Numbers delivered, not promises
Audits, training, AI agents, automations: what our clients changed in weeks, not quarters. Each case study covers the starting point, our approach and the results.
Book a call
100+ staff equipped with AI, around fifteen agents in production
A European Private Equity fund and 100+ staff. We rolled out Copilot and Perplexity Enterprise to every team, trained all levels, and built around fifteen specialised agents and skills per function.
6 AI agents on Dust across the whole investment process
An early-stage Venture Capital fund between Paris and Silicon Valley. 6 specialised Dust agents cover every dimension of deal analysis, from tech to exit, supporting investors and IC memo drafting.
Thousands of quotes processed monthly, no manual data entry
An industrial printer receiving thousands of PDF quotes per month, all processed by hand. AI extraction pipeline + ERP integration, operators freed from data entry.
Millions of AI executions a year, from the edge to the core business
France's B2B data reference, 450,000 decision-makers and 100,000 companies tracked. Its data experts stay on verification, Tandem operates the AI pipelines and agents on n8n on a continuous retainer, at the scale of millions of executions.
A PDF quote goes in, an ERP-ready file comes out
Industrial joinery, multi-leaf products, close to a thousand lines of business rules. An n8n workflow reads every quote, matches it to the catalogue and produces the ERP integration file in 30 seconds.
A field audit to sort what can be automated from what must be reviewed
An estimating department copying hundreds of quote lines by hand and sorting dozens of emails a day. A few weeks of audit to prioritise the use cases and structure eighteen years of documents.
One form, a complete SEO article published as a draft in the CMS
Source research, on-brand writing, cover image, multi-entity publishing: eight manual steps reduced to a single form and three to five minutes of processing.
A 15-day field audit, several costed use cases
A network of 22 dental centres, with practices that differ from one site to the next. Fifteen days alongside the teams to map the patient journey, spot the bottlenecks and prioritise use cases by volume and ROI.
Part of the support desk automated, without health data leaving the perimeter
Imperfect scanned documents, regulatory deadlines to meet, health data that must not be exposed. A self-hosted workflow reads, verifies and routes incoming requests, automatically falling back to human review whenever a case leaves the standard path.
Three AI opportunities, quantified on the brand's own data
No AI in production and little visibility on where it would actually pay off. An audit run on the brand's own data, delivering three costed, sequenced, build-ready use cases.
From a handful of pioneers to 90% building their own agents in three months
An AI agent platform already deployed at scale, but usage carried by a few. Three months on the ground to make a commercial organisation self-sufficient: more than 50 agents built by 20 people across 5 business teams.
Latest articles
AI for medical reception: what the agent handles alone, what it hands over
Picking up, identifying the caller, offering a slot and confirming can be handed to a machine. Judging an emergency cannot.
Read article →AI for insurance brokers: cancellations, claims and back-office work
Reading a letter, finding the policy, meeting a legal deadline and routing the file can be handed to a machine. Signing the advice cannot.
Read article →AI agents for financial reporting: closing, variances, forecasts
Consolidating exports, checking totals and commenting on variances can be handed to a machine. Deciding what goes to the board cannot.
Read article →