Domain model
One definition per concept, mapped to every source system beneath it.
Advisory work is easy to claim and hard to show. Each of these started as a process that was being run badly on paper, spreadsheets or a WhatsApp group — and became a product. They are also how PolyMat tests its own method: model the domain first, automate against the model second.
Inside most organisations the same real-world thing has a different name in every system. A person is a customer in the core platform, a party in CRM, an applicant in onboarding and a counterparty in risk. Nothing reconciles them, so every report, integration and AI pilot starts by re-deriving what should have been settled once.
The ontology layer settles it once. It is an independent model of a business's own concepts and the relationships between them, sitting above the systems rather than inside any of them — so applications, reporting and AI all reason against a single agreed picture instead of inferring meaning from table names.
You meet people constantly: clients, prospects, old colleagues, someone useful at a conference. The follow-up that would have mattered is always the one nobody remembered to make. Contacts sit scattered across a phone, LinkedIn, an inbox and memory, and not one of those can tell you who you have not spoken to in eight months.
RMS is a private record of relationships rather than contact details. Who you know, how you met, what you last talked about, what matters to them, and when it is time to reach out again. It is built for one person keeping track of their own network, not a sales team logging activity into a pipeline.
Most businesses and households lose the same thing repeatedly: the piece of paper that proves something happened. A receipt for a warranty claim, an invoice for a reimbursement, a delivery note for a dispute.
ScanKeep turns the phone camera into a filing system. Capture a receipt or document, and on-device OCR reads it, pulls out the merchant, date and amount, and files it somewhere searchable — so retrieval takes a search box rather than a shoebox.
Most apartment societies are run on a register book, a spreadsheet and a WhatsApp group. Nobody can say with confidence who has paid, which complaints are open, or what was agreed three committee meetings ago.
This puts the whole operation in one place: units and owners, maintenance billing and collection, complaints with an audit trail, and approvals that leave a record. The committee gets a real ledger; residents get visibility instead of a group chat.
Every productivity app has the same failure mode: it asks you to open a second place and build a new habit. Most people do that for a fortnight and then stop.
This one has no app to install. It lives inside WhatsApp, which is already open all day, and you talk to it the way you would message a colleague — add something, ask what is outstanding, mark it done. The interface is one you already know.
That is usually where the quickest return is. Tell us what it looks like today and we will tell you whether it is worth automating.