Insights
From the builders
Notes on AI agents, infrastructure and shipping real products — from Pharoah Technology Ltd.
-
We Built a 5-Minute Watchdog So We'd Never Hear About a Crash From a Customer First
Our own systems failed silently at 3am until we built a 5-minute watchdog to catch it — here's the three bugs we hit building it, and why every SME running unmonitored automation has the same blind spot.
Read more → -
The Silent Automation Failure Every SME Risks (And How We Built Ten AI Agents to Stop It)
Anonymous automation fails quietly — you usually find out from a client, not a dashboard. Here's how naming, monitoring, and holding our own AI agents accountable closed that gap, and what any SME can copy from it.
Read more → -
Your AI Bill Is Padded — One Client Cut Theirs by 62% Without Touching a Prompt
Most businesses run every AI task through the same expensive model, whether it needs it or not. One client's bill fell from £18,400 to £6,900 once theirs stopped.
Read more → -
GhostWire Is In Early Access — Here's the Vendor Checklist It Forced Us to Write
We put our own secure messaging tool into early access, which meant writing down exactly what we could and couldn't prove — including the six questions every SME should ask any 'secure comms' vendor before trusting them with client data.
Read more → -
We Scored 58/100 on Google's Speed Test. Here's the Cheap Fix That Got Us to 100.
Our product pages scored 58/100 on Google's speed test — the same test that quietly decides how many visitors stick around. Here's the low-cost, no-framework fix that took us to 100, and what your own score might be costing you.
Read more → -
The Silent 'Success' Bug That Hid Two Weeks of Missing Data
A function that returns true doesn't always mean it worked — ours said "success" for two straight weeks while our analytics silently failed. Here's the blind spot in dashboards and automated reports that every business owner should check for.
Read more → -
No Cookie Banner, No Tracker: How We Built Our Own Analytics in a Day
Most analytics tools mean a cookie banner and a data-sharing decision you probably haven't reviewed. Here's how we built a ~2KB first-party alternative instead — no tracker, no banner, done in under a day.
Read more → -
Why We Label Every Unfinished Product — And the One Question It Hands You For Free
Our product catalogue labels every build by real stage — live, early access, labs, by engagement. That same discipline hands you one question that exposes any vendor selling you a roadmap instead of a product.
Read more → -
The Forge Method: How Two People Ship a New AI Product Every Week
For a two-person studio, the most expensive bug isn't in the code — it's in the memory: you rebuild what you forgot you'd built. How a live reuse register, deliberately boring shared rails, and review budget spent by blast radius let a tiny team ship fast without shipping slop.
Read more → -
Safe by Design: What Building AI for Children Actually Requires
Kid-safe AI isn't an adult chatbot with a swear filter and a mascot. It's age assurance that persists across a session, privacy defaults set before signup, and an escalation path that knows a disclosure of harm at home must not be tipped off to the home.
Read more → -
The Compliance Layer Most UK AI Products Quietly Skip
Scattered API keys, no kill-switch, no audit trail. A governance gateway is the boring fix enterprise buyers and regulators actually expect.
Read more → -
I Built Pharaoh Vision Because I Was Tired of Guessing Which Light Was Red
Our co-founder is colour-blind. Here's why he built real-time colour correction instead of waiting for design teams to fix red/green dashboards.
Read more → -
Why We Build a UK AI Company from the Costa del Sol
The myth says serious software needs London or SF. Here's what actually happens when you build UK-facing AI from Malaga instead — the real advantages and the...
Read more →