Most SME businesses have a CRM in Exact Online that is barely used. Opportunities get entered, statuses set, and then… forgotten. Your sales team has no idea which deals are warm, your board has no visibility on the pipeline, and your forecast is guesswork.
Meanwhile all those opportunities are right there in your administration, with fields for expected value, close date and stage. With an AI assistant via MCP you query them in seconds — no dashboards to build, no Excel exports. This article: how to activate your sales pipeline with natural language.
What Exact Online stores per opportunity
Each opportunity (crm_opportunities_* tools) has at minimum:
- Customer or prospect;
- Expected value and currency;
- Close date (estimated);
- Pipeline stage (lead, qualification, quote, negotiation, won, lost);
- Owner (the responsible salesperson);
- Custom fields — often containing source, product line, etc.
Weekly pipeline review with AI
Turn the sales meeting from an Excel show into a data conversation. Questions to ask every Monday:
"Give all opportunities with close date within 30 days, ranked by value."
"Which opportunities have been in negotiation for 4 weeks already? What is their combined value?"
"Which opportunities had a close date last week but are still neither won nor lost?"
The AI fetches this directly from Exact Online and presents it cleanly. Stale deals get attention; forgotten deals get tidied up.
Conversion and weighted forecast
Not every opportunity converts. Each stage has a historical win rate. Ask your AI:
"Compute the weighted pipeline value for Q3: for each opportunity, multiply expected value by the win probability per stage (lead 10%, qualification 30%, quote 50%, negotiation 75%)."
For the first time you get a forecast that is more than addition. Calibrate the percentages on your own historical conversion data and you have an instrument leadership can actually use.
Source analysis: where do winning deals come from?
"Which lead sources produced the highest conversion to won deals in the past 12 months?"
The AI compares won versus lost deals per source (referral, website, event, cold outreach) and gives you a ranking. Marketing spend goes where it actually pays off.
Customer segments and cross-sell opportunities
"Which existing customers bought only product line A in 2025 but would be a fit for B based on similar customers?"
The AI reasons across patterns in your customer base. Combine with a targeted campaign and you open an upsell channel that was invisible without analysis.
Clean up stale leads
A messy pipeline gives a misleading picture. Periodic cleanup is essential:
"Which opportunities have been open for more than 6 months without activity and have no scheduled close date?"
Walk through them one by one. Mark them lost or update the close date. Your forecast becomes more reliable and your dashboard cleaner.
From opportunity to quote to order
A won opportunity wants a quote and ultimately a sales order. With your AI in one flow:
- "Create a quote for opportunity OPP-2026-0231 based on the product lines we discussed earlier."
- "Show me the PDF to review."
- "Customer signed off — convert quote Q-2026-0145 to a sales order."
Read our guide on creating quotes with AI for more on that step.
Getting started
The strength is not in any single question but in the routine. Define a weekly and monthly question set you ask consistently — just like you would draft an Excel report, but without the Excel. Check the prompt library for more CRM prompts, and follow the MCP setup guide to get started.