Our Work
This is what the next generation of software companies looks like.
Small teams. Precise specifications. Production platforms shipped in weeks. These are not demos or prototypes — they are the evidence that spec-driven, AI-augmented engineering works at a scale that used to require organisations ten times our size.
Every product below was built using our full lifecycle: human architecture, AI-augmented implementation, human verification, deployed in production.
Data Act Compliance · EU
Dativo
The Problem
The EU Data Act (Regulation 2023/2854) is in force since September 2025. Manufacturers of connected products must share product-generated data with customers and authorized third parties — or face penalties of up to 4% of global annual turnover. Most manufacturers have equipment data scattered across silos with conflicting identifiers, no customer-facing access layer, and no audit trail.
What We Built
Dativo is an operational compliance platform for connected product manufacturers navigating the EU Data Act. Equipment registry, smart data matching, customer self-service portal, fulfillment engine with SLA tracking, trade secret classification, compliance audit trail. Open-core — Community Edition on GitHub, Commercial Edition for enterprise.
How We Built It
Built by a two-person team in weeks. This is what becomes possible when AI handles implementation volume and senior engineers govern every architecture decision. The economics of building a platform like Dativo changed dramatically. The bar for what it needs to do didn't.
Agricultural Advisory · Tunisia
Filahti
The Problem
Tunisia has 516,000 farms, 87% under 20 hectares. Public agricultural extension services are functionally dead — offices can't cover fuel for field visits. No private advisory market exists. Pesticide dealers fill the vacuum, diagnosing and selling with a conflict of interest and credit terms reaching 50% interest. Climate conditions are changing fast, with prolonged drought since 2017, and inherited farming knowledge is no longer enough.
What We Built
Filahti connects Tunisian farmers with vetted agricultural experts, government extension agents, and certified input suppliers. Hybrid WhatsApp bot and companion mobile app. Credit-based consultations, AI-assisted diagnostics, digital prescriptions, on-site visit booking — in Tunisian Arabic, Modern Standard Arabic, and French.
How We Built It
A product that would have taken a funded team of eight a year to build. Specified and shipped by two engineers. The Jevons Paradox applies here too: AI making agricultural software viable for niche markets means the farmers who needed expertise the most can now access it. Launching 2026.
What the Specification Looks Like
The complete Filahti specification — system architecture, functional requirements, implementation plan, test results, and handoff documentation — is available as a live walkthrough. You can read the same documents our engineers authored, see the comments from stakeholders, and explore every phase of the delivery lifecycle. This is what every Unio Lab engagement produces.
What's next
We're always scoping the next product. If you have a problem that needs a platform — not just a feature — we should talk.
Have a system that needs engineering?
Describe the problem. We'll assess whether it's a good fit for spec-driven delivery and scope it within a week.