Your brand is the only thing AI can't copy: Opticon 2026

My biggest takeaways from Opticon 2026

You know the feeling - every conference, every LinkedIn post, every pitch seems to be competing to convince us that we are roughly five minutes away from handing over the keys to an army of autonomous agents while we sip coffee and watch revenue roll in.

What I appreciated about Opticon 2026, Optimizely’s annual flagship conference, was that, beneath all the AI hype (and yes, it was a lot of hype), there was a surprisingly human theme running through almost every session.

Ironically, the more sophisticated AI becomes, the more important it is to understand what makes your business uniquely human and where AI remains fallible.

Innovation isn’t a process; it has to be a culture shift 

One standout keynote session came from ASOS.

Whilst much of the industry is racing towards automation, ASOS' approach felt refreshingly grounded. Their perspective wasn't about building a better ecommerce engine but rather about building a better stylist ‘companion’ for their customers. 

Anyone can use AI to generate product descriptions, create landing pages or optimise campaigns. What AI can't replicate is the years of accumulated brand equity, cultural understanding, taste and customer trust that sits behind a great brand. ASOS knew that to fundamentally shift their performance, they’d need to leverage this, and that micro-optimisations of 1-2% alone wouldn’t get them to their ambitious goals.

ASOS shared how they'd run an aggressive programme of experimentation over the previous six months, releasing hundreds of features and changing the way teams worked to support much faster innovation. Some of those ideas won. Some lost. Some lost initially and were improved afterwards. 

The important aspect was that ASOS embedded a ‘fix-forward’ culture; if an experimental feature failed in testing, they didn’t bin it – they released it anyway. They didn’t wait for post-test results to make decisions and deploy, but rather released and analysed retrospectively. This helped prevent features customers started to use and love from suddenly disappearing while the results were analysed behind the scenes and business decisions were made about investing in that feature as a permanent addition. These features stemmed from ASOS’ non-negotiable ‘stylist first’ brand-led approach, and ASOS were willing to take the gamble that these ideas would work because they were true to the brand. Ultimately, the gamble paid off, though I do wonder whether this is an easier play when your brand is as large and well-established as ASOS.

My takeaway: spend more time understanding what fundamentally differentiates your brand. As AI inevitably lowers the barrier to creating content, experiences, and products, the things that become most valuable will be the things competitors simply can't copy. 

The economics nobody wants to talk about

One of the more pragmatic discussions came from Optimizely's focus on token economics - the less glamorous side of AI that focuses on measuring, governing and optimising AI consumption to ensure it delivers real-world ROI.

Everyone wants to talk about what AI can do. Far fewer people want to talk about what it costs to run at scale. And with 3 trillion tokens used by Optimizely customers in the last year, it’s getting harder and harder to sweep to the side. 

How are Optimizely supporting customers here? Well, they announced a suite of marketing-specific post-trained models designed to handle common marketing tasks more efficiently, alongside Mark-Bench, their new marketing AI benchmark. Early testing suggested they were achieving up to 10x lower costs whilst achieving a 7% quality increase vs. Claude. 

My takeaway: Many businesses are still experimenting, with pilot programmes and few, sporadic licences dotted around their company. But AI adoption tends to follow the same pattern as most successful technology, and once people start finding value, usage will explode. Organisations need to start talking about governance, cost management and infrastructure now to be in a much stronger position than those treating AI as a limitless resource. 

Marketer experience: please don't automate away curiosity

One of Optimizely's biggest announcements was its new Virtual Teammates. On paper, they're incredibly impressive. AI-powered teammates capable of analysing performance, surfacing opportunities, and carrying out marketing tasks.

But I found myself asking a slightly different question - if AI does more of the analysis, do marketers spend less time learning? The best marketers I've worked with have never just looked at dashboards. They've obsessed over them. They know their numbers. They spotted tiny changes in trends before anyone else. They've developed instincts because they've spent years immersed in the data.

Curiosity has always been one of marketing's greatest superpowers. As more repetitive work becomes automated, we need to ensure we're removing admin rather than removing understanding.

