From pilot to practice

Bring AI tools into everyday work across all roles of a 70‑person international team and increase overall productivity.

A series of six workshops, from AI fundamentals to custom agents for specific business tasks, with guest experts from the industry.

Give the strategy team working AI tools for research, analysis and board materials in 8 hours per person, on confidential data: live deals, financial models, internal research.

Each participant picked a real work case and measured the baseline time. Two workshops followed, on research tools and on desktop agents, with clear rules for what can go into cloud models. The program ended with a hackathon on real strategy tasks.

Help the finance team hand its routine to AI: reports, instructions, recurring analysis.

Eight sessions in which the team built its own AI agents and dashboards for its real tasks.

Make AI part of everyday work for 70+ employees in two countries, not a one-off initiative, and first find out where each person starts.

Every participant had a video interview with an AI avatar to assess their level. Then a three-day base workshop for teams in both countries, with live translation into Portuguese, followed by an AI community with an evangelist: regular tool reviews, deep dives, guest experts and hands-on Claude Code sessions.

Employees spent up to 15% of their working time searching 10,000+ technical documents, and classic search failed on contextual queries.

We built a semantic search system that answers questions in natural language based on the company’s documentation.

Competitors change prices, promotions, payment plans and programs daily. The company couldn’t track 20+ competitors across 50+ parameters and was losing its edge.

An automated system regularly analyses competitors across discounts, new products, promotions and features, and sends a table report by email or messenger.

One program for executives of businesses that barely overlap. The goal was not hands-on tricks but a management picture for decisions about AI.

Three blocks: foundations and the tools of an executive’s daily stack; business automation and agents; a strategic discussion of competition when companies run teams of agents around the clock.

The advanced track of an executive program, centred on Claude Code: no more than an hour of theory, the rest hands-on.

Setup moved outside the session with step-by-step guides for Windows and Mac. The module was built around one working scenario instead of a tour of features.

The bank’s product community already used AI. It needed live examples of AI in product work at large companies, from practitioners rather than trend overviews.

An online mini-conference of four talks by guest speakers, from the paradigm shift to concrete cases in product, engineering and fintech, followed by a long Q&A.

In one public broadcast, bring new users into AI, win back those who tried and dropped off, and give advanced users new scenarios.

A series of applied scenarios instead of theory followed by practice, a moderation team for a fast-moving chat, and a two-hour rhythm without a drop in attention.

The team that markets an AI product knew the tools but didn’t use them systematically in its own work.

Two onsite days of 2.5 hours each: AI foundations for marketing and advanced prompting, then automating marketing routines and analysing competitors and trends.

Most employees didn’t use AI, and many were worried about data security. The task was to teach them to use it safely and for real work.

A survey split the audience by maturity. We mastered the client’s closed corporate AI platform in advance and adapted every scenario to it, using real development cases.

Move the fund’s analytics, research, parsing and data aggregation from manual work to AI agents and automations.

Twelve sessions from Claude and connectors to Cursor, Claude Code and multi-agent setups. The second half was rebuilt around topics the participants proposed, and the program ended with a hackathon on the fund’s own data.

Speed the team up with AI and make sure the training actually changed how projects are delivered.

Seven sessions, from LLM basics and knowledge bases to n8n, multi-agent systems and agentic programming, including the economics of automation, followed by a final hackathon.

Everyone used chat LLMs, almost nobody used agents. The agency wanted a hackathon within a week to make agents part of its processes.

Short pre-recorded Cursor lessons and setup guides instead of a long course, a five-hour sprint with four experts, and judging on completeness, autonomy and business impact.

Teach back-office staff, almost 40% of them lawyers, to use AI on the company’s own platform, starting with contracts and internal documents.

First we addressed the two main fears, data security and accuracy. Then practice on real tasks: checking a legal document against internal requirements, reviewing NDAs.

Train the whole company at once, from the CEO to engineers and sales, strictly on work tasks and with a skeptical technical audience.

We surveyed every participant beforehand, adapted the cases to their roles, and left cheat sheets and a chat for questions after the intensive.

A course for the client’s education platform that takes paperwork off healthcare staff: official letters, protocols, responses to complaints.

Research first, with 90 healthcare specialists from 15 regions. Then turnkey production, from scripts to filming and editing.

Bring AI into the daily work of the finance, HR and legal teams.

A two-month program on the teams’ own tasks, with a final hackathon.