> Markdown version of [/videos/100123-agentic-employees-in-world-s-most-downloaded-fintech-app?t=8](https://www.wearedevelopers.com/videos/100123-agentic-employees-in-world-s-most-downloaded-fintech-app?t=8). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Agentic employees in world's most downloaded FinTech app

How does Trade Republic run autonomous AI teammates on decommissioned laptops for just $100 a month? Discover their LLM-agnostic architecture that safely automates 60% of routine engineering tickets.

- **Speakers:** [Sasa Fajkovic](https://www.wearedevelopers.com/@sasa-fajkovic)
- **Event:** World Congress 2026 Europe
- **Published:** July 9, 2026
- **Duration:** 30:37
- **URL:** https://www.wearedevelopers.com/videos/100123-agentic-employees-in-world-s-most-downloaded-fintech-app

## Summary

Trade Republic successfully transitioned from simple task automation to deploying autonomous AI teammates across nearly 50 engineering, compliance, and finance teams. Treating these AI models as digital employees rather than generic chatbots completely shifted their organizational capabilities. To navigate strict European banking regulations, the system prioritizes deterministic, code-driven logic over raw LLM invocations. This hybrid approach delivers highly auditable, sub-millisecond decision-making that avoids exposing sensitive financial data, resolves cross-platform incident triage, and halts unchecked generative API costs.

Rather than executing massive infrastructure modernizations, the architecture is astonishingly cost-effective—operating on decommissioned physical laptops for roughly $100 a month in infrastructure spend. It features an ingenious, channel-specific confidence gating mechanism: if the agent is over 80% confident, it acts autonomously; between 50% and 80%, it drafts a solution and tags human teammates for review; below 50%, it stays quiet to prevent alert fatigue. Furthermore, continuous learning mirrors human team dynamics. A simple thumbs-up or thumbs-down Slack reaction kicks off a self-correction workflow, resulting in the agent drafting a PR to independently update its behavioral logic and knowledge base.

To manage organizational scale without creating a monolithic bottleneck, Trade Republic implemented composable knowledge layers. Core compliance rules and security hooks are embedded at the global organizational level and cannot be overridden, whereas individual agile domains and specific teams maintain granular control over localized playbooks. By storing proprietary data primarily in Markdown formats and keeping the underlying orchestration completely LLM-agnostic, the company sidesteps vendor lock-in entirely. Ultimately, this virtual workforce automates over 60% of routine tickets and saves upwards of 300 collective hours a week—proving that heavily governed, strategic AI empowers developers to focus on higher-order engineering.

**Keywords:** autonomous AI teammates, fintech compliance automation, deterministic code execution, confidence gating mechanism, slack workflow automation, self-improving agents, incident triage automation, laptop server infrastructure, LLM-agnostic architecture, infrastructure cost optimization, continuous feedback loop, composable knowledge layers, markdown knowledge base, vendor lock-in prevention, cross-platform context sharing, pull request generation

## Chapters

1. **Building an AI employee for organizational agility** (00:08) — Creating an autonomous teammate from a small proof of concept to a fully deployed organizational asset.
1. **Mimicking human problem-solving with deterministic logic** (03:56) — Using non-LLM deterministic gates to evaluate confidence before acting or escalating issues to human colleagues.
1. **Automating scheduled tasks across variable organizational functions** (07:56) — Enabling cross-functional departments like finance and compliance to directly produce and trigger regulated workflows.
1. **Enforcing strict execution security and access management** (11:25) — Implementing hardcoded deterministic evaluation hooks and permission systems to safely govern automated agent actions.
1. **Composing organizational rules without dynamic runtime lookups** (14:04) — Pre-compiling markdown-based configuration logic to share knowledge context per team while strictly enforcing global company policies.
1. **Implementing frictionless feedback loops for agent self-improvement** (16:28) — Tracking simple emoji reactions to automatically trigger background pull requests that systematically refine agent behavior.
1. **Ingesting historical data for automated knowledge graph generation** (18:52) — Scanning repositories, conversational history, and documentation during initial onboarding to natively build specialized routing instructions.
1. **Using interaction analytics to discover institutional blind spots** (19:56) — Leveraging weekly automated blog posts and aggregated interactions to proactively identify and inject missing internal documentation.
1. **Observing agent metrics and optimizing localized infrastructure costs** (21:44) — Tracing detailed execution telemetry to maintain comprehensive auditability while radically lowering expenses through repurposed bare metal hardware.
1. **Building shared conversation context and provider-agnostic capabilities** (26:03) — Maintaining continuous task awareness across diverse platforms with native architectural design built to easily swap underlying language models.
1. **Addressing automation anxiety and prioritizing impactful engineering work** (29:40) — Reframing continuous execution engines as an opportunity to eliminate repetitive toil rather than simply reducing organizational headcount.

## Related Moments

- [Shifting developer workloads and realistic AI productivity gains](https://www.wearedevelopers.com/videos/1830-wearedevelopers-live-speculaitions) (from "WeAreDevelopers LIVE - SpeculAItions")
- [Balancing developer autonomy with the adoption of coding agents](https://www.wearedevelopers.com/videos/100198-the-last-mile-of-ai-from-prototype-to-production) (from "The Last Mile of AI: From Prototype to Production")
- [Exploring AI agent usage within the software engineering industry](https://www.wearedevelopers.com/videos/1814-wearedevelopers-live-markdown-liquid-and-checkouts) (from "WeAreDevelopers LIVE - Markdown, Liquid and Checkouts")
- [Rethinking team structures around AI agent capabilities](https://www.wearedevelopers.com/videos/1539-agentic-devops-how-ai-powered-automation-transforms-software-delivery-on-github-and-azure) (from "Agentic DevOps: How AI-Powered Automation Transforms Software Delivery on GitHub and Azure")
- [Security integration and AI skepticism in developer tooling](https://www.wearedevelopers.com/videos/1830-wearedevelopers-live-speculaitions) (from "WeAreDevelopers LIVE - SpeculAItions")
- [Managing new AI workloads for non-technical employees](https://www.wearedevelopers.com/videos/100015-7-ways-to-fail-at-building-a-platform) (from "7 Ways to Fail at Building a Platform")

## Related Articles

- [How we Build The Software of Tomorrow](https://www.wearedevelopers.com/magazine/120-how-we-build-the-software-of-tomorrow)
- [Never delegate the understanding](https://www.wearedevelopers.com/magazine/749-never-delegate-the-understanding)
- [What is Agentic Programming and Why Should Developers Care?](https://www.wearedevelopers.com/magazine/625-what-is-agentic-programming-and-why-should-developers-care)
- [Navigating the AI Shift](https://www.wearedevelopers.com/magazine/629-navigating-the-ai-shift)

## Related Jobs

- [Senior AI Agent Software Engineer (Go, Python) (m/f/x)](https://www.wearedevelopers.com/jobs/48277-senior-ai-agent-software-engineer-go-python-m-f-x) at **Dynatrace**
- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [Principal Product Manager, Agent Platform](https://www.wearedevelopers.com/jobs/ext/277541-principal-product-manager-agent-platform) at **GitHub**
- [Senior Engineer, Infrastructure Platform](https://www.wearedevelopers.com/jobs/ext/328836-senior-engineer-infrastructure-platform) at **Intercom, Inc.**