Advisory
Four areas in which we work
Indicative durations. Pricing is established case by case, in a meeting.
01
Framing and prioritising use cases 10 days
Your teams have already begun, often without you. We record the real use cases, qualify them by value and by risk, and establish the order in which to address them. A use case that shortens no deadline and lightens no workload does not deserve to be industrialised.
You leave with
- The inventory of use cases, including those outside IT's sight
- A value / risk / effort matrix
- A six-month roadmap, with decisions taken
02
Governance and compliance 15 days
The European AI Act imposes a classification of systems and the obligations that follow from it. We situate your use cases within that framework, articulate them with the GDPR and your security policy, and write the rules your teams will apply.
You leave with
- The classification of your systems and the associated obligations
- An AI usage policy, enforceable internally
- The control arrangement and its checkpoints
03
Sovereign AI infrastructure 20 days
Hosting models is an infrastructure decision: compute base, data location, applicable jurisdiction, access conditions, total cost. We treat it with the same method as any hosting decision, and we hand you the model.
You leave with
- Scenarios compared on five axes, sovereignty among them
- The techno-economic model, which belongs to you
- The tender specification
04
Cost and footprint of AI usage 10 days
Training, inference, corpus storage: AI is a line of spend and emissions. We attach it to your digital services using the Cigref model, and put it on your governance agenda rather than in a year-end report.
You leave with
- The cost and footprint of your use cases, by service
- The decisions to examine before committing
- Integration into Green PMO
The principle
The model comes to your data. Your data does not go to the model.
It is the question everyone asks and almost nobody answers properly: where do your documents go when an artificial intelligence analyses them?
With our solutions, the data stays under control
A dedicated connector lets the model reason on your data where it sits. Only the context windows required for a given analysis pass through, for the duration of that analysis. No document is transferred, none is retained, nothing is reused to train anything.
The architecture is proven: it runs in production on a corpus of several hundred thousand documents, in a setting where confidentiality was not negotiable.
Training
An arrangement your teams cannot steer will not last six months
Our six modules all address the same subject: running your work with AI. Project management, agile, Lean 6 Sigma, FinOps, GreenOps, cybersecurity, designed and delivered by the practitioners who apply them on engagements.
We also coach project managers and PMOs on their own files, once the training is done.
Artificial intelligence does not replace the project manager. It removes the part of the work that consumed time without calling on judgement.
Our tools
What we built for our own engagements
We are not a software publisher. These three tools were built because our engagements needed them, and they are included in our advisory offers. Above all they are the proof that we know what we are talking about.
Esiense Discovery In production
Document corpus analysis
Processes corpora nobody has time to read and extracts the facts. Internal investigations and forensic analysis, audits, contract reviews, due diligence, responses to regulatory requests, legal expertise.
Mailboxes · PDFs including scans · office documents · encrypted files
Actual status
In production at a client, on a corpus of several hundred thousand documents.
Esiense PreVista Pre-release
Governance, risk and compliance
Holds your control framework, your risks and your evidence in a single place, and maps them to the applicable standards. Establishes gaps on evidence rather than on assertions, and prepares reviews and audits.
Controls · risks · evidence · action plans · reviews
Esiense Carbon Pre-release
Cost and carbon management
Connects your consumption data, your IT budget and your emissions using the TBM model. Identifies the lines that really weigh, AI usage included, and tracks the effect of the actions taken.
Cost-to-carbon conversion by reference tables · carbon amortisation of assets · allocation to digital services
Design principle
A result on day one, not after six months of data collection
Most governance tools have to be filled in before they are good for anything: asset inventory, application map, control framework. That is what makes the majority of deployments fail in organisations with neither a dedicated team nor time to spare.
We designed our tools the other way round. You get a usable result from the first session, from what you already have. The data you add afterwards refines the analysis; it never gates access.
Hosting
On your premises, or on ours. Never in a shared environment.
Deployment in your environment
The tool is installed on your premises. No data leaves your information system. We provide a turnkey solution; you retain complete control of it.
Hosting by Esiense
On sovereign European infrastructure, with a dedicated instance per client. No shared database, no shared resources, no mere logical separation between organisations within one system.
Both options apply to all of our tools, with no consequence for functionality.
The commercial model
Included during the engagement. Yours if you keep it.
Our tools are built into our advisory offers: you have access at no extra cost for the whole duration of the engagement.
At the end, two possibilities. Either the arrangement we put in place lives without a tool, and you keep nothing. Or your teams want to keep steering continuously, and you retain access on a subscription.
That is consistent with what we promise you. You stop needing our advice; you keep what makes you self-sufficient.
Let's talk about your use cases
The partner you meet leads the engagement. A demonstration can fit into the same meeting.