AI consulting
Everyone tells you that you need to get on board; nobody tells you where to start
Walk away with a well-reasoned decision, rather than an impressive demo.
New tools come out faster than anyone can evaluate them, every vendor promises the same thing, and the trial launched six months ago never went any further.
The real risk is not missing out on artificial intelligence: it is putting your budget in the wrong place. A project that automates a rare task, or that runs into data you are not allowed to share, costs you twice: once to build it, and once to abandon it.
An outside, independent view
I look at how you actually work, identify where AI can genuinely help you, and put figures on the options that are worth it. Sometimes the answer is “none for now”: that is a useful conclusion too, and it saves you money.
I don’t develop anything as part of this engagement. If a project is chosen, you entrust it to whoever you like: that is what guarantees the independence of my advice.
- An assessment of your tasks, ranked by the time they really take
- For each option: what it brings in, what it costs, and what could go wrong
- The data question settled from the start: what may leave your premises, and what must stay
- A roadmap, from the most profitable project to the least urgent
- What is best not done, and why: that is often what saves you the most
What you receive
Documents that belong to you, and that you can use without me, including with another provider.
Assessment
Your tasks and your tools, with the time each one really takes you.
Costed options
Each option with its estimated gain, its implementation cost and its known limits.
Rules for your data
What can be entrusted to an external model, what must stay in-house, and on what conditions.
Roadmap
In what order to move forward, in steps short enough that you can stop along the way.
Debrief
A presentation to your team, with time to answer their questions.
Help choosing
The tools under consideration, tested on your own cases.
How it works
1. A free initial conversation, to check that the approach makes sense for you
2. Short interviews with the people who do the work day to day
3. The analysis, then costing of the options selected
4. The debrief, then you decide what comes next, with or without me
Still have questions?
I answer all your questions.
-
How is this different from training?
Training teaches people how to use a tool. Here, we work on your organization: which tasks, which tools, in what order and with what budget.
The two complement each other, but one doesn’t replace the other: knowing how to use a tool doesn’t tell you whether to adopt it.
-
Will you then sell me the development?
I can take it on, and often do. But the roadmap is written so that any provider can follow it. Advice that always ended with “give me the project” would just be a sales pitch.
-
Our data is sensitive. Is that a deal-breaker?
No, but it changes the available options, and that is precisely a good reason to take stock before getting started. Many projects fail on the day someone finally asks where the data goes.
Depending on how sensitive it is, we go for a model installed on your own servers, a provider that commits in writing, or a project limited to anonymized data.
-
How much time does it take?
On your side, a few hours spread over two weeks: the interviews with the people involved, and the debrief.
The total duration depends on how many departments are involved. I let you know after the first conversation, which is free and with no obligation.