How Sprint got its operators out of quote data entry
An industrial printer of 620 staff across 12 sites, thousands of inbound quotes per month in mixed formats (email, PDF, customer files). We replaced manual data entry with an AI extraction pipeline plugged into the in-house ERP.
Absorbing the quote volume without capping capacity
Sprint receives several thousand quotes every month in varied formats: emails, PDFs, customer order files. A team of operators dedicated most of their days to entering the data into the in-house system, a high-volume task that absorbed the team's full capacity and limited the natural absorption of growth.
Sites to serve
12 sites in France producing for large accounts, fed by a heterogeneous inbound quote flow.
Quote families
6 different format families (emails, PDFs, customer files) covering most of the volume, each entered by hand.
Entry on human time
The whole entry chain ran on operator time, tying growth directly to team size.
Sprint, GennevilliersGroupe Sprint
Groupe Sprint is a French industrial operator in print and digital communication, founded in 2007 as a small digital print shop before becoming one of the sector's leaders. The group exceeds €100M in revenue and operates nationwide.
With 620 staff across 12 sites and an assumed eco-responsible approach, Sprint produces for large accounts (publishing, retail, institutional). Absorbing inbound volume without quality drift is a central operational challenge.
Extract, control, integrate : without breaking the existing ERP
The goal wasn't to replace Sprint's ERP, but to replace the data entry. We built an upstream pipeline that takes raw quotes in and returns structured data ready to be ingested by internal tools, keeping a human in the loop on ambiguous cases.
Quote mapping
Analysis of a representative sample (formats, sources, ambiguity level). 6 quote families identified covering 90% of volume, entry point for the POC.
Pilot family POC
N8N workflow + AI extraction on the largest family, ERP integration, fast validation UI for borderline cases. Measurement of real precision and time saved.
Extension to 6 families
Progressive rollout family by family, with prompt and fallback rule tuning. Structured anomaly reporting for continuous improvement.
Run and improvement
Full production cutover, quality monitoring, operators redeployed to higher-value tasks (control, customer relation, exceptions).
Operators freed from data entry, across 12 sites
Pipeline rolled out family by family, then moved to production across all sites.
« The goal was to free the teams from repetitive tasks and refocus them on their real job. That's what automation brings us. »
Want us to map your requests?
Free audit of your processes, we identify high-ROI use cases together, and quantify the gains before deploying anything.
More case studies
All case studiesInfraVia Capital Partners
↗Internal GPT
How InfraVia equipped 100+ staff with AI using Copilot and Perplexity Enterprise
Newfund
↗Internal GPT
How Newfund tooled its investment process with 6 specialised AI agents on Dust
Nomination
↗Ops
How Nomination runs millions of AI executions a year, from the edge to the core business