Facilitated AI pilots · validation stage

A clear decision before deploying AI.

AURA helps leadership teams frame an AI decision before deployment: one defined process, visible risks, an accountable owner and the smallest useful test.

Frame my AI projectCompare the two primary formats

AURA / BABEL · INTERACTIVE EXPERIENCE

Build a shared AI language before deciding what to deploy.

Play solo or as a team: 1–20 players, six zones and guided challenges in English or French. No form before you begin.

Play BABEL

3 DECISION SITUATIONS

Recognise your situation before choosing a format.

AURA starts with the decision that must become defensible. These signals point to SHIFT or Mission 2032; final fit depends on scope and human review.

01Bounded workflow

Can AI improve this workflow without shifting the risk?

The workflow, owner and decision horizon are known, but value, data, dependencies or human controls remain uncertain.

ORIENTATION · SHIFT
02Use-case portfolio

Which use case deserves the next proof?

Several ideas are competing. Leadership must compare value, feasibility, risk, reversibility and ownership before allocating resources.

ORIENTATION · SHIFT or Mission 2032 after human review
03Cross-functional alignment

Is the organisation ready to make one shared decision?

Leadership, operations, technology, data and governance do not yet share the same view of opportunities, boundaries and ownership.

ORIENTATION · Mission 2032

Two clear starting points. Tailored work only after a call.

Choose one bounded workflow or a cross-functional leadership mission. Every other scope is discussed only after the decision has been clarified.

Focused decision01

SHIFT · Workflow decision

A concentrated workshop on one bounded process, its friction, potential value, risks and human controls.

OutcomeOne documented decision: PILOT, PREPARE, REDESIGN or STOP.

Duration
3 h 30 + D+30
Format
Up to 8 process stakeholders
  • Current workflow, boundary and accountable owner map
  • Friction, data, dependency and constraint register
  • Human-control checkpoints, permissions and stop rule
  • Reversible pilot hypothesis with baseline and counter-metric
  • Documented decision and D+30 evidence review
BEST FORAn accountable sponsor and process owner with one concrete AI question
Scope boundaries
  • Automation development
  • Systems integration
  • Cybersecurity audit
  • Production deployment

CHF 3’500 · View format and scope

Introduce my SHIFT need
Experience02

AURA — Mission 2032

A facilitated four-hour experience revealing readiness, friction, opportunity and risk across functions and levels.

OutcomeAn aggregated readiness snapshot and one responsible next decision: PILOT, PREPARE, REDESIGN, CONSOLIDATE, PAUSE or STOP.

Duration
4 hours + sponsor debrief + D+30
Format
12–40 participants
  • Aggregated readiness snapshot, with no individual or external /100 score
  • Prioritised portfolio of friction, opportunities, risks and dependencies
  • Cross-functional map of understanding gaps and accountable owners
  • Documented next decision and evidence gate
  • Sponsor debrief, reference report and D+30 review
BEST FORLeadership teams facing an AI decision
Scope boundaries
  • Individual employee scores
  • External /100 score
  • Automatic champion or blocker labels
  • Certified audit, ROI or adoption guarantee

From CHF 6’900 · View format and scope

Prepare Mission 2032

Compare formats and pricing.View the catalogue

A CONCRETE EXAMPLE

What does a decision prepared with AURA look like?

Read an illustrative note: workflow, hypothesis, measurements, human control and next step.

View a sample deliverable

TECHNOLOGY UNDER CONDITIONS

AI provides capability. Trust sets the conditions. Web3 is used only when it makes a required proof verifiable.

AURA starts with the decision and prefers the simplest sufficient architecture.

