Inside Edge Labs — making AI work, for humans
AI strategy and implementation for companies ready to move beyond efficiency and into genuine growth.
Start a conversation →Every major technology transition follows the same pattern. Companies start by using the new capability to do existing things more cheaply — digitise the brochure, put the catalogue online, make the call centre more efficient. Then a smaller group works out how to do things that weren't possible before. The second group tends to define what comes next.
We saw this with digital, then with mobile, then with social. The organisations that won weren't the ones that optimised fastest. They were the ones that found new sources of value — new capabilities, new markets, new ways of serving customers — that only existed because the technology made them feasible.
AI is following the same arc. Most corporate investment is concentrated in making today's operations faster and cheaper. That work is real and it's rational. But it means the growth layer — using AI to build genuinely new capabilities and reach new markets — is getting a fraction of the budget, the talent, and the senior attention. The executives who are starting to rebalance aren't abandoning efficiency. They've recognised that optimisation alone doesn't create competitive separation. Your competitors can automate the same processes with the same tools. The question is what you're building that they aren't.
Taking your existing processes and products and using AI to make them incrementally better. Better chatbots, smarter recommendations, faster reporting — useful, but you're still operating within the same business model.
Stepping back and asking: if we were designing for this customer need from scratch, knowing what AI can now do, what would we build? The starting point is the need, not the existing product. More often than not, it's a powerful extension of what you already do — a capability your customers need but that wasn't feasible to deliver until now.
The gap between these two approaches is where the growth opportunity sits. Most companies are well into the first. Far fewer have seriously explored the second.
Every platform shift has a period where the growth opportunities are visible but uncrowded — where most players are still focused on optimisation and a relatively small number are building new value. We saw this with digital in the early 2000s, with mobile around 2010, with social shortly after. In each case, the window lasted roughly three to five years before the landscape consolidated and the cost of entry rose dramatically.
We're about 18–24 months into mainstream AI capability. The technology is mature enough to build on, but most corporate investment is still concentrated in the efficiency layer. The growth space — new capabilities, new markets, new service models — is comparatively open.
That won't last. As more organisations exhaust the obvious efficiency gains and start looking for the next source of value, the growth layer will get crowded quickly. The question isn't whether AI matters — that's settled. It's whether you'll be exploring the growth opportunities while the field is relatively open, or competing for them once everyone else has arrived.
Most AI strategy work starts from one of two places: the technology ("what can AI do?") or the current operation ("which of our processes could AI improve?"). Both approaches anchor you to what already exists. The technology lens generates clever solutions looking for problems. The process lens generates efficiency gains. Neither reliably leads to genuinely new value.
The missing ingredient is primary, qualitative insight into what people actually need — not what they say they want in a survey, not what the data says they're doing, but the underlying need that often hasn't been articulated at all.
This is a capability that most strategy consultancies and most technology consultancies simply don't have. Strategy firms work top-down from market analysis. Technology firms work bottom-up from what the tools can do. Neither spends time sitting with real people, observing real behaviour, and uncovering the needs that no one has thought to address because they weren't technically addressable until now.
The approach that consistently produces the strongest opportunities follows a specific sequence, and the order matters.
Qualitative, often ethnographic research — going out and sitting with real customers or employees, observing their world, understanding their frustrations, their workarounds, the things they've stopped even trying to fix. This isn't a survey. It's not a focus group behind a mirror. It's primary insight work that uncovers needs the market hasn't articulated yet.
Beneath the surface-level requests ("I want this feature to be faster") sit deeper needs ("I need to make this decision with confidence and I currently can't"). The deeper need is where novel opportunity lives, because it's not tied to any particular solution.
Once you understand the genuine need, you apply the lenses to ask: how could this need be met in a way that simply wasn't possible before? Not "how do we use AI to improve the current solution" but "if we were designing for this need from scratch, knowing what AI can now do, what would we create?"
AI is often discussed as though it's a single technology. In practice, it's a collection of distinct capabilities, each of which unlocks different kinds of opportunity. When you look at a genuine human need through each lens, you see solutions that weren't possible before.
AI can now synthesise signals — behaviour patterns, context, timing, history — to anticipate what someone is likely to need next. The shift it enables is from reactive to proactive service. Where in your customer relationship would anticipation create disproportionate value?
The difference between an AI tool and an AI agent is who's driving. A tool responds when you prompt it. An agent pursues a goal on your behalf. This changes what a business can offer — not access to a platform, but outcomes delivered directly.
Specialised professional knowledge used to be scarce because it required expensive humans to deliver it. AI is making that expertise reproducible at scale. If you have deep domain expertise, you can now reach customers you could never have served before.
Individual lenses reveal opportunities. Combinations multiply them. Contextual Prediction plus Agentic Execution creates services that anticipate and act. The most powerful opportunities sit at the intersection — and they're unique to your specific assets and relationships.
Three failure modes show up again and again when organisations try to move from AI experimentation to strategic AI capability.
Individual teams explore AI independently, generating scattered results that can't be compared or built upon. There's no shared vocabulary, no coordination, and no way to identify which experiments deserve investment.
With so many possible applications, teams default to whatever's easiest to try rather than what would drive the most value. Effort is dispersed across dozens of small experiments, none of which gets enough attention to succeed.
