How an Industrial Products Company Put 88% of Its Supplier Email Work on Autopilot

Client: Industrial products importer & distributor (name withheld)

Industry: Industrial products import & distribution

Market: Israel

Solution: AI Agents + System Integration

At a Glance

The Situation

The client is a company that imports industrial products and distributes them across Israel. Day to day, the business runs on email — suppliers send a constant stream of price changes, stock updates, new-product specs, contracts and general notices. A large team spent every day reading those emails and acting on each one by hand.

The Problem

Every email had to be read, understood, and sorted before anyone could act on it. A price or stock update meant finding each item in the system and editing it by hand. A new product meant building a record field by field. And the most expensive failures were the quiet ones: when a supplier discontinued or stopped shipping a part, nobody noticed — until a customer order fell through. The team learned the item was gone only after it had already cost them a sale. Multiply that across hundreds of emails a day and a whole team, and a huge share of the company's capacity went to nothing more than keeping the inbox from falling behind.

Why the Usual Fixes Didn't Work

For several years, the company tried to fix this with traditional, rule-based automation — no AI. It never held. Every supplier sends documents in its own structure and format: some as PDFs, some in the email body, some as spreadsheets, some scanned. The moment a supplier tweaked a template or a new supplier came on board, the rules broke. And rules can't make a judgment call — telling a price update apart from a discontinuation notice, or knowing which of dozens of fields actually matters, takes understanding, not pattern-matching.

And an off-the-shelf AI tool wouldn't have solved it either. Such a tool can summarize an email, but it can't find the right item in your ERP, update the price, build a new product record, or act across your live systems correctly, every time. The real work was never reading the email — it was doing something precise inside the systems the business runs on, for every document type. That's where both approaches stopped.

What We Built

We built a set of AI agents that read every supplier email, classify it by type, and route it to the agent that specializes in that exact kind of work. One agent handles price and stock updates. Another builds new products from the supplier's specs. Another watches for discontinued or out-of-stock items and raises them before they turn into failed orders. Each agent learned the job from the people who used to do it by hand, and connects directly to the client's ERP and data sources to make the change right inside the systems themselves. When an agent isn't confident, it doesn't guess — it raises a flag and a team member steps in. Over time, the flags got rarer.

The Result

• 88% of incoming supplier emails are now handled end-to-end by AI agents, with no manual work.
• Discontinued and out-of-stock items are caught the moment the supplier's email arrives — not after a customer order has already fallen through.
• Price, stock and new-product changes land straight in the company's systems, applied consistently instead of one manual edit at a time.
• The team that used to spend the day inside the inbox was freed for higher-value work.

Instead of chasing the inbox all day, the team now spends its time on work that actually moves the business forward.

Ready to Build Something Similar?

If your team spends its day processing emails, updating records, and firefighting the ones that slip through, the same approach can work for you. Let's talk about what it would take.