AURANOM · INDEPENDENT RESEARCH PROGRAMME

Autonomous multi-agent systems for management consulting

AURANOM studies whether management consulting can be carried out autonomously by AI agents — and where the human has to stay in charge. The research is guided by five ISO standards, with a human decision at every critical step.

The conceptual framework is being tested in a field study; a prototype comes next. It is a conceptual framework, not a validated or finished product.

Pilot · 1 human Crew · AI agents Rings · 5 ISO standards

The research in numbers

  • 12months of AI-led development
  • 8experiment repositories, ~3,770 commits
  • 241analysed review-pipeline runs
  • HUMANEvery merge is a human decision

Figures as of October 2026, from the AURANOM research log and the project repositories.

THE RESEARCH QUESTION

Can management consulting be carried out autonomously by AI agents?

The programme asks which parts of consulting work AI agents can take over on their own, and where the human has to decide. It is explored in practice — as field experiments inside real development and consulting work, not in a lab.

An autopilot with a pilot: the agents are the crew — the human steers.

What the research looks at

  • Which consulting tasks AI agents can carry out end to end
  • Where a human decision is required, and how it is enforced
  • How requirements of five ISO standards can be checked by machines
  • How agent roles, sprints and approval gates work as a team
  • Whether named AI personas change trust and results

RESEARCH ARTEFACTS

Three research artefacts

All research artefacts
  • Status: Field study

    01 · CONCEPTUAL FRAMEWORK

    Conceptual framework

    for autonomous multi-agent systems in management consulting

    A 13-level model for structuring autonomous, AI-supported consulting processes, with the human as the final escalation instance.

    Learn more about the framework
  • Status: Prototype

    02 · CONTENT PIPELINE

    AURANOM Busina

    An agent pipeline for content, with human approval, checked images and distribution.

    Learn more about Busina
  • Status: Beta

    03 · SCRUM TEAM + REVIEW PIPELINE

    AURANOM Spektra

    An AI agent Scrum team that plans and builds sprints, and a multi-agent review pipeline that reviews them in VS Code. Merging stays a human decision.

    Learn more about Spektra

HOW AURANOM IS BUILT

One human, many agents

The whole project is built by AI agents — and the agents themselves are built by AI agents. A single researcher leads and makes every design decision; AI agents with fictional, clearly labelled personas do the specifying, coding, reviewing and auditing.

This development team is separate from the AI agents studied in the research artefacts.

  • Portrait of Christian MikolaschHuman leadSteers and decides
  • AI-generated portrait of the agent persona Lukas SteinerLukas SteinerPM / Coordinator · AI
  • AI-generated portrait of the agent persona Benjamin TanakaBenjamin TanakaDeveloper · AI
  • AI-generated portrait of the agent persona Dr. Lena RossiDr. Lena RossiAuditor · AI
  • AI-generated portrait of the agent persona Mag. Elara Vance-StanglElara Vance-StanglInnovation · AI
  • AI-generated portrait of the agent persona Prof. Dr. Friedrich SonderhausenF. SonderhausenResearcher · AI

GOVERNANCE

Guided by five ISO standards

The conceptual framework maps technically implementable requirements of five standards as machine-checkable rules. A governance layer checks every agent action in advance, and every step goes into an audit trail.

No ISMS and no certificate is claimed. Certification is not part of the project.

  • ISO/IEC 42001

    AI management: AI results are checked before they are released.

  • ISO/IEC 27001

    Information security: implementable requirements as checkable rules.

  • ISO 20700

    Management consulting: engagement steps checked and recorded.

  • ISO 21500/21502

    Project management: status always shown with icon, colour and text.

  • ISO 9001

    Quality management: defined processes, checked results.

Human approval at every critical step: no agent result is released unchecked, and rule violations go straight to the human.

Interested in the research?

Send us a request — about the research, a collaboration or one of the research artefacts. A human on the AURANOM team reads and answers every one.