Genuine AI fluency at every level of the firm, from analyst to partner.

Each engagement is scoped and priced individually: a six-person deal team and a forty-person firm are not the same programme. Programmes run monthly with no minimum commitment, so the decision you are making is a month's decision, not a year's.

Month one is four sessions covering the full arc, from how a model thinks through to building with Claude Code. From month two the cadence is monthly, going deep on the areas the team finds most valuable. Sessions run on your own deals, models, and memos, so the time is spent on the actual work rather than beside it.

Sessions run on your actual deals, models, and memos, so this question deserves a precise answer rather than a reassuring one. Data handling is settled during scoping, before any material is opened: whose accounts the work passes through, on what terms, and what, if anything, is retained afterwards. The answers are put in writing before the first session.

We train on the Claude suite: Claude.ai, Claude Code, and Claude in Excel, PowerPoint, Chrome, and Cowork. Solven is a member of the Anthropic Partner Network. That is a deliberate position, not an accident. On lock-in: session one teaches how any large model works, from tokens, attention, and probability to where hallucinations come from, and the verification habits we build are model-agnostic. If the market shifts, the fluency travels with your team.

Either, or both. We are based in Lisbon and work with firms from Lisbon to London, so sessions run in the language the room prefers. Written materials are in English, because the tools and the frontier are.

Three ways the firm can see rather than take on trust. The work itself: output that is defensible at investment committee, with the human contribution legible to senior reviewers. The token P&L from our token economics work: what each workflow costs and what it returns, in a form the partners can read. And the calibration reviews built into the methodology, which surface drift before it compounds.

Three honest reasons. The work is specific: we do one thing, for investment firms, and sessions run on your deals and memos rather than a generic curriculum. The cadence is the product: the technology shifts weekly, and a one-off training project, internal or external, is stale within a quarter; our engagement is monthly because the frontier is. And the thesis is different: most enablement measures adoption; we design against cognitive offloading, so judgment gets sharper rather than thinner. If what a firm mainly needs is change management across a thousand seats, a large consultancy is the right call, and we will say so.

The engagement is monthly with no minimum commitment, so ending is a decision you can make any month. What the firm keeps: the verification habits; the prompts and skills built in session four, version-controlled files that live in your systems; and anything produced in build hours. Nothing is designed to make leaving hard.

That risk is the reason the firm exists. The failure mode in most rollouts is not adoption failure; it is analysts becoming editors of drafts they cannot defend. Every session is built the other way round: the analyst produces a defensible first pass, then brings Claude in as a critic, a red-teamer, and a verifier. Used this way, the model sharpens exactly the skills that unconstrained use erodes. We call it augmentation without cognitive atrophy.

The whole firm, analyst to managing partner. Fluency is a firm-wide capability, not a specialism held by a few, and each session is paced to the people in the room. The partners' sessions matter most: senior-led review is how the verification discipline holds.