Campaign launch time fell about 75%, with 15% more conversions than the ads already running.
Otrium Ad Creative Agent
Agent Ready Award Agentic Achievement Award
Matalan Product Description Generation
100 product descriptions in 30 minutes, against 100 a day by hand.
An agent reads the product image, fills the gaps in the metadata, and writes copy in Matalan's own voice.
What the agent does
Reads the product photograph, fills the gaps in thin metadata, and writes the description in the retailer's own copy style.
01
The problem
Why this exists
Product copy is the bottleneck between a product existing and a product being findable. A retailer's copywriting team produces around 100 descriptions a day, and the catalog moves faster than that.
Read more
The usual workaround is templating from structured data fields, but the fields are sparse. What is actually distinctive about a garment is visible in the photograph and absent from the metadata, so a templated description says nothing a shopper or a search engine can use.
02
Outcomes
What changes when this is in place
- 100 descriptions in 30 minutes, against 100 per day from the copywriting team.
- Delivered and fully integrated within 4 weeks.
- Enhanced SEO and deeper customer engagement, with copywriter capacity freed for strategic work rather than volume.
Read more
- Copy is stylistically coherent with Matalan's own writing rather than generically correct, which is the difference between usable and merely present.
03
How it works
Architecture
Gemini analyzes the product image to extract visual features that the metadata does not record. Those insights are combined with the sparse structured product data that does exist, and the model generates copy that mimics the style of Matalan's own copywriters through prompt engineering.
Read more
The system runs on Google Cloud using Gemini and Vertex AI.
Valtech draws the distinction against both plain automation and predictive ML. Template-filling would produce a description from structured fields alone. A predictive model might flag which products need attention. This reasons over unstructured visual input, synthesizes it with sparse metadata, and generates novel copy, which is a generative act rather than a rule or a pattern match.
Open standards used
REST APIs
Tech stack
Google GeminiGoogle Vertex AIGoogle Cloud
Resources
Next step
Request an intro
Valtech is behind this solution. Ask about scaling product copy in your voice, and the MACH Alliance will put you in touch with the person who can answer.
Not ready to talk to anyone? MACH Insider is the Alliance mailing list. New solutions, program openings, and the research behind them, once a month.
Everyone credited on this solution
- ValtechSystems integrator
Every organization named here has a page listing the rest of its work on the Hub.
Related solutions
A shopping assistant that lifted conversion 25%, taken from a split test to all traffic.
Ask Frasers Shopping Assistant
Shoppers describe the room they are furnishing and get the right fixture back.
Lightopia Ask Cleo
Agent Ready Award Agentic Achievement Award
Verification
Published case study
The detail on this page is drawn from a case study published by the contributor or the MACH Alliance. Any figures on this page are the ones that case study states.
Where this came from
Solution detail is published from the contributor's 2026 Agent Ready submission to the MACH Alliance. Submitter names and contact details are not published.
Status: Live in productionAdded Jul 8, 2026A listing is an editorial decision, not a MACH Alliance certification.
Built something worth showing?
Put your agentic solution in front of the enterprise leaders who use the hub to decide what to build next. Submissions are open to MACH Alliance Members, