User Agents to User Agency: Why Optimizely is Right for Experience Innovators in an AI-disrupted Landscape 

Driving operational effectiveness with Opal AI 

I've been writing recently about how user needs are shifting in response to AI capabilities, particularly 3rd party AI assistants like ChatGPT, Gemini and Claude. And I've been laying out how CMSs need to adapt to new user needs from a platform-agnostic perspective. Now I'm going to shift to a more opinionated stance: here’s why I think Optimizely are already leading the shift into AI-driven customer experience, and are set to extend that led. 

Opal is leading this year's Opticon agenda (check out Mark Starling's report from Opticon New York to preview what's coming to Opticon London on September 30^th). The first version of Opal was announced in 2023, but most users encountered it primarily as a contextual CMS interface to Gemini, which didn't add enough value to lure them away from other AI assistants.  


But a fully revamped Opal started rolling out this summer, and it's a completely different beast. Out of the box, it provides a much deeper toolkit - properly integrated across the Optimizely One stack with functions that perform actions rather than just creating suggestions. 


What’s more impressive though, is Opal’s tool and agent creation capabilities. We saw this in action during a recent hackathon, in just a couple of days we were able to bring performance data from Microsoft Clarity into Opal, and within a couple of hours we had Opal’s analysing that data recommending optimisation changes and experiments. The new Opal just goes further, harder - setting up custom agents, with defined parameters and connected tools, gives you the ability to build custom AI applications that are directly hooked into your experience stack for ease of implementation.

Organisational commitment to adoption 

The UK Department of Trade and Industry recently trialled Microsoft 365 Copilot with disappointing results - staff were given minimal training, leadership wasn't committed, and they had just two months to find valuable use cases in a sceptical culture. The predictable failure made (misleading) headlines as proof that "AI doesn't improve productivity."  

 
But the real lesson? You can't just deploy AI tools without proper strategy, training, and organisational commitment. Senior leadership is critical to the AI adoption process. You have to provide the resource to properly build capability, the authority to redefine process, and the leadership to carry through change at pace. 

 
It's only fair to acknowledge that most CMSs are pushing through AI as toolkit implementations. Adobe have Sensei AI and Sitecore have Stream, and there are many others. But based on the latest Opal iteration and immediate roadmap, I think Optimizely have the edge in flexibility, extensibility and composability in AI. That's exactly what brand leaders need to make AI an operational success. 

AI in user experiences 

Another piece of research came out last week: OpenAI's study, with Duke and Harvard Universities, on how people are actually using ChatGPT. One particular statistic stands out: non-work messages now account for more than 70% of user engagement. That rise in consumer usage, combined with a meteoric rise in consumer engagements via Google's AI Summaries, starts to look mightily like a tipping point. AI is about to become one of customers’ primary tools, and you can choose whether to provide your own customer AI experiences and showcase your brand in third party AIs, or alternatively to be ignored by AI and bypassed by all of the customers using it. 

 
That means you need new experiences, new customer engagement models, not just internal tooling. It means you need content platforms that aren't limited to monolithic content models based on pages and articles as static outputs, but which enable content to be dynamically transformed as unique expressions that meet the needs of individual audiences, whether consumed directly by humans or intermediated through external AIs. 

 
Optimizely get this to an extent I haven't seen anywhere else, yet. They're already releasing out-of-the-box functionality addressing emerging SEO and content visibility needs that create different content expressions: not just simple content and asset tagging, but transforming content into Q&A pairs, into LLM-focused summaries. Optimizely Graph is already powerful enough to power custom expert knowledge chatbots (like CIPD's Buddy), but we know that they're looking into the future of content graphing and agent-focused interfaces to create entirely new experiences. 

 
Recently I predicted the CMS functionality that will become core requirements to deliver next generation, AI-driven user experiences. Nobody meets these yet, but I'd be surprised to find anyone getting there faster than Optimizely. And they're already providing enough innovation at pace to fully occupy any content and experience creation teams out there. 

 
If you're going to be at Opticon London and you'd like to know more, catch me on the 26 DX stand. Or contact us to arrange another time.