and end-to-end personalization.

The tone was clear from the opening keynote by CEO Alex Atzberger: if you don’t feel like you’re moving too fast, you may actually be falling behind.

Here’s a deeper dive into the biggest themes, product innovations, and practical insights from the day.

What It Takes to be a Digital Leader in 2025

According to Atzberger, digital leaders must build a foundation capable of absorbing constant change. With over 320 million terabytes of data created each day and AI adoption accelerating, digital transformation is no longer optional - it’s foundational.

Atzberger walked through 4 must do’s for digital leaders:

1. Content Supply Chain Optimization

Content is king - again. But now, it’s also fuel for GenAI systems. To scale personalized, intelligent experiences, you need a structured, scalable content pipeline.

Optimizely’s take: Create a foundation now so you can scale content production later, with platforms that support velocity, localization, and brand consistency.

2. AI-Driven Personalization

Personalization is evolving from segments to dynamic 1:1 experiences. Every visitor expects hyper-relevant content - instantly.

Optimizely’s solution: A dedicated Optimizely Personalization product that leverages AI to dynamically segment audiences and automatically deliver the best version of the experience.

3. Agents to Scale

The most forward-looking point: Optimizely is betting big on AI agents - autonomous tools that “work on your behalf.” These agents enable what Atzberger calls an infinite workforce, increasing both the volume and quality of execution across digital marketing and experimentation.

Why it matters: This is more than AI-enhanced workflows - it’s about full task delegation to intelligent agents that align with your brand and business logic.

4. Data-Driven Results

Conversion rates alone are no longer a sufficient metric. True digital leaders need to measure the real ROI of experiments, content, and personalization efforts.

Optimizely’s move: The acquisition of Netspring, now rebranded as Optimizely Analytics, brings warehouse-native analytics into the core platform - linking experimentation outcomes directly to your organization’s data ecosystem.

A New Optimizely One Flywheel

I really have liked Optimizely’s “flywheel” showing it’s “marketing operating system” Optimizely One as a continuous loop across these stages:

Plan → Create → Store → Globalize → Layout → Publish → Personalize → Experiment → Analyze

The recent platform updates - particularly the integration of Analytics - close the loop and turn insights into continuous optimization. Other updates included launching personalization as a dedicated solution (Optimizely Personalization) and accelerated AI innovation to create brand-aware agents (Optimizely Opal).

Also, the Optimizely team didn’t shy away from highlighting that they’ve now surpassed Adobe in the Gartner DXP 2025 report (check out my analysis) - just in case you missed the campaigns and the 5,000 LinkedIn posts about it. To be fair, the vision they set years ago - and more importantly, their ability to execute on it - has been impressive. Compare that to vendors like Sitecore, where the vision was solid but the execution fell short (in my opinion). Getting this right is incredibly difficult, and I think Optimizely has done a strong job. Their evolving vision continues to align well with the needs of digitally mature organizations.

From Hype to Hero: The AI Opportunity for Marketing

Optimizely’s CPO Rupali Jain gave a keynote focused on how AI is evolving from shiny object to business-critical infrastructure. Her framing:

• 2022–2024: “AI wrappers are dead. Long live LLMs.”

The market was flooded with generic AI tools with limited business application.

• 2025+: “LLMs are commoditized. The value is in the application.”

Competitive differentiation now comes from how well platforms apply AI to real business workflows.

The 3 Levels of AI Maturity

Jain introduced a model describing how AI matures across marketing organizations. Note that this was described as being within Optimizely’s product ecosystem.

1. Embedded AI

• Described as GenAI.

• The marketer interacts with Opal (powered by Google Gemini), built natively into their Optimizely One workflows.

• They define the criteria as pre-defined features, generic output, triggered by human action and one-time use.

• Examples include variation generation and content generation.

2. Enriched AI

• Described as GenAI plus your organization’s data across Optimizely One.

• Opal enriches its responses using real data (unique to that customer) from across Optimizely One to improve the quality of its output.

• Some similarities to Embedded AI except rather than generic output, there’s enriched/customized output.

• Examples include their Experiment Summarizer and Branded Content Generator tools.

3. Agentic AI

• Described as GenAI plus your organization’s data across Optimizely One, plus autonomous action.

• Opal applies logic and reasoning to take proactive action on your behalf. Their customers can define these specific to their needs.

• The criteria they define for this level of AI mature includes custom action, enriched/customized output, autonomous (triggered by chat, time, or event), always-on.

• Examples include their Experiment Ideation Agent and Industry Marketer Agent, which are discussed in the next section.

