Kensai for CPG - Verified Product Data Your Models Can Trust

Supplements · Functional Foods^* · Functional Drinks^* · Skincare & Cosmetics^* · Personal Care^* · Baby Care & Toys^* *Coming soon Supplements · Functional Foods^* · Functional Drinks^* · Skincare & Cosmetics^* · Personal Care^* · Baby Care & Toys^* *Coming soon

Why this matters

The models you want to build need data you don't have yet.

You want to read consumers, predict what wins, and decide faster what to launch and where. All of that runs on product data, and today it's locked in labels, certificates, and supplier PDFs, unstructured and unverified.

Until it's clean and trustworthy, nothing you build on top of it can be either.

Half right

Is not good enough. Your predictions are only as good as their inputs. Models can't read consumers from data that's half right.

Once

Fixing it compounds. Structure it once, and every model, category, and market you move into builds on the same trusted base.

The problem

A claim isn't just text. A certificate isn't just a logo.

Raw product data can't be dumped straight into a model. The signals that matter carry context that gets lost:

A claim needs proof

"Clinically proven" means nothing unless the evidence and its source sit right behind it.

A certificate needs context

Who issued it, what it covers, and when it runs out.

A lab result needs a limit

A number only means something against the threshold it's measured against.

Proof goes stale

A product can sit unchanged while a rule shifts or a certificate lapses. By the time anyone notices, the packaging is printed and the inventory is shipped.

Miss that context and every model built on the data inherits the errors. This is the layer that has to be right first.

What we do

One clean, verified record for every product. Ready for your teams and your models.

Step 01

Collect

We pull in everything behind a product (labels, supplier docs, certificates, lab reports, claims, sourcing) and read the rules straight from the regulators (FDA, EFSA, MHRA and more).

Step 02

Verify

We tie every claim to its evidence and source, match ingredients to doses, check certificates, and validate it all against each market's rules.

Step 03

Deliver

Each product becomes one clear record: what it is, what it can claim, where it can sell, with a link back to every source, ready for your teams and your models.

Step 04

Keep current

When a rule changes or a certificate expires, we catch it and update the record, so the data your models run on is always current.

Who it's for

Built for the teams behind the product.

AI & Digital

Own what AI says about your products.

People ask chatbots what to buy, and the answers come from whatever product data they find. Give them a verified record, so what's said about your products is accurate and yours.

Marketing & Commercial

Know what you can say, before the campaign is built.

Know what each product can claim in each market, with the proof behind it, before the campaign is built.

Consumer Insights

Connect each product to how people actually buy.

One trusted base to connect each product to consumer behaviour and predict what will land.

Regulatory & Compliance

Spend your attention where it counts.

Spend your time on the products and markets that actually need a look.

Proof

Every side of the market is asking for this.

"We used an AI tool to rewrite our product descriptions and ended up with 4,000 items disapproved by Google."

CEO, supplements brand

"We stepped away from all paid media because of the challenges with claims and wording."

CEO, women's health brand

"The innovation process is still largely manual. Long feedback times, high pressure. I'd spend my budget to iterate much faster."

Product lead, multinational health retailer

"The know-how lives in different places, and no one connects it."

C-level, longevity brand

Design partnerships underway with health & wellness brands, retailers, and market-entry partners across the EU, UK, and US.

The outcomes

What changes when your product data is clean and verified.

Better predictions

Your models finally run on data you can trust, so what they tell you about what sells is worth acting on.

Faster to launch

Know where a product can sell and what it can claim on day one, not after the money's spent.

Less busywork

Teams and models start from one verified record instead of chasing PDFs and emails; weeks of back-and-forth become minutes.

Fewer surprises

Catch unsupported claims or expired certificates before they cost you a delisting or a reformulation.

Built by operators who lived the problem

We spent years wishing this existed.

Karina Valencia

Co-Founder

Karina has spent 18 years moving fast across borders under real pressure. 3× scale-up operator who directed market entries at Palantir in the run-up to IPO and managed an $11M CPG portfolio at P&G.

Dr. Leonie Bode