--- title: "Custom sales cockpit connected to the ERP · Client case · Quentin Andrieux" description: "After two abandoned CRMs, a sales cockpit connected to the existing ERP: targets by brand, season plan built in blocks, AI projection. Anonymised client case." url: https://quentinandrieux.com/en/client-cases/sales-cockpit-erp/ lang: en author: Quentin Andrieux updated: 2026-11-11 --- > Version texte de https://quentinandrieux.com/en/client-cases/sales-cockpit-erp/ (the HTML page is the reference). Author : Quentin Andrieux, https://quentinandrieux.com/ [Home](https://quentinandrieux.com/en/) [Client cases](https://quentinandrieux.com/en/client-cases/) Sales cockpit Events and media group · international sales team # Two abandoned CRMs, one cockpit on the existing ERP. An events group whose salespeople work remotely, across several countries. The tool reads their in-house ERP, adds no data entry, and gives each salesperson their season plan and signed figures. **1 h** between the ERP and the salesperson's screen: signed figures refresh every hour **4 weeks** to build the production tool, after two months of prototypes **0** extra data entry: the tool reads the ERP, read-only The flow 1. **In-house ERP** Read-only, signed figures every hour. 2. **Targets** By brand: target, signed, planned, still to find. 3. **Plan in blocks** Who, which offer, which actions, where, when. 4. **AI projection** Revenue, customers, hours, confidence level. 5. **Management** Plan sent, written and dated reply. From the existing system to a shared season plan. ## The context The group runs trade fairs in several countries. Its salespeople work remotely, each on their own portfolio. A first off-the-shelf CRM had been judged too heavy and dropped, and a second one never took either. The internal IT team had ended up building an in-house ERP. The audit found, as reported by the team: only a small share of emails attached to customer records, untracked calls, activities typed by hand, and management unable to follow the time and projections of a scattered sales team. Season plans lived in one spreadsheet per person, with no projection attached, and management feedback went by email. ## What was built A personal cockpit for each salesperson, connected read-only to a copy of the ERP database. Each person sees only their own targets, portfolio and plan; managers see their team. - **Targets**: by brand, the season target, signed, planned and still to find, in dollars or euros. - **Season plan in blocks**: the salesperson describes who they target, with which offer, through which actions, where and when. Only one field per block is mandatory. - **AI projection**: for each block, expected revenue, new customers, conversations and hours needed, with a confidence level and three notes (how I get there, watch out, worth a look). A known account is valued on its last signed season, the rest on average basket sizes and the person's conversion rates. - **AI assistant**: it knows the salesperson's plan, targets and portfolio, and answers questions such as "is this plan realistic given last season?". [More on the assistant](https://quentinandrieux.com/en/client-cases/ai-assistant-sales-plan/). - **Sync**: signed figures come up from the ERP every hour, with an alert after three hours without success. Cockpit · season targets *Fictional example* #### Targets by brand Brand A 72% Brand B 48% Brand C 31% #### Block · Southern Europe fairs Projected revenue **€38k** New customers **6** Hours needed **54 h** Confidence **medium** Watch out: six cities in one week will not hold. Fictional data. ## How Discovery in April, audit and presentation in May, a working day on site with management in July. Two months of successive prototypes on real data, then a six-point definition of "delivered", approved by the client before the first line of production code. Four weeks of build, acceptance testing with twenty pieces of feedback handled, a feedback session with two salespeople and the sales manager, and a one-hour training for the whole sales team. ## What changed The season plan, targets and signed figures sit in one place, up to date without data entry. The salespeople picked the season view as the right level to agree with management, and the AI comments as the most useful part: on one plan, for instance, it points out that six cities in one week cannot work. Their next request came from using it: the plan says where to go, it does not say whom to call. It led to a [prospect list generation module](https://quentinandrieux.com/en/client-cases/from-sales-plan-to-prospect-list/). The client's IT team then took over development in-house. ## Further reading - [The AI assistant connected to the salesperson's data](https://quentinandrieux.com/en/client-cases/ai-assistant-sales-plan/) - [From the sales plan to an enriched prospect list](https://quentinandrieux.com/en/client-cases/from-sales-plan-to-prospect-list/)