Greg Coleshill, Chief Commercial and Operating Officer
Not sure I should be saying this, but I'm going to anyway.
For a pretty long time, I had absolutely no idea what my co-founders were actually talking about.
I’d sit in meetings with Mark and Michael and they’d be deep in the weeds. They’d be talking about "MCP Servers," "retrieval context," and "non-deterministic outputs." I’d nod along. I’d get bits of it. But internally I was thinking, "This sounds great, but how on earth do I explain this to a COO who just needs their team to move faster?"
It took me a while to really work it out.
In my defence, it wasn't just me. It was all of us. Translating the technical into clearly understandable concepts is always the hard part. But taking a technical solution and articulating the real-world benefit is vital.
Because we aren't building software.
We’re building a way to hire intelligence.
Customers aren't working with a "program."
They are working with intelligence.
I've posted before about the role of words in shaping the future of AI. My focus then was encouraging people to stop thinking about "agents" and start thinking about "teammates."
The mental model matters because it builds trust it enables and powers adoption. But it goes beyond that. The framework we provide doesn't just govern how people perceive a solution; it provides a control framework they already understand. It’s the beginning of how people move from "personal productivity" to building organisational assets that actually help the business.
A customer recently said something that stuck with me: "We’ve managed non-deterministic intelligence for hundreds of years. It’s called people."
Operations leaders already know how to manage and ensure controls around people. We should use that same muscle for AI.
The big mistake I see everywhere right now is firms treating AI like a software update. They buy a few hundred licences for a chatbot, give them to the team, and hope for the best.
But at best, that just creates operational debt. If you provide a generic tool to someone, they might get a bit faster, but that "speed" is trapped in their head. If they leave next month, that efficiency goes with them. The business hasn't gained anything permanent.
Instead of the old Digital Transformation model (requirements gathering, building, and testing to "prove" trust), we use the Onboard, Collaborate, Trust, Scale framework.
The adoption breakthrough happens when you move the AI out of a private browser tab and into your collaboration environment.
When an AI lives in the space where your team is already talking, it stops being a "tool" and starts being a participant. It enables a 1-to-many interaction rather than a lonely 1-to-1 chat.
When a team "coaches" the AI in a shared environment, that entire team sees the improvement. You aren't just training a piece of software; you are scaling your best people’s expertise across the whole department.
The most exciting part of this shift is that it empowers the people who actually know the work. We’ve moved the barrier to entry.
You don’t need your SMEs to become "Prompt Engineers." You don’t need them to learn to code or spend weeks in technical workshops. Because the AI is onboarded like a human, they can interact with it like a human. They mentor it, correct it, and guide it using the expertise they already have.
This isn't a traditional, heavy-lift IT project that takes 12 months to see value. We build at pace because we start with Collaboration. We get the teammate "in the room" quickly, allowing it to earn trust through performance rather than through endless theoretical testing.
In regulated industries like lending or banking, you can’t just "move fast and break things." You need the adult in the room. Not just to ensure safety but to give our best people the ability and safety to innovate.
That’s why we built with controls at the heart of our product. Real-time guardrails that check every single interaction against your specific policies keeping the business safe and providing an audit trail of every decision delivering the safety that Compliance demands, but at the speed that Operations needs.
It took Mark, Michael and I a while but now it's clear.
We aren't just trying to "optimise" a journey or shave a few seconds off a task. We are building the capacity for your best people to actually do the work you hired them for.
Teams are stretched. People are incredibly busy. Asking them to learn something new is more work.
Ultimately, this is about freeing your team to do the work they were actually hired for.
We're providing a framework people understand to safely innovate whilst making it easy for your best operators. We're helping people to perform and build the businesses of the future.
March 5, 2026
Greg Coleshill, Chief Commercial and Operating Officer
March 5, 2026
Not sure I should be saying this, but I'm going to anyway.
For a pretty long time, I had absolutely no idea what my co-founders were actually talking about.
I’d sit in meetings with Mark and Michael and they’d be deep in the weeds. They’d be talking about "MCP Servers," "retrieval context," and "non-deterministic outputs." I’d nod along. I’d get bits of it. But internally I was thinking, "This sounds great, but how on earth do I explain this to a COO who just needs their team to move faster?"
It took me a while to really work it out.
In my defence, it wasn't just me. It was all of us. Translating the technical into clearly understandable concepts is always the hard part. But taking a technical solution and articulating the real-world benefit is vital.
Because we aren't building software.
We’re building a way to hire intelligence.
Customers aren't working with a "program."
They are working with intelligence.
I've posted before about the role of words in shaping the future of AI. My focus then was encouraging people to stop thinking about "agents" and start thinking about "teammates."
The mental model matters because it builds trust it enables and powers adoption. But it goes beyond that. The framework we provide doesn't just govern how people perceive a solution; it provides a control framework they already understand. It’s the beginning of how people move from "personal productivity" to building organisational assets that actually help the business.
A customer recently said something that stuck with me: "We’ve managed non-deterministic intelligence for hundreds of years. It’s called people."
Operations leaders already know how to manage and ensure controls around people. We should use that same muscle for AI.
The big mistake I see everywhere right now is firms treating AI like a software update. They buy a few hundred licences for a chatbot, give them to the team, and hope for the best.
But at best, that just creates operational debt. If you provide a generic tool to someone, they might get a bit faster, but that "speed" is trapped in their head. If they leave next month, that efficiency goes with them. The business hasn't gained anything permanent.
Instead of the old Digital Transformation model (requirements gathering, building, and testing to "prove" trust), we use the Onboard, Collaborate, Trust, Scale framework.
The adoption breakthrough happens when you move the AI out of a private browser tab and into your collaboration environment.
When an AI lives in the space where your team is already talking, it stops being a "tool" and starts being a participant. It enables a 1-to-many interaction rather than a lonely 1-to-1 chat.
When a team "coaches" the AI in a shared environment, that entire team sees the improvement. You aren't just training a piece of software; you are scaling your best people’s expertise across the whole department.
The most exciting part of this shift is that it empowers the people who actually know the work. We’ve moved the barrier to entry.
You don’t need your SMEs to become "Prompt Engineers." You don’t need them to learn to code or spend weeks in technical workshops. Because the AI is onboarded like a human, they can interact with it like a human. They mentor it, correct it, and guide it using the expertise they already have.
This isn't a traditional, heavy-lift IT project that takes 12 months to see value. We build at pace because we start with Collaboration. We get the teammate "in the room" quickly, allowing it to earn trust through performance rather than through endless theoretical testing.
In regulated industries like lending or banking, you can’t just "move fast and break things." You need the adult in the room. Not just to ensure safety but to give our best people the ability and safety to innovate.
That’s why we built with controls at the heart of our product. Real-time guardrails that check every single interaction against your specific policies keeping the business safe and providing an audit trail of every decision delivering the safety that Compliance demands, but at the speed that Operations needs.
It took Mark, Michael and I a while but now it's clear.
We aren't just trying to "optimise" a journey or shave a few seconds off a task. We are building the capacity for your best people to actually do the work you hired them for.
Teams are stretched. People are incredibly busy. Asking them to learn something new is more work.
Ultimately, this is about freeing your team to do the work they were actually hired for.
We're providing a framework people understand to safely innovate whilst making it easy for your best operators. We're helping people to perform and build the businesses of the future.
