Optimizely

Like most content marketers using AI, I've been training my tools to write in my tone of voice. Great news : it worked.  Bad news : apparently I say "actually" approximately 400 times per doc. My AI has absorbed this completely. It now opens sentences with it more confidently than I do, which I didn't think was possible, and yet... here we are. Also, food metaphors. Everywhere. Do I constantly write copy when I'm hungry? Because it bloody well seems like it. The em dash conversation (/ heated debate ) Okay, so here's the one that really got me. I love em dashes. I have always loved em dashes. They've been in my content guidelines at more than one job, and part of that is professional convention, but in all seriousness... part of it is that I think punctuation is beautiful and I like how it makes a sentence look. In my totally humble (and perhaps a lil bit content-nerdy) opinion, the em dash is a very useful and very pretty piece of punctuation. Alas , then the great em dash witch

The "Model Wars" have entered a new phase. In the early days of generative AI, success was measured by who had the most parameters or the lowest latency. But as AI moves from the lab into enterprise product stacks, a new challenge has emerged: delivery. It's no longer enough to have a great model. You need a way to deploy it, test it, and govern it without breaking your product. To scale AI, you don't just need better GPUs; you need agentic infrastructure. At the center of that infrastructure is a critical component that most teams are overlooking: the feature flag. In traditional software, feature flags control whether code runs. In AI systems, feature flags control how intelligence behaves. The delivery layer: Scaling beyond the model We are moving away from monolithic AI, where a single model powers an entire feature, to agentic systems where multiple models, prompts, and tools interact dynamically. The model is just one part of the equation. The real bottleneck is the delivery laye

You've just spent three weeks building the case for a new AI solution for your marketing team. The deck is beautiful. You've got trend graphs, competitor analysis, a whole slide on how the tool will "transform your content output." You're ready. Then the CFO is like: What's the payback period? Why this one and not the cheaper tool we've already got? Cue: s i l e n c e. Here's the thing: Your instincts as a marketer are actually working against you in that room. You're trained to lead with reach, resonance, and relevance. You think in audiences and narratives. The CFO thinks in ratios and risk. Neither is wrong, but you're essentially just speaking completely different languages. In fact, only 22% of marketers say they have the measurement insights needed to justify their value to the CFO. The fastest way to get budget approved? Stop trying to get them excited about marketing, and start speaking their language. Marketing leaders, welcome to your translation guide. Here's how to speak CF

Here's what the Opal U | AI Marketing University April cohorts proved: five days is enough. Enough to build an ABM team from scratch. Enough to retire the Slack thread where everyone asks "what's happening with our tests?" Enough for a solo web marketer at a medical device company to absorb a colleague's entire workload during maternity leave and not drop a single blog post. The April cohorts ran across four weeks and pulled in 51 builders from 40+ companies. Salesforce, KPMG, FedEx, Vodafone, NatWest, HMRC, Birkenstock, Canon Europe, ODEON Cinemas, ClassPass, Republic Services, Sandvik. Companies you'd expect. ... and a few you really wouldn't — a UK government department, a retail bank, a medical device company, an indirect procurement manager who had no business being this good at this. Cohort 7 ran as an experimentation specialty track, and delivered 264 agents in five days. That's the highest volume of any cohort to date.  ~600 agents total, but (just) five stories belo

My interest in statistics started with a joke I heard as a kid. Someone asked: "What's the chance of seeing a dinosaur on the metro?" Being a data-minded person, I'd say "0%. Dinosaurs are extinct." They'd say "No, it's 50%. You either see one or you don't." The logic felt oddly airtight, yet completely wrong. I'd push back : "I've ridden the metro more than a hundred times and never seen a dinosaur. So, how can it be 50%?" That small puzzle stuck with me, and it captures something essential about how we think about probability. Today, I want to use that intuition to break down the three statistical approaches used in A/B testing — frequentist, Bayesian, and sequential and explain how they differ in practice and help you decide which one fits your team best. First, a quick grounding in probability Probability is the branch of mathematics that quantifies uncertainty by measuring how likely an event is to occur. It's typically calculated by dividing the number of favourable outcome

Let's start with some numbers that should make you at least a little uncomfortable. 91% of organizations have no formal structures or processes in place to govern AI use internally. 71% of companies include ethical principles in their AI strategies — but only 36% have actually formalized those principles into policy. And of the companies that do have an AI policy? Only 41% make it accessible to employees or require acknowledgment of it. Long story short: most organizations are running AI programs that exist almost entirely on vibes. This isn't a dig. It's the reality of where most marketing teams are right now — and it's exactly why 95% of AI projects fail to scale. Not because the technology doesn't work. Not because people aren't using it. But because there's nothing underneath it. No ownership, no guardrails, no shared standard for what 'good' looks like. " Building an effective AI program means sitting with the pendulum swings, the false starts, and the moments where yo

Hands up if you (and the rest of your marketing team) have been here: A team gathers, someone asks "what would you want an AI agent to do for you?" and within minutes the room is full of ideas. People want to be shouting their ideal marketing agent (and often, some non-marketing agents too) so the world can hear it. And this totally  f e e l s  like a productive session. Except , nothing changes. Enthusiasm behind dream agents has never been an issue. But when you're getting started with agents, you need to rethink the question. When you ask people what they want AI to do, you're going to get wish lists; ideals or solutions in search of problems. This leads to agents being built that demo beautifully, but don't actually deliver anything for your team. But you want to start with the problem-fighters first. No wonder 74% of companies are yet to show tangible value from AI, right? Here at Optimizely, we've spent the last several months running AI agent discovery workshops across

Every marketer wants the same thing: to make every prospect feel like the only prospect. You know the pitch you'd give if you had time. The one that leads with their exact pain point, references the initiative they're probably running right now, speaks to the CFO differently than it speaks to the head of ops. The one where the landing page they land on doesn't feel like it was built for a segment of 50,000 — instead, it feels like it was built for them. Only them. You know exactly what that would do to your conversion rates, and you know that it's the ideal (read: right ) way to work. But alas, you just can't produce it. Not at the volume your pipeline demands anyway. The math that breaks most marketing teams Here's the problem laid (painfully) bare: 4 different personas x 500 target accounts = 2,000 tailored assets. O u c h. And that's before you factor in message variants, campaign phases, or the fact that half of those accounts need refreshing the moment a new signal comes in

I cried at a marketing presentation once. Twenty-something, bad coffee, someone else's work on someone else's screen. The room fluorescent and forgettable. And something in it found me anyway, reached right through the slide deck and pressed on a place in my chest I did not know was open. I have been chasing that feeling ever since. Something crystallized in me that day. The kind of crystallization that does not ask permission and does not reverse. It metamorphosed into the marketer I am now. Into the reason I stood on a stage in Copenhagen, years later, and said the part the room had been carrying for too long without anywhere to put it. We have spent so long optimizing the machine that many of us have forgotten the point of it all. I am going to tell you about that room. About what I think happened to us, and what I think is waiting on the other side of this particular moment. About agents and change and the psychology underneath all of it. But first, I want to talk about why we are

Quarterly Report filed 2026-05-01

Material Event filed 2026-05-01