Furze Newsletter: Feb 25

At this time of year, most people are recovering from the bleakness of January. It’s rained for what feels like an eternity and all the good Saturday night TV has vanished. But at Firehills, it’s been bums on seats, heads down, and full steam ahead on our first workstreams from the tech road map. It’s gone by in a blur and we’re making great progress. Here are some of our January and early February highlights.

Doubling Down on Core Functionality

Good advice only works if you listen to it and act on it. One of our network supporters, who ran a prop tech business for over 20 years, recently shared some hard-won lessons that have really stuck with us.

He explained that when his business started winning clients, the technology gradually became heavily over-customised. Several clients ended up with bespoke functionality built just for them. Maintaining all of it was heavy, costly, and in hindsight, it’s unnecessary.

From my own experience in professional services technology, this is something that’s avoided at all costs. Building technology that only one customer can use is neither scalable nor resalable.

So we’re making a conscious set of choices to bring in as much customer voice as possible around our core product. To give you a flavour of where we’re heading: Furze is being built around four connected capabilities. It starts with organisational analysis, using the Firehills Canvas model to build AI-generated insight from public and private data. From there, we layer on market mapping to structure and process complex competitive landscapes. Then we treat those markets as systems, understanding how your organisation’s value connects to others to surface growth opportunities. Ultimately, all of this feeds into generating strategic growth scenarios — complete with deployment planning and outcome modelling — so it’s not just about understanding where you are, but where you could go next.

Right now, we’re focused on getting the first two right before layering on the rest. Nail the foundations, then build.

If you’re interested in becoming a potential test case and would like to explore what this looks like, we’d love to hear from you. Use the contact form on our website and we’ll be in touch.

Mapping the World’s Organisations Is Harder Than We Thought — But Not Impossible

One of our critical workstreams for building out market maps is finding every organisation that exists within a given market space. We’ve created early prototypes for this before, but now we’re building a battle-hardened approach that can be repeated and scaled.

This is a multi-layered process. We use a number of open-source tools to continually refine and qualify the organisations we find. For example, if we discover both “Acme Inc.” and “Acme Incorporated,” how do we resolve those into a single entity? Our method folders in multiple technology steps which behind the scenes work their magic.

Many CRM providers have solved for this problem and for anyone in a corporate organisation will know how painful this job is. Especially merging data together, often it just ends up in the too hard box and never gets done. It’s right for us to invest now and never feel this pain later.

As with everything, technology can’t fully solve this on its own. We’ll be reviewing this all piece by piece of have a high quality tech process we can rely on for years to come.

One for the techies: We’re using Agentic Coding to Accelerate Development

We’ve been using AI-assisted coding since our inception, but there are different levels of productivity to be gained from it. You’ve probably seen people on LinkedIn showcasing how non-technical founders have built an app in four hours using tools like Codex or Claude Code, which let you write code in natural language.

What we’re doing is similar in principle, but we’re pushing it further. We’re exploiting additional productivity gains through AI agents, reusable skills, and more recently — GitHub worktrees, which let us distribute several code builds across separate agents working in parallel. The result is genuine scale: we’re seeing 3–5x productivity multiples, and that number keeps climbing as we refine our approach.

So, When Will Furze Be Ready?

Our aim is to have the core functionality of Furze ready as an MVP by spring 2026. As of today, we’re on track to meet that milestone. As we get closer, we’ll be planning a more formal launch campaign plus, we’re currently setting up our first in-person event.

Watch this space — and if you want to stay in the loop, sign up for email updates at firehills.io.

By Rob Read: 20/02/2026