Workshops: Agentic Coding for Roles & Industries

Agentic Coding in Banks & Insurers: Workshop under DORA and BaFin

Development teams in banks and insurers want the same gains from coding agents as everyone else, but they work under the EU Digital Operational Resilience Act (DORA), BaFin supervisory expectations and strict outsourcing rules. This workshop shows how to bring tools such as Claude Code, GitHub Copilot or OpenAI Codex into regulated development with controls internal audit and supervisors can follow.

Who is this workshop for?

  • Development and architecture teams at banks, insurers and financial service providers
  • IT risk management, information security officers and ISMS teams
  • Outsourcing and third-party risk management functions
  • Internal audit teams with an IT audit focus
  • IT service providers that build software for financial institutions

What you will take away

  • Classify coding agents and their providers within your ICT risk management and third-party framework.
  • Map agent-assisted development onto change, test and release processes, including the four-eyes principle.
  • Define which data classes and repositories agents may access and which environments stay off limits.
  • Set up logging and evidence so agent contributions stay traceable for audit and supervisors.
  • Prepare the key questions for contracts and exit strategies with AI tool providers.

Agenda

  1. Module 1

    Regulatory frame

    • EU DORA: ICT risk, ICT third-party risk, incident handling, resilience testing
    • BaFin expectations: BAIT and VAIT as background, and how DORA shifts the picture
    • EU AI Act: coding assistants versus high-risk systems such as credit scoring
  2. Module 2

    Tools, providers and data

    • Enterprise offerings, data processing terms and data residency
    • Coding agents as ICT third-party services: register of information, contracts, exit
    • Data classes that must never appear in a prompt
  3. Module 3

    Controls in the development process

    • Four-eyes principle and segregation of duties for agent-written code
    • Permission modes and sandboxes without access to production or customer data
    • Exercise: extend your change process for agent contributions
  4. Module 4

    Traceability and audit readiness

    • Labelling agent-assisted commits and pull requests
    • Agent logs, review evidence and test records
    • What internal audit is likely to ask
  5. Module 5

    Practice on a bank-style codebase

    • An agent-assisted change to a core service with tests and review
    • Security and dependency checks in the pipeline
    • Lessons learned for your rollout plan

Why financial institutions need their own approach

In banking and insurance, a coding agent is more than a developer tool. Depending on the setup, it is at once a third-party ICT service, a new data flow to a provider and a change to your development controls. DORA puts ICT risk, third-party risk and operational resilience at the centre; here DORA means the EU regulation, not the DevOps metrics of the same name.

We explain the requirements in practical terms and show how agentic coding fits into existing frameworks. The legal and supervisory assessment stays with your compliance, legal and outsourcing functions.

Typical controls we work out

Each control is assigned to an existing process owner so it survives beyond the pilot.

  • Approved tools only, via enterprise contracts, no private accounts
  • Agents in isolated environments without production credentials
  • Mandatory human review and approval before merge
  • Logging of agent sessions in line with your retention rules
  • Regular review of the provider in third-party risk management

Background and format

jocapps, the company behind neooutput, is certified to ISO 27001 and ISO 9001 and works with auditable processes itself. Our AI Solutions Architects and AI Code Engineers train remotely or on-site, in German or English. If code must not leave your institution, we use sandbox repositories that mirror your stack.

Frequently asked questions

Are regulated institutions allowed to use coding agents at all?

There is no general ban. What matters is that the tools are integrated into ICT risk management, third-party risk management and development controls, and that data flows are documented. The binding assessment for your institution remains with compliance and legal.

Can agents work on core banking or policy administration systems?

Yes, under clear conditions: isolated environments, synthetic or anonymised test data, no production access and full review. Long-grown systems in COBOL or older Java benefit in particular from agents that explain and document existing logic.

Which tools are suitable for banks and insurers?

Usually enterprise offerings with contractual data protection commitments, admin controls and audit logs, for example from Anthropic (Claude), GitHub, OpenAI or European providers such as Mistral. We compare options against your requirements, vendor-neutral.

Workshop

Get in touch

Want to use coding agents in a way your auditors and supervisors can follow?

Tell us roughly how many people will join and pick dates that work – we will get back to you within 24 hours with a proposal for content, format and schedule.

Prefer to talk first?

15 minutes, free and without obligation – with our AI Trainers and AI Code Engineers.

Book a call for free (opens in new tab) +49 761 519 751 50 (24/7)

Request a training / workshop

Workshop: Agentic Coding in Banks & Insurers: Workshop under DORA and BaFin · 1 day

Dates that work for you (optional)

Suggestions about 2 weeks ahead – pick as many as you like

Reply usually within 24 hours · jocapps is ISO 27001 certified · Privacy