WeAreDevelopers Logo WeAreDevelopers
Sign in
For You LIVE
World Congress 2026 Europe - Virtual Stage • Jul 2, 2026 • Session details

Taming the Beast: Building Autonomous Agents to Solve German Tax Bureaucracy

Mohamed Dhiab

The future of AI agents isn't bigger models—it's better engineered loops. Discover how Mika built a strict graph architecture to safely automate complex German tax bureaucracy.

Pause
Mute Enter Fullscreen
#1 about 3 min

Automating the complex German tax bureaucracy system

Navigating legacy structures, contradictory rules, and unstructured data requires high legal correctness.

#2 about 2 min

Measuring automation success with touchless booking rates

Achieving a 97 percent automation rate requires a rigorous definition of zero human edits from raw documents to final entries.

#3 about 2 min

Structuring code with deterministic graphs and smart nodes

Replacing hallucination-prone single guard prompts with strict deterministic graphs ensures legal step order while leveraging AI only for reasoning.

#4 about 2 min

Processing documents via deterministic layers and reasoning agents

Raw documents pass through duplication filters before specialized AI agents extract line items and assign standard accounting codes.

#5 about 1 min

Validating tax compliance using composable decision trees

Deterministic rules form decision trees that verify booking candidates against specific tax law paragraphs for audit readiness.

#6 about 2 min

Converting human decisions into permanent structured memory rules

A gatekeeper agent surfaces necessary to-dos, turning every human resolution into a persistent rule without requiring model fine-tuning.

#7 about 2 min

Scaling accountant capacity through AI workflow automation

Persistent memory rules enable accountants to manage eight times more customers while maintaining a high touchless booking rate.

#8 about 2 min

Modeling document-to-transaction matching as an NP-hard problem

Matching multiple invoices to multiple payments creates a computationally explosive subset matching challenge that requires simplification assumptions.

#9 about 3 min

Solving transaction matching with a layered pruning system

A layered pipeline uses deterministic filters and gradient-boosted relevance rerankers to prune pairs before an agent handles complex edge cases.

#10 about 3 min

Essential patterns for building reliable agents in regulated domains

Designing predictable loops, prioritizing cheap pruning, and replacing single prompts with specialized agents ensures compliance in high-stakes environments.

Matching moments

2:06 min

Rethinking team structures around AI agent capabilities

Mike Mike · World Congress 2025

3:20 min

Automating regulatory compliance through distributed AI agents

Olga Woschitz Olga Woschitz +1 · Europe 2026 Virtual

3:17 min

Building reliable AI agents for business value

Kesha Mykhailov Kesha Mykhailov · World Congress 2026 Europe

1:05 min

Key takeaways for architecting reliable software agents

Nimrod Kor Nimrod Kor · World Congress 2025

1:19 min

Turning repetitive developer tasks into automated AI agent skills

Markus Eisele Markus Eisele · World Congress 2026 Europe

2:34 min

Balancing developer autonomy with the adoption of coding agents

Clemens Wasner Clemens Wasner +4 · World Congress 2026 Europe

More from World Congress 2026 Europe - Virtual Stage

One Does Not Simply Make Architecture Decisions Alone thumbnail
36:15
2 months ago

One Does Not Simply Make Architecture Decisions Alone

Vladas Diržys

From Black Box to Glass Box : Bedrock AgentCore Observability thumbnail
21:15
2 months ago

From Black Box to Glass Box : Bedrock AgentCore Observability

Yasemin Aktürk

Leading in Rhythm: How Women Make Better Decisions with Biological Intelligence thumbnail
30:41
3 months ago

Leading in Rhythm: How Women Make Better Decisions with Biological Intelligence

Julia Grosse

Better Together: Leveraging Your Observability Tools as a SIEM thumbnail
39:06
3 months ago

Better Together: Leveraging Your Observability Tools as a SIEM

Mathias Palmersheim

ChatGPT Apps: From Chat to Checkout thumbnail
30:29
3 months ago

ChatGPT Apps: From Chat to Checkout

Aaron Koivunen

Interior Decorating (TypeScript Decorators) thumbnail
29:22
3 months ago

Interior Decorating (TypeScript Decorators)

Michael Dowden

Upcoming sessions on this topic

Open session

World Congress 2026 North America

September 25, 2026 · 12:20–12:50

Stage 4

Give the Agent a Budget, Not a Token

Sachin Malhotra

MTS @Anthropic

Sachin Malhotra
Open session

World Congress 2026 North America

September 24, 2026 · 11:40–12:10

Mainstage

I Don't Trust AI Agents (And Neither Should You): Building Production-Ready Architectures

Darko Mesaros

Distinguished Developer Advocate at AWS

Darko Mesaros
Open session

World Congress 2026 North America

September 25, 2026 · 11:40–12:10

Stage 6

Codifying Trade-offs: Security, Cost, and Compliance as Agent Guardrails

Suzanne Daniels

Chief Developer Advisor at Microsoft

Suzanne Daniels
Open session

World Congress 2026 North America

September 24, 2026 · 14:50–15:20

Stage 7

When Agents Became Users: Rearchitecting Identity and Permissions for AI at Scale

Yoav Gal

Product Lead

Yoav Gal
Open session

World Congress 2026 North America

September 24, 2026 · 15:30–16:00

Stage 5

The reviewer can't be the author: independent verification for AI-generated code

Manish Kapur

VP, Product and Solutions at Sonar

Manish Kapur
Open session

World Congress 2026 North America

September 23, 2026 · 13:15–15:15

Stage 9

DeepAgents: Build Multi-Agent AI Systems That Actually Work

Anagha Rumade, Anjana Umapathy, Apoorva Jaiswal

Anagha Rumade
Anjana Umapathy
Apoorva Jaiswal

Related videos

Agentic employees in world's most downloaded FinTech app thumbnail
30:37
World Congress 2026 Europe • 2 months ago

Agentic employees in world's most downloaded FinTech app

Sasa Fajkovic

The Missing Layer Between Enterprise Data and AI Agents thumbnail
30:17
World Congress 2026 Europe • 2 months ago

The Missing Layer Between Enterprise Data and AI Agents

Jannis Eickenroth, Sebastian Klenke

Hard Problems Hide in Boring Places: Turning Accounting Workflows into AI Products thumbnail
21:52
World Congress 2026 Europe • 2 months ago

Hard Problems Hide in Boring Places: Turning Accounting Workflows into AI Products

Oleksandr Korotkykh, Tolga Sümer

When Should You Use an Agent? Architectural Trade-offs in Agentic Systems thumbnail
32:48
World Congress 2026 Europe • 2 months ago

When Should You Use an Agent? Architectural Trade-offs in Agentic Systems

Matheus Guimaraes

Event-Driven AI Agents: Orchestrating Long-Context Legal Processing at Scale thumbnail
32:26
World Congress 2026 Europe • 2 months ago

Event-Driven AI Agents: Orchestrating Long-Context Legal Processing at Scale

Luca Bianchi

When Agents Meet Legacy: Never Change a Running System thumbnail
28:57
World Congress 2026 Europe • 2 months ago

When Agents Meet Legacy: Never Change a Running System

Michael Friedrich

Related articles

From learning to earning

Jobs that call for the skills explored in this talk.