What Every Digital Leader Needs to Know from Opticon 25 New York

Attending Opticon 25 in New York confirmed what we're already witnessing across industries: a fundamental shift in how digital experiences are built and consumed. AI isn't just changing what's possible—it's changing what customers expect, how teams work, and what competitive advantage looks like. The organisations that thrive won't necessarily be the ones with the most advanced AI, but those that most thoughtfully integrate AI into their culture, processes, and customer experience. The question isn't whether these changes are coming—it's whether you'll lead them or be disrupted by them. Here are the key themes every digital leader needs to understand.

Your Website is No Longer Your Front Door

The role of a company website is shifting from the main digital entry point to a content source. LLMs and AI agents are becoming the primary channel, meaning your customers will increasingly interact with AI assistants that pull from your content rather than browsing your site directly. This fundamental shift requires rethinking how you structure and organise your content. Instead of solely optimising for human visitors navigating pages, you need to ensure your product data, policies, and brand information can also be easily accessed via APIs and understood by AI systems. This transition involves significant technical and content architecture work—from restructuring information hierarchies to implementing site-wide optimisation that serves both human and AI consumption.

At Opticon, we saw how Optimizely are providing tooling to automate much of this heavy lifting through their Opal product, learning from existing content patterns and language used across repositories to help organisations adapt their architecture for AI accessibility. The companies preparing for this transition are those asking whether their content architecture supports AI consumption and whether they're ready for conversational AI to handle first-touch customer interactions.

The Language Gap That Derails AI Customer Experience

Preparing for your website to be consumed by AI systems is important, otherwise it can amplify systemic issues with the website if they're not addressed first. Your content organization must match how customers think and speak, not how your business is structured internally, because AI agents trained on your content will communicate using the patterns and language they learn from your systems. This language gap can completely derail AI customer experiences when agents use internal jargon, follow organizational hierarchies that confuse customers, or navigate information in ways that don't match customer mental models. Our recommendation: audit your content architecture through a customer lens before implementing AI agents.

Optimizely customers mentioned how they could use the analytics capabilities within Opal to identify where these language gaps exist by surfacing customer behaviour patterns and then use their experimentation platform to test different content

taxonomies and organisational hierarchies with real customers to fine-tune the foundations that AI agents will inherit. Success requires ensuring site navigation reflects customer thinking rather than internal org charts, writing product descriptions in customer language rather than industry jargon, and organizing information in ways that make sense to the people you're trying to serve.

Fear Stifles AI Innovation

The biggest risk organisations face isn't AI failure—it's paralysis from overthinking or waiting for ideal conditions. Fear stifles innovation by preventing teams from experimenting, learning, and building the confidence needed for larger AI initiatives. Companies succeeding with AI start small, learn fast, and build trust through validated wins before tackling bigger challenges, creating psychological safety for experimentation and learning.

Optimizely's Opal enables this incremental approach by allowing teams to start with simple AI-assisted tasks from within the UI and gradually build capability and confidence. As teams become more comfortable, Opal can progress to running end-to-end processes—such as on-page experiments—without requiring development effort or manual intervention, transforming what was once a technical bottleneck into an autonomous capability. This progression recognises that competitive advantage goes to those willing to iterate and build confidence through practical experience rather than waiting for perfect conditions.

AI Mandates Fail Without AI Literacy

C-suites are mandating that organisations can't increase headcount without first proving AI can't do the job, but this pressure from leadership only succeeds when teams have the skills to make it work. AI success is fundamentally "a people issue, not a technical issue," and organisations that simply deploy AI tools without systematic literacy training see disappointing results. The key is teaching teams how to "talk AI"—not just asking questions, but collaborating with AI as a thought partner while avoiding "satisficing" (settling for good enough instead of optimal outcomes).

At Opticon, it was clear that Optimizely has designed Opal specifically with non-technical teams in mind, recognising that AI literacy needs to be built through accessible interfaces and guided interactions rather than technical complexity. Executive AI mandates fail without ground-level AI literacy. The companies succeeding are those whose leaders model AI usage, create safe spaces for learning and experimentation, and invest in developing capabilities that turn top-down pressure into genuine competitive advantages.

The AI Agent Promise Depends on Data Quality

The expectation with AI is that gains become exponential as you progress from individual usage to setting up agents to deploying teams of agents working together, with expectations of up to 20x improvements compared to a single user. However, this exponential progression isn't achievable if the foundations aren't in place.

Every successful AI implementation story shared at Opticon started with clean, structured, accessible data. The companies seeing 79% velocity increases haven't just deployed agent orchestration platforms; they've invested in data architecture first, ensuring their customer data isn't scattered across disconnected systems and that product information, customer insights, and performance data can be easily surfaced for AI agents to use. Context is air for AI, and we saw how Optimizely's Opal addresses this challenge by automatically enriching agents with contextual data from across systems without requiring manual input, while connecting to third-party tools to ensure agents have access to the full data ecosystem they need to make intelligent decisions.

Without high-quality, well-organised data, even sophisticated agent teams can only make poor autonomous decisions. The 20x promise of AI agent orchestration requires rethinking entire service delivery models around what becomes possible when multiple AI agents have access to the structured, contextual data they need to work together intelligently without constant human intervention.

Composable Systems Supercharge AI Agent Integration

Flexible, API-first architectures have become essential for unlocking AI agent potential. Composable systems supercharge AI agent integration by enabling rapid deployment of agents that can tap into existing system capabilities, facilitating easy connections between multiple AI services and seamless scaling as capabilities evolve. Monolithic systems create bottlenecks that slow AI agent rollout and limit collaborative potential between different AI tools, primarily because they lack the granular API functionality that agents require. Composable architectures allow organisations to quickly test new agent combinations, integrate emerging AI capabilities, and adapt their tech stack as the AI landscape changes.

The companies positioning themselves for success understand that their architectural choices today determine how quickly they can deploy AI agents tomorrow, how easily they can connect new AI services to existing workflows, and ultimately whether they'll be able to capitalise on the rapid evolution of AI agent technologies.

It's Never Been Easier to Embed Experimentation in Your Company

The pace of change at the minute is relentless, and in an age of uncertainty, testing is key. The companies thriving are making experimentation a core capability across all teams, not just dedicated optimisation specialists. This shift means moving beyond treating testing as a special project to building it into regular workflows, giving non-technical teams the tools and training to run their own experiments, and measuring tool adoption as well as conversion improvements.

Optimizely is accelerating this democratisation through Opal AI, which enables users to create new UI elements and alternate page versions directly through the Optimizely interface—elements that seamlessly integrate into code repositories without requiring development resources. This capability allows non-technical team members to create test variants and run experiments independently, removing the traditional bottlenecks that kept experimentation locked within technical teams. The transformation requires viewing experimentation as a fundamental business discipline rather than a marketing department function, embedding the testing mindset throughout your organization's daily operations.

Want to get the most out of Optimizely's latest AI features? We're offering free Opal enablement workshops to help your team harness the power of AI-driven optimisation.