First attempts with AI are rarely polished. Without structured iteration and support, teams hit friction, conclude that AI isn't ready for them, and quietly abandon tools that — with a little more patience and a better methodology — could have transformed how they work.
We've built and run innovation programmes inside large organisations. We've seen what kills them. It's almost never the quality of the ideas.
Random experiments without strategic focus. Teams generate interesting ideas that don't connect to anything the business actually needs to do next.
Innovation as a separate function — a lab, a skunkworks, a tiger team — that gradually loses its connection to commercial reality. The work becomes impressive but irrelevant.
An innovation programme that requires permanent, expensive external support becomes a target the moment budgets tighten. If it hasn't built internal capability by then, everything it created disappears.
Every element of how we work is designed to avoid these three failure modes.
Inside Edge Labs brings together three things that most AI consultancies treat separately: primary insight work that uncovers genuine human needs, the strategic frameworks to identify AI-first ways of meeting them, and the human-centred facilitation that gets organisations aligned and moving.
The approach follows a Double Diamond structure. The first diamond opens up the problem and opportunity space — deep discovery, qualitative insight, understanding what people actually need. It converges on the highest-value opportunities through structured analysis and AI Capability Lens mapping. The second diamond opens up the solution space — designing, prototyping, and proving AI-first solutions grounded in those real needs.
Before any AI work begins, we map four competing dynamics: the pain of the current way of working, the pull of what's possible with AI, the anchors that keep people attached to existing processes, and the anxiety about what change might mean.
At every checkpoint, the team makes a deliberate decision about each experiment. Stick — it's working, scale it. Twist — it's promising but needs adjustment, iterate. Fold — it's not working, kill it and redirect resources. Named after a card game decision, it normalises the act of stopping things that aren't working.
Structured rounds of rapid qualitative testing. The first round tests what we believe is the right answer. The second iterates on what we've learned. The third deliberately pushes into radical territory — testing provocations that stretch what's possible. The fourth brings it all back together into something grounded, ambitious, and informed by the full range of what we've explored.
Most AI consultancies start from the technology or from your existing processes. We start from your customers and your people — primary, qualitative research that uncovers the needs nobody has thought to address, because they weren't addressable until now.
During the design phase, we work alongside your team in their actual workflow — not observing from outside, but embedded in the work. This is how you capture the tribal knowledge, the workarounds, and the real process that determine whether an AI solution will actually work in practice.
The explicit goal of every engagement is to build your internal AI capability. Through the Build Lab, your key innovators learn by actually building the solutions identified through discovery — working hands-on with specialist build partners to prototype real AI applications. The measure of success is that you don't need us anymore.
We've built innovation programmes that succeeded and ones that failed. We've navigated the competing incentive structures, the internal politics, and the gap between executive ambition and middle-management reality. We won't pretend this is easy. We'll make it structured, achievable, and focused on outcomes that matter.
Every engagement is modular. You can begin with a focused sprint and expand as you see value — or commit to the full journey from the start.
Diagnose where you are and align on where AI can create the most value. Includes AI readiness assessment, 4 Forces analysis, and strategic path recommendation.
The full first diamond — from diagnosis through to a prioritised opportunity map. Includes immersion workshops, capability lens analysis, and workflow story mapping.
Hands-on design and prototyping of your highest-priority AI opportunities. Includes structured concept testing cycles and Stick/Twist/Fold validation.
Your key innovators don't just observe — they build. Working with specialist partners, your team constructs real AI solutions with expert support.
The complete journey from diagnosis to scale. Everything above — including Build Lab — plus scaling support, champion development, and quarterly retrospectives to sustain momentum.
The work is hands-on and practical. Sessions are structured and facilitated, but designed so your team genuinely enjoys them — building confidence alongside competence, not sitting through slide presentations.
Honesty is built into the process. Not everything is an AI problem, and we won't pretend otherwise. Where AI can create real value, we'll find it. Where it can't, we'll tell you — and we won't recommend work you don't need.
Every stage produces concrete artefacts your team can use immediately — prompts, workflows, prototypes, decision frameworks. The measure isn't whether the strategy document is impressive. It's whether anything actually changes on Monday morning.
Rose has navigated the inside of Amazon, Sky, and Trustpilot, and the hardest part of any technology transformation is never the technology — it's getting people aligned, confident, and moving in the same direction. That's where the real work happens, and it's where we spend most of our time.
Fifteen years leading product, innovation, and AI strategy for brands including Sky, Amazon, Trustpilot and LOVEFiLM, alongside numerous start-ups and scale-ups. A product leader who's built and shipped at scale across media, e-commerce, and consumer technology.
A founding member of Sky Labs, Sky's innovation team — developing collaborative approaches for rapidly identifying growth opportunities and prototyping with new technologies, and learning first-hand what makes corporate innovation programmes succeed and what kills them. Trained in design sprints with AJ&Smart in Berlin, with deep expertise in product management, design thinking, and the organisational dynamics of technology adoption.
Co-founder and Chair of the 2,000-strong Brighton AI network, and founder of the No Code AI communities. She also sits on the Innovation Council for BIMA, the British Interactive Media Association.
Now running Inside Edge Labs — helping ambitious companies use AI to build what's next, not just optimise what exists.
If you're ready to move beyond efficiency and into genuine growth — if you want to find the opportunities your competitors haven't seen yet — let's have a conversation. Drop a few lines below and we'll come back to you within a day or two.