Capability map
Everything we do, and the sectors we do it in
Four services and six sectors. The service pages describe how the work runs — what we build, what you keep, when we stop. The sector pages describe the constraints: the regulation, the data and the KPIs that industry actually manages to.
Not sure which page applies? AI strategy and innovation is the usual starting point.
Our services
Each page carries its own metrics, method, systems list and the cases where we are the wrong choice. Read the one that matches the work you have in mind.
AI strategy that ends in a shipped use case, not a deck
We build the three things that decide whether AI spend returns anything: a measured baseline, a use-case portfolio with a written kill criterion per case, and an EU AI Act position that matches what you actually run. We will not open with an enterprise-wide data maturity programme — readiness is assessed only for the critical data elements your first three use cases consume, because the enterprise version takes a year and produces a radar chart.Read the service →Next-generation AICustom AI systems that hold up after the demo
We build retrieval, document extraction and agent systems on top of frontier model APIs — together with the eval harness, the guardrails and the tracing that decide whether any of it survives contact with your real documents. We ship an agent only where a workflow provably cannot do the job, because multi-agent orchestration burns roughly 15x the tokens of a chat call and a single agent about 4x — Anthropic’s measured figures. Against a retrieval workflow the gap is smaller, but it is real, and it comes with failure modes a workflow does not have.Read the service →Customer experience AIVoice and chat agents that resolve, escalate — or admit they don’t know
We build voice and messaging agents that live inside an 800ms voice-to-voice budget, answer only from passages retrieved out of your own knowledge base, and write the outcome back into Zendesk, Salesforce or whatever is your system of record. The constraint we hold to: nothing irreversible — a refund, a cancellation, an eligibility decision — happens without a human approving it, and the route to that human is designed before the agent is.Read the service →Operations & EfficiencyAutomation designed from your event log, not from a workshop diagram
We build back-office automation for finance, purchasing and order intake — document intake with confidence routing, validated writes through released ERP interfaces, and long-running cases held in a durable orchestrator instead of a cron job and a status column. The constraint we respect is that nothing gets built before the event log says what the process actually does: below roughly 80% of cases carrying complete timestamps and resolvable IDs, we fix the measurement before we touch the process. End to end that is 9 to 15 weeks from signature to a first process live in production, and 13 to 21 including handover — if someone quotes you six, ask what they are skipping.Read the service →Pick the service if you know what you want built. Pick the sector if you want to see whether we understand your constraints.
Sectors we work in
Sector pages exist because the hard part is rarely the model. It is the meter data, the release process, the safety case or the school year.
AI for keeping power reliable through the 15-minute market
Forecasting, grid analytics and automated operations for energy players.Read the sector page →Logistics and transportAI for freight that runs loaded and arrives when promised
Route planning, dispatch and tracking, automated where the data supports it.Read the sector page →Retail and e-commerceAI for keeping the promise your product page makes
Demand forecasting, order handling and customer flows, automated step by step.Read the sector page →Manufacturing and industrial operationsAI for plants that ship good parts, shift after shift
Predictive maintenance and OEE data capture on brownfield lines.Read the sector page →Media and publishingAI for journalism that still pays for itself
Retention and next-best-offer models for subscription publishers.Read the sector page →Education and EdTechAI that keeps children in school and teachers teaching
Learning powered by AI — personalized teaching, automated administration.Read the sector page →
Still not sure where your problem sits?
Describe the process in two paragraphs. We will point you at the page that fits, or tell you that none of them do.
We answer inquiries ourselves — there is no qualification funnel in front of us.