Services
What you can buy.
Four pillars. Read the one that matches where you are, not all of them.
CONSStart here
Consultancy
Judgement about systems you already have, or ones you are about to commit to. Most engagements begin here and many end here.
- ERP fit assessment: whether Odoo is the right answer, what it will really cost, and what it will not do. We sell no licences, so the assessment is not a sales call.
- Software architecture: designing the system before it is built, and the boundaries that keep it maintainable afterwards.
- Architecture review and technical due diligence, including pre-investment review of someone else's codebase.
- AI readiness: what is worth automating in your business, what is not, and what it would take to find out cheaply.
- Digitalisation of compliance-heavy processes, where the constraint is regulation rather than technology.
- Interim technical leadership for teams carrying more system than they have seniority to steer.
| Evidence | What it demonstrates |
|---|---|
| Platform strategy | Building a traceability platform commodity-agnostic while serving one customer first, a decision that is cheap early and unaffordable later. |
| Architecture governance | A written, non-negotiable data-access boundary that held across multiple applications and contributors. |
| Compliance automation | Reading a regulated Belgian document process and deciding what could safely be machined, and what could not. |
| Operator's view | We run our own SaaS and our own ERP in production, so cost, upgrade and migration advice comes from operating the thing, not from a vendor deck. |
Typical engagement
Day rate, or a short scoped review producing a written recommendation you can act on without us.
R&DIf it needs proving
Applied research & proof-of-concept
For technology with no off-the-shelf answer, including the funding proposal that pays for it.
- Feasibility studies against a hard constraint: cost, latency, offline operation, regulation, or data you do not yet have.
- Evaluation harnesses: how you prove an AI system is good enough to deploy, rather than a demo that worked once.
- Grant and programme proposals: technical specification, workplan, budget.
- Expert review of other organisations' research proposals.
| Evidence | What it demonstrates |
|---|---|
| Agent evaluation | A domain copilot with an MCP server, voice interface and a dedicated evaluation suite. Agent quality treated as an engineering measurement, not a vibe. |
| Speech under a cost ceiling | A 660-hour Conformer-CTC speech model made to run CPU-only and deployed live, because GPU inference would have made the use case uneconomic. |
| Regulated payments PoC | Purpose-bound payment rails: money that cannot become cash, designed around the regulation rather than retrofitted to it. |
| Smart-contract escrow | Escrow-as-a-service on NEAR. Blockchain feasibility work with a clear-eyed verdict. |
| Funded programmes | Applications, functional specifications and funding-landscape maps for agri-tech and knowledge-management programmes. |
| Peer review | Expert reviewer for a regional research funding programme. |
Typical engagement
Fixed-scope study or PoC, 4–10 weeks. You get a working prototype and the written argument for or against building it.
ENGIf it needs building
Engineering
Production systems, built and deployed. Not prototypes handed over as slide decks.
- Odoo: custom module development, ERP implementation, hosting, integrations, version migration.
- Full-stack products: Next.js and FastAPI over Postgres, Stripe billing, containerised, deployed.
- AI systems: LLM agents, MCP servers, speech pipelines, and the evaluation that keeps them honest.
- Mobile apps: Flutter front ends over Odoo or your own API, shipped to both stores.
- AI assistants: domain copilots that answer from your own data, with evaluation behind them rather than a demo.
- Agentic software engineering: agents doing real work inside a codebase, and the guardrails that make that safe.
- Data platforms: traceability, monitoring, analytics.
| Evidence | What it demonstrates |
|---|---|
| Syntone | A complete subscription platform in production since April 2026: Next.js 15, FastAPI, Supabase, Stripe and Docker. Built, deployed and still operated by us. |
| Traceability suite | Six Odoo 18 modules covering producers, intake, processing batches with lot contribution, and quality-based bonus distribution. |
| MCP platform | An MCP server as the single data boundary, with LLM applications as pure clients. Architecture that survives new contributors. |
| Odoo theme & product work | Repositioning and theme development for another ICT consultancy's flagship mobile suite. |
| ERP operations | Odoo 19 on containerised infrastructure with the real traps solved: proxy mode, filestore persistence, and a guard that refuses to drop a company's books on deploy. |
Typical engagement
Fixed-price project or monthly retainer. You get a running deployment, the repository, and the runbook.
TRNThen you own it
Training
Your team becomes able to operate and extend what was built, or to adopt a technology we already run in production.
- Odoo for the people who use it: functional training per app, and developer training on writing custom modules.
- Agentic software engineering for existing teams: tooling, guardrails, and where it actually pays.
- LLM agent architecture: MCP, tool design, evaluation.
- Handover as a deliverable, attached to a build rather than sold separately.
Typical engagement
One to three day workshop, or a training block attached to a build.