Enterprise workflows.

The workflows that cost an organisation the most are rarely the ones a product can be bought for. They span several systems, depend on knowledge that was never written down, and need a person's judgement at specific points — which is exactly why generic tools bounce off them. We build the system that operates one of them.

The problem

A workflow like this is usually held together by people: someone knows which exception is normal, someone else knows where the real record lives, and the software involved was never designed to talk to itself. Automating the easy half tends to move the work rather than remove it.

What we build

Document intake and extraction, retrieval over institutional knowledge, and orchestration across the software already in place — wrapped in the governance layer that decides what the system may do on its own: permissions, approval gates, audit logs, and continuous evaluation against real examples.

Who it's for

Enterprises and institutions with a workflow valuable enough, and complex enough, to justify building rather than buying.

Where we have built it so far.

The same practice each time: understand how the work actually runs, build the system that operates it, and keep a person on the decisions that carry consequences. What changes is the domain.

Clinical trials

Runs the recruitment pipeline for a trial, so enrolment keeps moving whether or not anyone at the site has time to chase it this week.

The problem

Clinical trials are frequently delayed or abandoned because sites cannot enrol participants quickly enough. Recruitment work is high-volume and repetitive — screening enquiries, checking eligibility criteria, chasing responses — and it competes with clinical duties for the same people's time.

What it does

Engages inbound enquiries, screens participants against protocol eligibility criteria, manages communication across channels, qualifies candidates for site review, and keeps follow-up moving so interested participants are not lost between steps.

Who it's for

Trial sponsors, contract research organisations, and research sites running recruitment at volume.

Industrial manufacturing

Holds a plant's production, quality and maintenance picture in one place, and turns each shift's decisions into a record at the moment they are made rather than at the end of the week.

The problem

A plant's operating knowledge is spread across a scheduling system, a quality binder, a maintenance log, and the experience of whoever is on shift. When a line goes down or an order changes, the reconciliation happens in people's heads and over the radio — and the record of what was decided gets written up afterwards, if at all.

What it does

Draws production, quality and maintenance records into a single operating picture, drafts the schedule changes and documentation a shift produces, and routes anything consequential — a deviation, a hold, a changeover — to the person accountable for approving it.

Who it's for

Manufacturers running multi-shift production where scheduling, quality and maintenance decisions have to stay in step.

Childcare

Carries the administrative weight of a nursery — enrolment, ratios, records, families — so it stops being carried by the people who are supposed to be in the rooms.

The problem

Childcare runs on records that have to be right — who is enrolled, which staff are qualified, what ratio a room is at, what happened to a child today — kept by people whose actual job is looking after children. The administration ends up done twice: once in the moment, and again to satisfy the regulator.

What it does

Handles enrolment and waitlists, tracks staffing against required ratios as the day changes, captures daily records and incidents at the point they happen, and keeps families informed — so the documentation a regulator asks for is the same record the day already produced.

Who it's for

Nurseries, daycare centres and multi-site childcare operators accountable for ratios, staff qualifications and daily records.

Commercial real estate

Answers the research question behind a commercial building once, and keeps the answer current as ownership, tenancy and timing move underneath it.

The problem

Selling into commercial buildings means answering the same questions for every property: who owns it, who actually decides, what is already installed, and when the timing is right. The research is slow, the data is scattered across sources, and it goes stale quickly.

What it does

Assembles property and ownership data, identifies the decision-makers behind an asset, and runs the pipeline around them — with agents doing the research and drafting outreach that a person reviews before it is sent.

Who it's for

Contractors, service providers and vendors whose customers are commercial buildings and the people who own or manage them.

Building operations

Keeps a building's requests, work orders, vendors and inspection schedules moving as one operation, rather than as whatever the manager on duty happens to be holding.

The problem

A building generates obligations from every direction at once — a tenant reports a fault, a unit starts failing, an inspection falls due, a vendor contract comes up for renewal. They arrive by email, phone and spreadsheet, they are tracked in someone's head, and the first reliable sign that one was missed is the consequence of missing it.

What it does

Takes in requests from tenants and staff, opens work orders and routes them to the right vendor, tracks preventive maintenance and inspection schedules against each asset, chases what has gone overdue, and keeps the record an owner, an insurer or an inspector would ask to see.

Who it's for

Property managers, facilities teams and owner-operators running the day-to-day of a building or a portfolio of them.

Hiring and talent evaluation

Makes interviews comparable, so the decision at the end rests on what candidates actually demonstrated rather than on who remembers which conversation.

The problem

Hiring decisions are often made from inconsistent evidence. Interviews vary by interviewer, notes are thin, and the comparison between candidates ends up resting on recall and impression rather than on a common standard.

What it does

Conducts structured interviews, records what a candidate actually said and demonstrated, and evaluates it against a defined rubric — so hiring teams compare evidence rather than impressions.

Who it's for

Talent teams and hiring managers running structured, high-volume, or distributed hiring processes.

Small and midsized business

Gives a company with no AI team the thing an AI team would build first: a system that knows how the business actually runs, and can act on that.

The problem

Most small and midsized businesses cannot afford an internal AI team, and generic assistants know nothing about how a specific company actually runs. The knowledge that matters — who approves what, which exceptions are normal, where the real records live — is undocumented and lives in people's heads.

What it does

Learns a company's operating model from its existing tools and its people, then runs agents against that context: handling intake, moving work between systems, preparing documents, and escalating decisions to a human when they matter.

Who it's for

Small and midsized businesses with real operational complexity and no dedicated AI engineering team.

Retail and bespoke commerce

Closes the gap between a bespoke order and the supplier purchase it implies, without either end of it being retyped by hand.

The problem

Buying something genuinely bespoke online usually means a form that cannot express what you want, or a phone call the business does not have time to take. Behind the counter, the same order then has to be translated into supplier purchases by hand.

What it does

Turns the storefront into a conversation, so a customer can describe and refine what they want, then carries that order through into operations — drafting the supplier purchases it implies and holding them for a person to approve before anything is committed.

Who it's for

Retailers whose product is made or assembled to order, where a sale starts as a conversation and ends as a supplier purchase.

Is one of these close to your problem?

If a workflow in your organisation looks like something here — or looks like nothing here — we would like to hear about it.

Start a conversation See how we work