We work with operators whose operation runs ahead of its information, from low-tech industrial plants to high-end hotels. We capture the work as it happens, show it live to management, and rebuild the decisions that matter with AI doing the heavy lifting and your people in charge.
Most software makes you fit your problem to the product. We do the reverse. We arrive with questions about how the work is done, where it slows, and what the people doing it would change if they could, and we build to that.
We do not bill for an idea and leave you to work out what to do with it. We learn how you work, see the workflows, find the gaps, then design and build the system that closes them.
From a factory floor in Côte d'Ivoire to a hotel in Lisbon, we are there with a solution: on site, until it runs, and until the team trusts it.
AI changes what a good process looks like. Where a process can be rebuilt, we design it from first principles, as it would be designed with AI available from the start. Where it cannot, because of equipment, regulation, habit or people, we start from the workflow you have and enhance it without breaking what already works. Either way we start small: one process, proven in weeks, then grown with you.
We walk the operation with the people doing it: the workflows, the handoffs, where it slows, and what an ideal day would look like from where they stand.
Where things are, what is stuck, what the numbers are, and in what form and rhythm managers want that visibility.
Rebuilt or enhanced, with ownership, controls and review points fixed before any technology is chosen. Then one small system, built in your team's language and put to work in weeks, not months.
Measured against the baseline, with the team using it every day. When it works, we expand it one process at a time, and we stay with you through each step.
A mill's daily production report. Today a clerk collects the station sheets and re-types them into a spreadsheet, and management sees the figures for a lot days or weeks after it was processed, once the sheets have been found, typed and reconciled. Designed with AI from the start, each reading is captured once, where it is made, and checked as it lands. The figures are live all day, the report compiles itself, and a manager who wants yesterday's extraction rate by station simply asks for it and gets the answer with its source. The report stops being a job.
A hotel's morning rate check. The revenue manager opens the same rate sheet at eight, as before. What has changed is what is already on it: overnight competitor moves by channel, pickup against the same date last year, the concert nobody had priced, and a suggested rate with its reasoning and its cost in occupancy. The habit stays. The decision stays with the revenue manager.
Work captured once, where it happens. No re-typing, no reconstruction at the end of the shift or the month.
Management sees the operation as it runs, in the form it asked for, on the same figures as the floor.
Decisions taken on the full set of inputs, with the reasoning kept, and your people making the call.
The vocabulary changes from a mill to a hotel. The method does not.
Plain-language questions answered with sources; daily and weekly reports; control checks as leads.
Read Industry · 02Lot-level mass balance, live yield and grade mix, piece-rate payroll from the same record.
Read Hospitality · 03One platform for leads, offers, contracts and check-in, fed directly by the website.
Read Hospitality · 04The morning rate check and the weekly meeting, kept, with the market assembled in advance.
ReadThe people you deal with design and build the solution, and remain accountable for it in operation.
Software engineer and founder of Aperture Management, a London consultancy in software, product management and data analysis. Background in finance and data technology, including data engineering on financial-crime datasets. Responsible for capture, data, live reporting and the AI layer. Dean to verify
Twenty years as a C-level executive and in business development, much of it in agro-industry and processing. Brings the operator's perspective: how a mill, a plant and a board work in practice, what the figures mean, and which decisions move them.
Describe the part of the operation you cannot see clearly, or the decision you want better inputs for. We come on site, map the work, and return with a scoped design for the first system.
Lisbon · Abidjan. English now, French edition in preparation.