01

AI-first

SITUATION
Analysis, decision support, productivity or personalisation within a controlled organisational boundary.
AI ROLE
Assist, generate, predict or automate one bounded task.
TRUST EVIDENCE
Data quality, access rights, logs, human control, baseline and counter-metric.
WEB3 CONDITION
No Web3 by default.
02

Trusted AI

SITUATION
Traceability, security, identity, provenance or governance are critical.
AI ROLE
Operate within measurable, reversible and attributable boundaries.
TRUST EVIDENCE
Dated source, human owner, audit trail, permissions and stop rule.
WEB3 CONDITION
Only when independent organisations must verify the same state without a sufficient central authority.
03

AI × Web3

SITUATION
Identity, ownership, provenance, reputation, coordination or payment must remain verifiable across independent parties.
AI ROLE
Intelligence or execution within an explicit scope.
TRUST EVIDENCE
Shared state, rights and verifiable events.
WEB3 CONDITION
Only after legal, architectural, economic and security review, under a separate scope.

If the protocol, network, token or smart contracts are the business model itself, the need belongs in the Startup path or a separate venture scope, not the default AURA format.

METHOD

The AURA evidence loop.

  1. 01

    Reveal

    Make concepts, friction and assumptions visible.

  2. 02

    Assess

    Read data, governance, security and human capacity together.

  3. 03

    Prioritise

    Compare use cases on value, feasibility, risk and reversibility.

  4. 04

    Decide

    Choose PILOT, PREPARE, REDESIGN, CONSOLIDATE, PAUSE or STOP.

  5. 05

    Prove

    Run the smallest useful test with an owner and criteria.

  6. 06

    Learn

    Review evidence at D+30 and D+90 before expanding.

DECISION GUIDE

How do AI use cases become a bounded pilot decision?

A method for comparing value, feasibility, risk, reversibility and accountability before deployment.

Read the AURA guide

PUBLIC TRACK RECORD · 2023–2025

Build ventures. Share knowledge. Inform decisions.

Event organisers and innovation organisations document the public activities of Agartha and its cofounder Arnaud Depres.

2025

GEC · Blockchain and AI · 3 April

GEC lists Arnaud Depres, Agartha’s cofounder, as a speaker at its Geneva session on blockchain and artificial intelligence.

2023

UNIGE · Entrepreneurship · 15 November

The University of Geneva’s agenda lists Arnaud Depres as Agartha’s cofounder among the speakers at “Pitch and Networking”.

2023

swisstech · CES Las Vegas

Agartha Labs is listed among the announced swisstech pavilion exhibitors at CES 2023 in Switzerland Global Enterprise’s press release.

These records document speaking and entrepreneurial activities. They are not client testimonials or measured AURA outcomes.

Frame my AI projectView a sample deliverable

FAQ

Useful questions before we start.

Is AURA a certified AI audit?

No. AURA is an accompanied decision pilot producing an aggregated session snapshot, indicative readiness signals and testable hypotheses. It does not certify compliance, maturity or performance.

Who is the priority fit?

The priority cohort is a knowledge-intensive service organisation with 50–249 employees, a concrete AI decision within 90 days, an accountable sponsor, a process owner and a cross-functional group.

Does AURA score individual employees?

No. Outputs are aggregated. AURA does not infer champions, blockers or resistance from facial expressions, silence, hesitation, tone or speed, and it does not transmit individual scores to the employer.

What happens after the session?

The target cadence is a sponsor debrief 24–48 hours later, a reference report within five working days and an included D+30 review of action taken. Timing and delivery capacity are confirmed in each written proposal. D+90 is a separate evidence gate.

Does AURA deploy an AI solution?

No. AURA diagnoses and structures the next proof. Architecture, integration, cybersecurity, legal review, licences and production deployment require a separate scope with qualified specialists.

What can the decision be?

Before a pilot, the decision can be PILOT, PREPARE, REDESIGN, CONSOLIDATE, PAUSE or STOP. Progression is never automatic and value remains a hypothesis until measured against a baseline and counter-metric.

NEXT STEP

A facilitated validation method—not a certified audit or a promise of ROI.

Frame my AI project

Turn AI ambition into a clear next decision.

Frame my AI project

Form: about 4–6 min · Human review

A question before you get started?

Let’s talk about your AI project.

Describe what you need to prepare a conversation. Agartha reviews each request before proposing a next step.

arnaudeste@icloud.com+33 6 70 61 58 20Agartha · Geneva

You’re requesting a conversation. Submitting this form does not book a time. Privacy.