Business and Sales · News · 🇧🇷 Brazil
Manufacturer ZEN uses AI agents to tell in 10 minutes whether it can make a part
The Santa Catarina metalworking company built agents that run the pre-assessment of whether a part can be made on its presses, a step that took up to 17 days, and an engineer validates the result.
- Profession
- Engineer, Salesperson
- Tools
- Agentes de IA
- Source
- joinville360.com.br ↗
- Country
- 🇧🇷 Brazil
- Report date
- October 31, 2026
- Verified on
- November 3, 2026
See more: AI for engineers · AI for salespeople
What they did
ZEN, a metalworking company with more than six decades in business, built AI agents fed with the company's technical information, equipment documents and the knowledge its engineers have accumulated. When a new part comes in, the agents check whether it can be produced within the capabilities of the plant's presses, list the requirements and deliver an assessment along with the technical references used. The project is part of the Data & AI Hub the company started setting up in January 2026.
How it went
According to the company, the manufacturing feasibility pre-assessment, which took up to 17 days, now comes out in under 10 minutes. The case was presented by Paulo Mortari, director of Engineering, Excellence and Innovation, at VDI-Brasil's DEBA 2026, as a way to answer requests coming from the sales team faster.
Good to know
The AI doesn't decide on its own: an engineer validates every assessment before it is used. The numbers come from ZEN itself and were released by FIESC.
The prompt
I'm an engineer at a [type of product] plant and customers ask whether we can produce a part. Based on our equipment capabilities [paste spec sheets, force limits, dimensions and materials] and the part's drawing or specification [paste], tell me whether it looks feasible, which requirements need to be confirmed and what data is missing. Cite which document each limit came from and flag as open anything that isn't in the documents.
Replace what is in [brackets]. Prompt written by us from the account, not a literal copy of the source. Do not paste other people’s personal data (privacy laws such as LGPD and GDPR apply).
Summary in our own words. Results as stated in the original account; we have not tested or audited them. Read the source ↗ Found a mistake?
Similar cases
01An n8n agent researches trade show exhibitors and scores each one as a potential client
A consultant built a workflow that scrapes the exhibitor list, researches each company online and writes size, score and reasoning to a Google Sheet.
02Pre-qualify leads before they reach a salesperson
Futurepedia built an agent that researches each company, writes a fit summary and forwards the best leads.
03An agent makes dealerships compete on the price of a car
AJ Stuyvenberg let an agent research prices, request quotes and pass the best offer on to rival dealers.
04An automatic 6 a.m. briefing on who you're about to meet
A Zapier agent reads your calendar, researches external attendees and sends the summary to Slack.
05Automatic preparation for sales meetings
Annika Helendi used a template that triggers on events with "discovery" in the title and researches the contact.
06Researching 50 target companies and filling in the CRM
A Dutch consultant filters the list by hand, then the agent researches each company and creates records ready for outreach.
07A neighborhood auto shop reminds customers about quotes without becoming spam
An auto repair shop in Curitiba set up automatic WhatsApp quote follow-ups using AI.
08Automate the WhatsApp sales funnel with PIX payments
A developer automated customer service and billing on WhatsApp to stop losing customers to slow replies.