Juniors are too expensive — or you can’t find them.
BTW work has become commodity. Juniors are scarce. Senior partners do work far below their billable rate. And every quarter it starts over.
Your clients upload receipts and bank statements. Utisha scans, categorises, and reconciles them. Your accountant reviews the result in minutes — then one click books it in Exact Online.
~1 hour saved per client, per quarter, on BTW work.
Design partners wanted for Q2 2026.
Built by Dennis Dam, freelance senior software engineer at Rabobank, ING, KPN and KLM.
// SHIPPED · DUTCH BTW · EXACT ONLINE · EU-SOVEREIGN
BTW work has become commodity. Juniors are scarce. Senior partners do work far below their billable rate. And every quarter it starts over.
Receipts by WhatsApp, email and post. Bank statements in PDF. You wade through everything for three months — for one filing.
Every booking documented. Every correction logged. Every file person-independent. Hours per client, per quarter — invisible to the client.
Working on AI compliance in a regulated organisation? See our compliance approach →
A complete quarterly BTW workflow for Dutch accounting firms. Clients upload documents through their own portal. Utisha handles scanning, categorising, and reconciliation. Accountants review, correct, and push the bookings to Exact Online. Live with design partners.
We are co-designing the next vertical with audit firms: working-paper preparation, sample selection, evidence cross-referencing, and exception triage — reusing the same primitives that power BTW. Talk to us if you want to shape it with us.
Document scanning, bank-statement parsing, reconciliation, customer portals, accountant review, and direct push to Exact Online and other APIs. The building blocks behind BTW are ready to assemble into the next vertical.
Every organization works differently. Utisha lets you define multi-step agent workflows — with human oversight built into every critical gate.
AI agents that draft risk assessments, analyse contracts, and generate regulatory reports — with mandatory human approval gates for all external-facing outputs. DORA and MiFID II ready.
Learn more →AI agents that process clinical documentation, generate summaries, and support research — with automatic patient data protection and full MDR compliance.
Learn more →AI agents that manage requirements traceability, generate test documentation, and support safety analysis — fully auditable across ASPICE and ISO 26262.
Learn more →AI agents that process freedom of information requests, assess grant applications, and draft regulatory impact assessments — on your infrastructure with air-gapped deployment. No data leaves.
Learn more →AI agents that review contracts, extract clause risks, and track regulatory changes — with mandatory human approval for all legal advice. Attorney-client privilege preserved.
Learn more →AI agents that assemble tender submissions, automate health & safety documentation, and streamline subcontractor prequalification — with full audit trail for every decision.
Learn more →AI agents that compile audit evidence for SOC2 and ISO 27001, generate incident post-mortems, and triage security vulnerabilities — fully integrated with your DevOps and security toolchain.
Learn more →// Self-hosted anywhere. Docker Compose, Kubernetes, or managed EU cloud. Every model swappable. No vendor lock-in.
We’re partnering with regulated organizations to deploy Utisha before general availability. You get early access to the platform and a direct line to the team. We handle the integration.
Or reach out directly: info@utisha.com
DORA gave banks a framework for digital resilience. But the EU AI Act introduces obligations that ICT risk management alone cannot satisfy — from fundamental rights assessments to training data governance and human oversight. Here's what the gap looks like, what the technical controls should include, and how to sequence the work before August.
Most organizations use AI for research and recommendations. But the real value appears when AI executes work autonomously — with governance, oversight, and compliance built in.
Most AI governance failures don't happen because organizations lack structure. They happen because the structure was designed for software — and AI doesn't fail like software. Here are five questions that expose the gap.
Interested in Utisha? Have questions about compliance? Drop us a message.