Short, honest pieces on running businesses with AI employees, agents and automation: what we are building, what we have learned, and what we would tell a friend. Written by the people doing the work. Press releases and media coverage live here too, filed under Media as they happen.
Insights
Claude Duet, Meta Muse, OpenAI Dots and Grok Bot all launched between August and October 2026. We have run our own always-on AI operator since March. Here is what these agents do, what they still get wrong, and what it means for a small business.
Read moreInsights
A US executive order now tells federal agencies to call AI 'super intelligence'. It changes no UK law. Here is what the rebrand and accompanying safety accord actually signal, and what to check in your own business regardless.
Read moreInsights
UK businesses with 100+ staff are almost twice as likely to use AI as smaller firms, 67% versus 36%, and company size predicts adoption far more than where a business is based. Here is what actually closes the gap.
Read moreInsights
On 17 September 2026 the King brought Nvidia, Google DeepMind, OpenAI and Anthropic to Dumfries House to discuss AI principles. Nothing binding was agreed, but trust and governance are becoming part of how customers judge small business AI use.
Read moreInsights
On 10 September 2026 we ran an AI visibility report on our own will-writing business: 20 prompts, four assistants, 80 measured cells. The domain was the most-cited domain in the sample and the brand was named far less often.
Read moreInsights
Anthropic released Claude Fable 5.1 on 1 September 2026. We switched our own AI operator onto it the next day and tracked seven days of usage: 160.8M tokens a day against an 89.3M baseline, and nothing throttled all week, on a counter we have never yet seen fire.
Read moreInsights
On 9 September 2026 we ran NVIDIA SkillSpector over the 58 skill directories in the two skill folders of our operator as they stood that day. Four scored CRITICAL. Nothing we read was malicious. Here is the routine, and the day we broke it.
Read moreInsights
A plain-English map of the Scottish government support that can pay toward AI adoption work, what each route actually covers, and where to check current status before you apply.
Read more
Insights
Deploying Claude Code across a company is eight decisions, not an install. Here is the deployment decision document we use, the security questions a client will ask you, and what Anthropic assesses before it issues its Claude Code partner badge.
Read more
Insights
The Claude Certified Associate exam is proctored, closed book and scored out of 1000. The surprise is what it weights most heavily: not writing prompts, but knowing when the answer is wrong and when a human has to check it.
Read moreInsights
Claude Enterprise buys single sign-on, provisioning, compliance tooling and admin control. For most businesses under fifty people the Team tier is the right floor and Enterprise is money spent early. Here is where the line actually sits.
Read moreInsights
Custom software used to fail the small business test on time and cost, so owners bought the off-the-peg tool and changed how they worked to suit it. That trade-off has stopped being necessary.
Read moreInsights
What building compliance-grade systems for an oil and gas well intervention firm taught us about AI automation for every other business: design the system so the record produces itself.
Read moreInsights
We handed Anthropic Fable 5 and OpenAI GPT-5.6 the same brief and design system and asked each to build a complete website in one shot. Both produced a full 14-page site with zero build errors. Here is how they differed, and why the one-shot production website is already here.
Read moreInsights
Most businesses have not adopted AI. Their staff have, on personal accounts the business cannot see. What that actually exposes, what it does not, and the one-week fix.
Read moreMore to come
We publish when we have something worth your time, not on a content calendar. Every post here is real work, really shipped, or a straight answer to a question clients actually ask.
Not a discovery call. Not a pitch with a calendar link. Five questions about how your business actually runs, and it costs nothing. Ross reads every enquiry himself and replies within one working day with a straight first answer: what looks worth automating, and what doesn’t.
If it looks like we can genuinely help, the next step is the AI Operations Day: one working day inside the business, followed by a written Opportunity Map. It shows what is hurting, what should stay human, and the best one or two jobs to prove first. That part comes later, and only if it makes sense for you.
Free to ask. No obligation. Ross replies personally within one working day.