Industries
The same engineering, organised by who is buying it. Fourteen sectors and 151 verticals, each page written for the person who has to answer for the outcome in that industry rather than for a general reader.
Two ways into the same work.
The services tree is organised around what gets built — models, agents, pipelines, governance. That is the axis a technical buyer navigates. It is the wrong axis for a hospital operations director or a logistics COO, who does not start from "we need a computer vision engagement" and never will.
So these pages start from the sector's own problems and route into the engineering. They are not restatements of the service pages with an industry noun substituted throughout, and where a sector genuinely needs nothing different from the general case, the page says so and links onward rather than inventing a distinction.
What transfers between industries, and what does not
Most of what we do is identical in a bank and in a bakery. A minority of it is completely different, and being clear about which is which is the difference between useful sector work and a consultancy pretending to know your business better than you do.
The engineering transfers
The engineering. Evaluation design, baselines, leakage detection, data quality, deployment, monitoring, drift, human review design, cost per request. A churn model in telecoms and a churn model in insurance are the same problem wearing different vocabulary, and the failure modes are identical.
Process patterns transfer partly
Process patterns. Document-heavy intake, dispatch and scheduling, demand forecasting, inspection, claims-like workflows. The shape recurs across sectors; the tolerances, exception rates and what counts as an acceptable error do not.
Domain knowledge does not transfer
What the data means, what a wrong answer costs, which regulator cares, what actually happens on the floor at 3am, and which of your own numbers people already distrust. That is your expertise. We ask for it early, in writing, and we do not pretend to supply it.
How we staff sector work
Which is why every sector engagement is staffed with your domain experts inside it rather than alongside it. Where we have not worked in a vertical before, we say so — that is a reason to involve your people harder, not a reason to decline the work or to imply experience we do not have.
Regulated sectors
Healthcare, financial services, retail, manufacturing and technology. Three of the five carry duties under the EU AI Act, so on those the governance work is part of the build rather than a review at the end. The engineering is ordinary; the evidence requirements are not.
Healthcare & Life Sciences
Hospitals, pharma, medical devices, payers and the diagnostics chain — where an error has a clinical consequence and a regulator.
Explore the sector →Financial Services & Insurance
Banking, insurance, capital markets, fintech and lending — where models decide who gets access to money.
Explore the sector →Retail & E-commerce
E-commerce, physical retail, CPG, fashion and grocery — high volume, thin margins, and demand that moves weekly.
Explore the sector →Manufacturing & Industrial
Discrete and process manufacturing, automotive, aerospace, semiconductors and chemicals — where AI meets machinery that has mass.
Explore the sector →Technology & Telecom
SaaS, startups, telecoms, cybersecurity and data centres — buyers who will read your architecture before they read your case study.
Explore the sector →Physical and operational
Energy, transport, property, professional firms and the public sector. Assets with long lives, processes with physical consequences, and buyers who need a system to keep working on a night shift rather than to impress in a demonstration.
Energy & Utilities
Oil and gas, renewables, power generation, grid, water and EV infrastructure — long asset lives and unforgiving physics.
Explore the sector →Transport & Logistics
3PL, freight, maritime, air cargo, rail, warehousing and supply chain planning — where a forecast error becomes a truck in the wrong place.
Explore the sector →Real Estate & Construction
Residential and commercial property, PropTech, construction, infrastructure and facilities — long cycles, fragmented data.
Explore the sector →Professional Services
Legal, accounting, consulting, recruiting, staffing, agencies and BPO — firms whose product is expert judgement.
Explore the sector →Public Sector & Education
Government at every level, defence, justice, policing, schools, universities and EdTech — public accountability as a design constraint.
Explore the sector →Commercial and public
Media, food and agriculture, travel and consumer services. Rights, provenance, perishable inventory and fragmented operators — commercial pressure rather than regulatory pressure, and different failure modes.
Media, Entertainment & Sports
Publishing, broadcast, film, music, streaming, sport and the creator economy — rights and provenance are the hard part.
Explore the sector →Agriculture, Food & Environment
AgriTech, crop and livestock, food processing, restaurants, climate and ESG — biology, weather and traceability.
Explore the sector →Travel & Hospitality
Airlines, hotels, OTAs, cruise, car rental and events — perishable inventory and demand that swings on a headline.
Explore the sector →Consumer & Lifestyle Services
Fitness, salons, home services, dealerships, childcare and franchises — many small operators, thin IT, real volume.
Explore the sector →Regulation is what actually differs by sector
The EU AI Act classifies systems by what they do and to whom, and several of its high-risk categories map directly onto sectors rather than onto technologies. Seven of the fourteen clusters carry meaningful exposure, and on those the documentation, logging and oversight requirements shape the build from the first week.
| Sector | Indicative EU AI Act exposure |
|---|---|
| Healthcare & Life Sciences | Annex I medical devices; Annex III health insurance pricing |
| Financial Services & Insurance | Annex III creditworthiness and life/health insurance pricing |
| Manufacturing & Industrial | Annex I safety components in regulated products |
| Energy & Utilities | Annex III critical infrastructure |
| Transport & Logistics | Annex III critical infrastructure |
| Professional Services | Annex III employment and recruitment |
| Public Sector & Education | Annex III education, law enforcement, migration, justice, essential services |
On the current timetable
High-risk obligations for stand-alone Annex III systems now apply from 2 December 2027, and Annex I systems embedded in regulated products from 2 August 2028, following the Digital Omnibus package. The transparency duties already apply. This is our reading as at September 2026 and we work alongside your counsel rather than in place of them — see EU AI Act compliance readiness.
Before you pick a sector
Do you actually have experience in our industry?
In some sectors deeply, in others not yet, and we will tell you which before you engage rather than after. What we bring in every case is the engineering — evaluation, data quality, deployment, monitoring — which is genuinely sector-independent. What we do not bring is your domain knowledge, and any supplier claiming deep expertise in all fourteen of these sectors is describing a sales position rather than a team.
How is an industry page different from a service page?
A service page describes an engagement: what gets built, how long it takes, what you receive. An industry page starts from the sector's problems and routes you to the right engagements, with the constraints that sector imposes — regulatory exposure, data availability, what an error costs — stated up front. If you already know you need a forecasting model, start with services.
Which sectors carry EU AI Act obligations?
Seven of the fourteen have meaningful exposure: healthcare, financial services, manufacturing, energy, transport and logistics, professional services through employment and recruitment, and the public sector. The categories are defined by use rather than by industry, so the mapping is indicative — your actual classification is settled per system with your counsel. See EU AI Act compliance readiness.
We are a small operator in a fragmented sector. Is this relevant?
Sometimes, and the honest test is volume. AI work has a fixed engineering cost that does not shrink with your size, so the question is whether the process you would automate happens often enough to repay it. Where it does not, we will say so — and several sectors here are full of operators for whom a well-chosen off-the-shelf product is the better answer than anything we would build.
Can you work across several of our divisions at once?
Yes, and it is often where the value is, because the same underlying capability serves more than one division and the data work is shared. What changes per division is the domain input and the acceptable error rate, not the platform. Start with AI strategy and roadmap where the scope spans a group.
Not sure your sector is the right way in?
If you already know what needs building, the services tree is the faster route. If you know the problem but not the technology, start here or just describe it to us.