Agents

Agents that work on your catalog by your rules.

Every agent has a goal, a set of permissions and an activity log. It reads what it needs, proposes changes as drafts and waits for a person to approve.

Request a demo How it works

Prepare the French product sheets for the missing valves.

I've drafted 9 of the 14 sheets. The other 5 need the supplier's spec sheet, which isn't on record. I've flagged them for you.

9 translation draftsAgent · today
FR technical glossaryKnowledge base · Oct
5 sheets with no sourceWaiting on supplier

Illustrative example

Agent catalog

Start with these. Then build the ones your process needs.

Translate

Product sheets in your languages, with your glossary.

Translate the Valves family into Italian.

52 sheets in draft. I used the IT technical glossary and flagged 4 terms for native review.

IT technical glossaryKnowledge base
52 draftsAgent · today

Enrich

Completes attributes from supplier PDFs and sheets.

Generate

Titles, descriptions and SEO copy for each channel.

Always from verified attributes.

Predict

What to enrich first and which anomalies to watch.

Validate

Nothing goes live without passing the rules.

Required attributes per channel
Compliance and certifications
Minimum images and documents

Syndicate

Feeds for every channel, no manual templates.

How it works

Detect, propose, approve, publish, log.

01
Detect
Reads the catalog within its permissions.
02
Propose
Leaves the change as a draft.
03
Approve
A person with the right permissions.
Human approval
04
Publish
Each channel in its own format.
05
Log
What, who and when.

Example · an agent's permissions

What it can and can't do.

An agent is just another user: it has its own permissions and logs everything it does.

Reads product records, attributes and supplier PDFs.
Proposes translations and attributes, always as drafts.
Can't see purchase prices or margins.
Doesn't publish without a person's approval.

Request an assessment

Questions about the agents

Do agents publish on their own?

No. They leave changes as drafts, and a person with the right permissions approves them.

What can an agent see?

Only what its token allows: role, rows, columns and actions.

Can I use my own agents?

Yes. They connect through the same MCP server or the API, with their own permissions.

How are they priced?

Per agent and by scope, following an assessment.

Connect your first source and start asking your catalog.

In 30 minutes we work through a real business question using your own data. You leave with a defined scope and a fixed price.