My takeaway: The boring stuff absolutely should become easier, but not at the expense of the curiosity that drives genuinely great marketing. Find the balance, start small, and check in with yourself: what are you losing when you hand over a task to AI?

The boring stuff

The most consistent message throughout Opticon wasn't actually about AI, but you guessed it - it was about data. I know that's not particularly exciting.

Nobody wants to stand on stage talking about connectors, governance structures, data hygiene or source management. But almost every successful AI story at Opticon ultimately came back to those foundations.

The promise of one-to-one personalisation is incredibly exciting. For years, personalisation programmes have often stalled because organisations simply couldn't operationalise them at scale.

Now, AI is beginning to remove some of those barriers, and Optimizely is promising genuine 1:1 limitless personalisation. But the challenge remains – personalisation, even if powered with AI, can only be as effective as the foundations beneath it.

Poor data will still create poor outcomes, and disconnected systems will still create disconnected experiences. Messy governance will still create risk.

The brands doing the boring work now will almost certainly be the brands seeing the biggest returns later. Both UNICEF and Sompo reinforced this point in different ways. The message was remarkably consistent: start by automating repetitive, low-risk processes. Learn. Build confidence. Improve the foundations. Then scale.

These are the foundations we’re working with our client partners on as part of their GEO strategies; ensuring technical readiness, a strong, data-led content strategy with consistent optimisation and scaling based on learnings and wins. Our whitepaper ‘The Persuasion Factor’ reinforces the importance of strong data structuring; our research found that as customer prompts get more specific, AI models favour brands where product data is complete, consistent and explained. 

Frankly, it's also a much more effective way to bring people along on the journey if you concentrate on low-risk processes first. Most people aren't worried about AI helping them with tedious administrative work – they’re terrified of being replaced. Leading with augmentation rather than replacement creates a far healthier path towards adoption.

As Alex Atzberger, Optimizely CEO, puts it:

"Cutting headcount is not a growth strategy. Unlocking talent is." 

It felt like one of the most powerful lines of the entire event.

A new challenger approaches: Agent Experience (AX)

Opticon used to entirely centre on the marketer; how could Optimizely support brilliant marketing experiences – or MX, if you will. MX ties naturally to CX, and Optimizely has a great product set to support that too. But guess what? We now also must also consider AX - agent experience.

Cloudflare CEO Matthew Prince has reported that between 52% and 62% of internet traffic now comes from bots. So, we’re entering a world where AI agents increasingly evaluate, consume, and interpret digital experiences before humans ever see them.

It radically changes how we think about discoverability, how we think about content and how we think about optimisation. At Opticon, Optimizely claimed they’re ahead of the competition here; their product offering doesn’t just identify AI agents on your site, but Virtual Teammates use those insights to optimise the site for better LLM crawlability. Robots telling robots how to better robot. 

But at the end of the day, a strong AX will always lead back to the human experience; whether its providing a marketer precious hours back to focus on creativity, or shortcutting initial research and product consideration for customers. So, the interesting challenge for brands is now going to be how MX, CX and AX strategies are balanced to provide the best outcome for humans (rather than our robot overlords!).

Parting thoughts

The talks that resonated with me most were the ones talking about what AI can’t replace; where boundaries are being set and principles set in place. 

• UNICEF refused to ever use AI to generate imagery, and didn’t feel confident in turning their experimentation programme over to autonomous optimisation products; instead favouring the cultural shift across markets that embedding experimentation would provide.

• Bain has a hard-and-fast rule – humans always write content.

• Optimizely’s rebrand came from an agency that is obsessed with ‘different’; something AI can’t manage.

I left Opticon feeling more optimistic than when I arrived. The marketers, brands and businesses who understand where humans create value and where AI creates leverage will be the ones to pull further ahead – I could feel it. The companies chasing AI for AI's sake will struggle.

The companies willing to do the boring groundwork, build strong foundations, and double down on the parts of their brand that genuinely can't be copied will be the ones that win. 

And therein lies some lovely irony – the more AI becomes embedded into marketing, the more integral the human becomes. 

If you'd like to find out more about how true can support you with Optimizely, drop us a note here.