This “Agentic” model is what powers Opal, Optimizely’s new AI agent ecosystem.

Why Opal Beats Generic AI

Optimizely showed how they want to differentiate Opal from AI alternatives:

Opal is enriched, brand-aware, intuitive, powerful, and agentic - designed for marketers, not engineers.

Opal in Action: Real-World Use Cases

Opal isn’t a chatbot or a glorified autocomplete. It’s an AI agent platform built into Optimizely One, designed to act, not just assist. Jain and Optimizely’s VP of Solution Architecture went through a few demos of different use cases for Opal, the main thing highlighted was that it dramatically shortens time-to-execution.

Opal as a Campaign Manager (Coming H1 2025)

• Before: It would take a team 8-12 hours to produce a campaign (topic research, theme ideation, campaign brief creation, task and activity planning, setup and assignment)

• With Opal: The work was done in seconds. Opal was already trained on data and brand awareness, so it knows the formatting. Jain showed how Opal generated tasks like creating a blog post, case study, social media campaign, email marketing, and paid advertising, then it in action executing on these tasks. The idea is that the marketer only reviews and adjusts.

Estimated savings: $40-60K per year, assuming 100 campaigns/year and $100k/year per employee.

Opal as an Industry Marketer (Coming H1 2025)

• Before: The end-to-end process would take a team 8-12 hours (content audit and selection, industry specific angle ideation, plan and brief creation, adaptation and customization, layout and distribution)

• With Opal: Identifies highest performing assets, conducts research and identifies top trends, creates content aligned to brand, the marketer reviews and refines the output, Opal layouts content and publishes to the CMS. An example was taking an article on personalization and tailoring it for healthcare.

Estimated savings: $70-90K per year, assuming 10 hours/asset and $150k/year per employee.

Another demo was done for Opal as an Experiment Advisor. Several other agents were discussed as well, most of which are coming in H1 2025:

• Opal Website Analyzer - Creates experiments that be run right away. Marketers don’t need to come come up with things to analyze. Thousands of experiments have been run and fed into Opal as a baseline.

• Experiment Ideation Agent - Opal will dynamically and autonomously review site pages to suggest hypotheses and auto-generate variants.

• Experiment Setup Agent - Opal will proactively offer hypothesis and appropriate metrics to use for testing. It applies guardrails to prevent bad behavior and incorrect experiment setup from occurring. The value is that it sets up high quality tests quickly with less effort.

• Experiment Dev Agent - Opal will assist in building experiences for each variant of your experiment. It’s a fully no code and agent-assisted experience creation. The value is that it reduces dependency on tech resources and increases testing velocity.

• Experiment Insights Agent - Opal will automatically analyze each experiment and provide a simple summary in shareable format. It reviews follow-up experiments to validate patterns. The value is that it gets critical information out of each test and gives you confidence in each decision.

Agent Orchestration + Autonomy

Another announcement was the upcoming Agent Orchestration + Autonomy feature, slated for release in H1 2025. This new capability promises to change how marketing teams scale operations by enabling the creation and coordination of autonomous AI agents. Users will be able to build their own agents, design workflow-based automation, and enable collaboration between multiple agents - all within a visual orchestration interface.

The preview showed agents like Content Analyzer, Case Study Agent, and Content Planner working together to analyze inputs, generate content, and assign tasks, ultimately driving productivity without manual intervention. This could unlock infinite scale across marketing organizations - increasing both volume and quality of output.

The AI Era of Marketing Has Arrived

Optimizely closed out the keynote with a message that AI agents aren’t a future concept - they’re already reshaping marketing today. From their perspective, the most immediate and transformative opportunities lie within marketing functions, where agentic AI can unlock unprecedented scale. Unlike traditional assistants that support individual tasks, AI agents can operate autonomously, collaborate across workflows, and execute at a level of speed and precision that’s simply not achievable with human effort alone. For CMOs, this shift represents a leap beyond incremental gains - toward a future where scale, compliance, quality, and creativity coexist. The move from “assistant” to “agent” marks the next major frontier in digital experience.

Final Thoughts

I think one point Optimizely tried to get across was that the future of digital experience isn’t just headless or composable - it’s autonomous.

Whether you’re optimizing content workflows, scaling experimentation, or delivering 1:1 personalization at scale, AI agents will play a central role in how digital leaders operate in 2025 and beyond.

This is very much on par with some takeaways from the Adobe Summit.

One thing I appreciate about Optimizely’s Roadshow format is that, despite their growth, the events still feel intimate enough that their leadership is accessible and actively engaging with customers and partners.