Azimut AI
Illustrative scenario. The company on this page doesn't exist: it's a typical profile for the sector, built to explain how Azimut would be used. It doesn't describe any customer or claim measured results.
The company in this scenario
- 180 people: 140 consultants across four practices and 40 in support functions.
- Runs on Microsoft 365: SharePoint for deliverables, Outlook and Teams, plus a CRM for the sales pipeline.
- Around fifteen proposals a month, half of them with a one-week deadline.
- Brings in a new class of junior consultants every September.
The problem
In a consulting firm the product is what the firm knows, and the firm doesn't know where it keeps it. Every proposal starts by searching SharePoint for a similar one, asking on Teams who ran that pricing project, and rewriting credentials that have already been written ten times. Client meeting minutes depend on whoever stays late. When a manager leaves, the context of their accounts leaves too. And many consultants already use an AI tool on their own to draft, pasting client information into a tool the firm doesn't control.
How it would work with Azimut
Azimut would connect to SharePoint, email and the CRM with the permissions that already exist: each consultant would only see the projects they can already see. On that foundation, the firm would publish four shared agents.
| Agent | What it does | Who uses it | What gets measured |
|---|---|---|---|
| Proposals | Writes the first draft from the RFP or the client's email, reusing the firm's proposals, methodologies and CVs. | Managers and partners | Hours per proposal and win rate. |
| Credentials and experience | Answers "what similar projects have we done, and who ran them?" by searching deliverables and the CRM. | The whole firm | Time to find the reference. |
| Client minutes | Turns a meeting transcript or notes into minutes, agreements and next steps, and files them in the project folder. | Consultants | Writing hours per week. |
| Onboarding | Answers newcomers' questions about methodology, templates and how each practice works. | Junior consultants | Weeks until the first deliverable that doesn't need redoing. |
The proposals agent would use the top-capability model, because it has to hold together a long, coherent document; the minutes agent a mid-tier one, faster and cheaper per use. Every consultant picks the model per task from OpenAI, Anthropic, Google and Mistral, and the agents keep working if the firm switches providers.
Where you'd start
You'd start with the credentials and experience agent. The whole firm uses it from day one, it doesn't change any process, and it forces the deliverables repositories to be connected with the right permissions, which is exactly what the proposals agent will need later. First you find what the firm knows; then you reuse it.
What's out of scope
Azimut doesn't replace the partner who signs the proposal, and it doesn't set the price: a person reviews the draft and makes it their own. It isn't a project management or time-tracking system either; if the firm wants to ask about hours, they're queried where they're already replicated, for example a data warehouse. The specific integrations are confirmed during rollout, and only the ones tested end to end get activated.
How it would be measured
Before turning on any agent, you set the baseline, with the company's own numbers rather than industry averages.
Week 0. Hours per proposal over the last ten, win rate over the last year, ramp-up weeks for the latest junior class, and how many consultants already use some AI tool on their own, which one and at what cost.
Month 3. The same numbers, next to Azimut's usage panel by user, practice, model and tool. The comparison is against the firm itself, not against an industry average.
Frequently asked questions
What about confidentiality between clients?
Agents inherit SharePoint and CRM permissions: a consultant who has no access to a client's folder gets nothing from it through Azimut either. Model providers don't train on the data sent to them.
Is it useful if we already have Microsoft 365?
Yes. Azimut uses SharePoint, Outlook and Teams as sources and adds what isn't in Microsoft: the CRM and other repositories, model choice, and agents the firm publishes and governs.
Who builds the agents?
The first ones, Azimut together with the firm during rollout. After that, the consultants who use it most build their practice's agents and publish them for everyone else.