Services
AI consulting
AI on top of the work you already record — not a new system of record.
Practical AI for small and boutique IT clients. Drafting, classification, and internal assistants that read from Siebel, a .NET application, or the documents the team already keeps.
Who it is for
- A firm of tens or low hundreds of people, with an IT staff that is already busy.
- A repeated writing or sorting task whose source data already lives in a system you trust.
- Leaders who want a human to send, approve, or file the result.
What the work includes
One workflow, named
A draft or a classification that saves time this month. Not an assistant that “knows the company” in the abstract.
Grounded in your records
The fields, documents, or exports the team already treats as current. If those are wrong, the work goes back to Siebel or the .NET app before anyone calls it an AI project.
A person in the loop
Review before send. A stop when the source data is missing, instead of a fluent sentence that fills the gap.
A data boundary
What may be sent to a model, and what may not: customer data, credentials, and anything a contract already forbids. Decided in the scope, not after a pilot.
Beside the system you have
The CRM or the line-of-business app stays the place people check. The assistant does not become a second, unofficial database.
How an engagement starts
We start from a task someone did twice this week. If the source data is not good enough, the recommendation is to fix that — in Siebel or in the .NET application — before a model is introduced.
If AI consulting is the project, send a paragraph.
Name the system, what is stuck, and any date that already exists. That is enough to tell whether this practice is the right one.