%402x.svg)
When finance asks why a number moved, the answer usually waits in the analytics queue until after the meeting it was for. In one day your analysts and business teams build three AI agents that find the cause, set it against a benchmark and draft the next action for a person to approve.
The workshop at a glance
One day at your office
Data analysts and BI developers, analytics managers, and the finance or commercial leaders who use the answers
From EUR 9,000. The cloud environment, travel inside the EU and the UK, and 30 days of Microsoft Teams support are included
Our cloud environment with a synthetic dataset. Your production systems are untouched
Engineers or a mixed data team? See the workshop for data engineers or Data AI in a Day.
What governed finance data and AI did at a global professional services group, and the day itself
Overdue receivables jump 35% in a month. The agent breaks them down by client, region, ageing bucket and payment terms, finds the two clients behind most of the rise, checks their invoice history, flags a payment-term change nobody approved and compares DSO with its six-month average.
Follow-up emails to account managers, with the invoice data filled in
The agent splits revenue growth by category, region and channel, sets it against external category growth rates, separates real outperformance from a market tailwind, and finds fast-growing segments where the company is barely present.
An opportunity brief posted to Slack and a calendar invite for the category review
When a supplier cannot deliver, the agent maps the affected products, stock and production schedules, calculates revenue at risk on committed orders and forecast, and sets out two scenarios, waiting or emergency sourcing, with the cost and customer impact of each.
On the manager's decision: a Slack alert to production, a Jira ticket for procurement, emails to key account managers
Every action waits for a person
The agent drafts each action, and nothing goes out until a person approves it. The figures in the scenarios come from the workshop's synthetic dataset.
Which dimensions to drill first, in what order and at what threshold, so the agent reaches the cause a senior analyst would.
Historical averages, industry benchmarks and internal targets that decide when a deviation is worth flagging.
When to send an email, open a Jira ticket, post to Slack or draft a calendar invite, always with a person's approval.
Tests, feedback loops and monitoring that keep the agent inside what it knows and make it say when it does not know.
Scoping call, 30 minutes
We agree the questions your participants bring and the people who should attend.
Proposal
We send the fee, the scope and the date.
Before the day
We set up the cloud environment with a synthetic dataset.
The day and after
Your team builds the three agents and one for each participant's question at your office, then gets 30 days of Microsoft Teams support from our engineers.
No. The day runs on our cloud environment with a synthetic dataset, so your production data stays outside the engagement.
One real business question each, the kind that now waits in the analytics queue. Each participant leaves with a working agent for it.
The playbook does that work. Give 1,000 people the same ERP extract and an AI assistant and you get 1,000 answers, the lesson self-service BI already taught with dashboards. Decomposition rules, benchmarks, tests, feedback loops and monitoring keep the agent inside what it knows, and it says when it does not know. Every action it drafts waits for a person's approval.
Choose this day if the business waits on your analysts for answers: analysts and the people who use their answers build agents that explain why a number moved and draft the next action. Choose the AI workshop for data engineers if delivery is the bottleneck: it builds the warehouse, CI/CD and a Jira agent. Choose Data AI in a Day if a mixed data team, analysts included, should take one project from request to merged pull request. Each is one day from EUR 9,000.
On site at your office. Inside the EU and the UK, travel is included in the fee; elsewhere it is a separate line in the proposal.
You talk to Michał Dębski, co-founder, about the questions your analysts answer by hand, and leave knowing which agents the day would build.
