Your R&D team did not join to retype supplier data sheets. Your quality team did not join to hunt through folders for last year’s specification.
Chef AI does that part. It is the assistant built into beCPG PLM: you ask in plain language, it works on your own product data with your permissions, and it proposes. You decide.
Ask from the product, project or document you are working on. Chef AI calls the beCPG tools it needs and shows which one it used.
Answers come in the language you choose.
Drop the PDF: Chef AI reads the values and pre-fills the raw material. You check and confirm.
“Which finished products contain palm oil?” Answered across your whole portfolio, without building a query.
Names, descriptions, marketing messages, translations, organoleptic profiles: proposed, then edited by the person who knows.
Spelling, missing fields, allergens to declare, consistency of the label: flagged in the Suggestions panel of the product.
Chef AI compares an artwork with the specification and lists the differences to review.
Cost breakdowns, composition trees, a cheaper recipe under constraints: asked in one sentence, calculated by the beCPG engines.
On every product and document, the Suggestions panel lists what Chef AI proposes: values read in a document, fields to complete, translations, allergens, a review of the label.
You accept them one by one or all at once. Nothing is saved in the PLM until you confirm it, and an accepted change is tracked like any other change in beCPG.
It does not decide what goes on your label.
Nutritional values, Nutri-Score, ingredient and allergen labeling, CLP classification, cost optimization: the beCPG engines compute them, the same way they have for years. Same inputs, same result, every time, including six months later in front of an auditor.
If an AI answer and a calculated value differ, the calculated value prevails.
In the new supplier portal, the AI reads the technical sheet a supplier uploads, proposes the values and the allergens, cross-contamination included, and points out what changed since the previous version.
The supplier applies each proposal. Your quality team reviews the sheet as usual.
Claude, Gemini or the agent your IT team is building can work on beCPG data through the MCP server, with the permissions of each user.
Which calculations involve AI, where your data goes, who approves each change and how it is traced.
A 30-minute demo on your own products is usually enough to know whether it fits your process.
Chef AI is based on Google Gemini. Your instance can also be configured to use an AI model that you host yourself. Chef AI works on your beCPG data with the permissions of the user who asks.
Chef AI is an optional module of beCPG, available with version 26.10. It is enabled at your request.
No. In beCPG, every creation or change proposed by Chef AI is a suggestion: nothing is saved until a user confirms it, and accepted changes are tracked like any other change.
No. Nutrition, Nutri-Score, labeling, CLP and costs are calculated by the beCPG engines from your formulas. Chef AI helps you find, read, draft and check.
Yes. The AI module can be installed with an on-premise beCPG. It connects to the AI service: Google Gemini, or a model that you host yourself.
No. beCPG does not use your data to train AI models.
Another question about AI in your PLM? A PLM expert answers you, with a demo on your own products.
Bring a supplier data sheet and a product: we will show you what Chef AI proposes.
© 2026 beCPG.
