This is the story of a mechanical manufacturing company that spent more time searching for information than making decisions. The production manager had become the company's "human router" — everyone asked him because he was the only one who knew where to find the data.
The company
- Sector: Precision mechanical manufacturing, contract production
- Employees: 45
- Location: Vicenza province, Italy
- ERP: Zucchetti Ad Hoc Revolution
The "before": the daily problem
Every Monday morning, the production manager arrived at 7:30 and started compiling the weekly report. He opened the ERP, navigated through job order screens, exported data to Excel, added weekly comments, formatted everything, and sent it to management.
Time: 45 minutes, every week.
But the report was just the tip of the iceberg. During the day:
- Sales reps called for job order status → he opened the ERP
- Management asked for revenue by client → he exported from ERP to Excel
- Logistics wanted to know what to ship → he checked the ERP then Outlook for confirmations
- Technical drawings were in network folders with no logic → he was the only one who knew where to look
All the information was there. But it lived in 4 different places: ERP, email, network folders, Excel files. And the intersection point was always the same person.
What we did
Phase 1: Audit (1 day)
We spent a day at the company, following our structured method. We observed the real workflow — not the theoretical one, the one that actually happens. We mapped systems, identified bottlenecks, and calculated time lost.
Audit result: 6 high-impact areas, with a potential savings of 35+ minutes/day.
Phase 2: Aitaky Brain integration (4 weeks)
We connected Aitaky Brain to:
- Zucchetti Ad Hoc (job orders, billing, inventory)
- Microsoft Outlook (customer and supplier emails)
- Network folders (technical drawings, specifications)
No modifications to the ERP. No data migration. Read-only connection — like all our integration solutions.
Phase 3: Automations
- Automatic weekly report: generated every Monday at 6:30 AM, sent via email
- Proactive alerts: late orders, missing materials, budget deviations
- Natural language queries: "Bertani job order status?", "Q1 revenue for top 10 clients?"
The "after": the numbers
| Metric | Before | After |
|---|---|---|
| Weekly report | 45 minutes | 30 seconds (auto-generated) |
| Job order status | 5-8 minutes (ERP navigation) | 10 seconds (natural language query) |
| Drawing search | 3-5 minutes | 15 seconds |
| Time saved/day | — | 35 minutes |
In one month: over 12 hours returned to the production manager.
KPI framework: measuring before and after
For companies evaluating a similar project, here are the key performance indicators we tracked:
- Time-to-information: How long it takes to get an answer to a business question
- Report generation time: From manual compilation to automated delivery
- Error rate: Manual data aggregation vs. AI-generated reports
- Decision latency: Time between question and informed decision
The quote
"It's not that I was working badly before. It's that I spent half my time searching for information instead of using it. Now I ask and I have the answer. The Monday report is in my inbox when I arrive. It's not magic — it's that someone finally connected the pieces."
— Production Manager
The lesson learned
The problem wasn't the technology. The ERP worked, email worked, folders worked. The problem was that information lived in 4 different places and the only way to bring it together was a person.
AI didn't replace anyone. It replaced the manual search-copy-paste-format work that nobody should be doing.
Want to find out what AI can do in your manufacturing company? Let's talk →
Related articles: