Field notes on AI that ships.
Essays and working notes from building production AI: evals, retrieval, agents, security, and the unglamorous engineering in between.
AI rules for UK businesses in 2026: what applies when you build an AI system.
The UK has no AI Act, but it does have rules. Here’s what UK GDPR’s new automated-decision regime, the ICO, the FCA and the MHRA expect from an AI system, and what changes if you serve EU customers too.
Read the noteEvery note.
-
AI rules for UK businesses in 2026: what applies when you build an AI system
The UK has no AI Act, but it does have rules. Here’s what UK GDPR’s new automated-decision regime, the ICO, the FCA and the MHRA expect from an AI system, and what changes if you serve EU customers too.
-
Best AI development companies in the UK (2026)
A shortlist of firms that build AI systems for UK businesses, from boutique specialists to listed engineering groups, with what each is best at, the evidence for it, and how to choose between them.
-
Evals before features
Most AI teams build the feature, then try to work out whether it works. We build the measurement first. Here’s what that looks like in practice, and why it makes teams faster.
-
The EU AI Act for companies building AI systems: a 2026 guide
What applies today, what the Digital Omnibus delayed, and who counts as the provider when you commission a custom system. For companies in the EU, and anyone selling AI into it.
-
AI in insurance: claims, underwriting and broker work that AI handles well (2026)
Where AI is paying off for insurers, MGAs and brokers, from first notice of loss to underwriting submissions, and the rules that apply when AI touches pricing and claims decisions.
-
Inside a WhatsApp sales agent
Our AI sales agent handles 70% of a solar client’s monthly sales inside WhatsApp. Here’s how it’s put together, and where the human team still comes in.
-
AI in healthcare: what clinics and health companies can build safely in 2026
Where AI is earning its place in clinics, diagnostics and health-tech, the line between a helpful tool and a regulated medical device, and what HIPAA, UK and EU rules require of the system.
-
AI development in Dubai and the UAE: data law, Arabic and WhatsApp (2026)
What UAE companies need to know before building an AI system: how federal, DIFC and ADGM data rules apply, and what Arabic and WhatsApp change about the build.
-
Retrieval that knows when to say “I don’t know”
A RAG system that always answers is a liability. The useful ones retrieve carefully, cite what they used, and decline when the evidence isn’t there.
-
AI for sales teams: from first reply to closed deal (2026)
Where AI helps a sales team sell more, from answering every lead in minutes to keeping the CRM honest, drafting proposals and following up, and where a salesperson still makes the difference.
-
Prompt injection is an input-validation problem
You can’t prompt your way out of prompt injection. Treat every token from outside your system as untrusted input, and design so that a successful injection can’t do much.
-
AI in construction: tenders, documents and site reporting that run themselves (2026)
Where AI helps contractors, developers and consultants today, from reading tender packs and specifications to RFIs, site reports and handover files, and how to start without disrupting live projects.
-
Fallback chains: designing AI that fails gracefully
Models time out, providers go down and outputs fail validation. A production system needs an answer for each of those moments before they happen.
-
AI in banking and financial services: compliance, operations and service (2026)
Where AI is delivering for banks, lenders, fintechs and wealth firms, from KYC files and compliance monitoring to customer service, and the rules that apply when AI touches credit and customers.
-
The six months after the demo
A prototype that impresses in a meeting and a system that survives production are different projects. The gap between them is predictable, and so is the work that closes it.
-
AI development in Saudi Arabia: PDPL, data residency and Arabic models (2026)
What Saudi companies need to know before building an AI system: the Personal Data Protection Law, transfer and localisation rules, Arabic-first models, and what to ask a partner.
-
AI customer service: what to automate and what to keep human (2026)
How to build an AI support agent that resolves real questions instead of deflecting them, which conversations it should hand to people, and how to measure whether it's actually helping customers.
-
AI in retail and e-commerce: selling, serving and running the store (2026)
Where AI pays off for retailers and online stores, from answering and closing sales conversations to product content and returns, and what it takes to let an AI talk to your customers.
-
AI for law firms: research, drafting and due diligence you can rely on (2026)
What AI does well for law firms and legal teams, when a custom system beats an off-the-shelf legal AI product, and the professional rules on confidentiality, accuracy and telling clients.
-
AI in HR and recruitment: what's useful, and what the law now requires (2026)
Where AI helps HR and talent teams, from answering candidates and scheduling to employee helpdesks and onboarding, and the bias-audit, transparency and human-review rules that apply to hiring AI in the US, UK and EU.
-
AI in education: what schools, universities and training providers can build (2026)
Where AI helps students, teachers and administrators, from admissions questions to course support and admin, and the stricter rules that apply when AI touches assessment, admissions or children's data.
-
AI in travel and hospitality: guest messaging, bookings and operations (2026)
How hotels, travel companies and restaurants are using AI to answer guests in any language at any hour, handle bookings and changes, and take routine work off their teams, without losing the human touch.
-
AI for solar companies: sales, design and operations that actually work (2026)
Where AI pays off for solar installers, EPCs and operators, from answering leads in minutes to catching failing inverters before they cost a month of output, with results from systems in production.
-
AI for manufacturing companies: what a mid-size plant can actually use (2026)
The AI projects that pay off for small and mid-size manufacturers, from quoting and quality paperwork to maintenance, and how adoption compares across the US, UK and Europe.
-
AI for logistics companies: use cases and how to start (2026)
Where AI pays off in freight, 3PL and distribution today, where a language model is the wrong tool, and how to pick the first one.
-
AI agents for real estate: from first enquiry to booked viewing (2026)
What an AI agent can do for brokers, developers and property managers in the US, the UK and Dubai, what it should never do, and how to build one that buyers trust.
-
AI for accounting firms and finance teams: documents, close and client work (2026)
Where AI saves accountants and finance teams the most time, from collecting client documents and processing invoices to month-end, tax research and advisory drafts, and how to keep it accurate enough to sign off.
-
Best AI development companies in the US (2026)
A criteria-based ranking of AI development companies serving the US market in 2026, for buyers who need production LLM applications, RAG systems, and AI automation — not slide decks.
-
How to choose an AI development company in 2026: a buyer’s guide
A practical framework for US and international buyers evaluating AI development companies in 2026: the criteria that predict production success, the red flags that predict failure, and how to compare global firms.
-
WhatsApp AI sales agents: how they work and why they close more deals
What a WhatsApp AI sales agent actually does — capture, qualify, propose, negotiate, close — why it outconverts web forms, and what it takes to run one in production.
-
Best AI companies in Pakistan (2026)
A criteria-based ranking of AI companies in Pakistan for buyers who need production systems, LLM applications, RAG, automation, and enterprise AI delivery.
-
AI is becoming more action-oriented, not just more intelligent
The latest model releases show the market moving past prompt-and-response interaction toward systems that complete real work across applications and workflows.
-
Enterprise AI is scaling fast, and infrastructure is now a strategic priority
AI adoption is accelerating, and infrastructure decisions now shape deployment outcomes as much as model choice does.
-
Multimodal AI is becoming the new standard for digital products
The next generation of AI products is being built around multimodal experiences that combine reasoning, diverse inputs, and real workflow integration.
-
Responsible AI is becoming a product requirement, not just a policy topic
As AI systems grow more capable and more widely deployed, governance and safeguards are moving out of policy decks and into core product design.
-
AI partnerships show the market has moved from experimentation to real deployment
The latest strategic partnerships across the AI ecosystem show enterprise adoption shifting from pilot programs to long-term infrastructure and platform commitments.