← // SELECTED WORK

Tax returns. Minus the repetition.

How we moved a tax workflow from manual data entry to production automation, freeing thousands of accountant hours per month.

Result

≈ +6,000 hours/month freed

Conservative estimate: ~1,000 returns/month × ~6h of manual accountant work each. Actual production volume is higher.

Situation

A US VC-backed fintech startup aimed to fully automate the US tax return filing workflow. The main challenge wasn't just interacting with legacy software via computer vision — it was making the system scalable at volume to generate real economic return. Going from code that works on one machine to massive operational scale is where most automation projects fail.

Decision

After a 30-day PoC, we engineered an autonomous agent combining OCR and Computer Vision to navigate Windows desktop environments and interact with tax software. To guarantee horizontal scalability, we focused on two technical pillars: 1. Stateless Architecture: We made the agent completely stateless. Session state is irrelevant and fully cleared between runs, allowing the system to scale without memory or process conflicts. 2. Structured Observability: Scaling means managing thousands of interactions. We implemented rigorous log management to avoid being flooded with indecipherable output. When errors occur — inevitable when interacting with software designed for humans — tracing lets you find the root cause and act fast.

Result

In three months, the workflow moved from manual work to production automation. Accountants no longer had to spend hours moving data through legacy software by hand. The company gained the capacity to process far more returns without hiring at the same rate, freeing an estimated 6,000+ hours per month.

OCR / Computer Vision / AI Agents / Python / LangSmith

Have a project in mind?hello@41dots.com ↗