A serious AI automation agency starts by measuring what the AI and the automations already in place return, then automates only what has a proven return on investment. Idun Group, a consultancy specialised in AI and built on tooling, analyses your real usage from your logs, recommends what to industrialise, consolidate or cut, then builds downstream the use cases proven profitable. 62 % of companies are testing artificial intelligence but only 23 % have deployed it at scale (Ifop-Converteo 2026).
A no-code scenario runs a chain of fixed rules; an AI agent reads the context and adjusts its decision. The signal that gives away a no-code provider in disguise: it automates without ever measuring. No sizing of what already exists, no estimated return on investment per use case, no recommendation to stop anything.
30 to 50 % of AI initiatives fail to return what they should (Gartner, McKinsey). Across the estates we analyse, 20 to 40 % of the AI licences paid for go unused. The stake is no longer to automate more: it is to industrialise what pays and cut what costs without serving.
1. Analysis of real AI usage: logs and exports from the platforms you administer (Azure OpenAI and in-house models, Dust, Copilot, custom agents), installing nothing and re-instrumenting nothing. Deliverables: a map across every tool, return on investment per use case, the list of what to abandon, the consolidated full cost of AI. 2. Advice and a 6-month to 2-year roadmap, licence arbitration, standardisation. 3. Development downstream, on evidence: the code we write belongs to you. Continuous insights extend the engagement. No pricing is published: every quote follows the analysis and is given in a meeting.
Demand a debrief that decides (industrialise, consolidate, cut), coverage across every tool, and a respectful method: read-only, pseudonymisation and aggregation by design, no access to your data before signature. We analyse usage, never people. The European AI regulation remains a substantive argument: transparency obligations (article 50) due on 2 August 2026, high-risk systems (annex III) deferred to 2 December 2027.
LVMH: design and operation of a generative AI agent platform in production. Airbus: citizen development at scale, AppSheet platform, governance, support agents. Leeto: strategic audit then production rollout of AI agents, automated classification under human review and validation. HynoTech: design and development of a marketplace.
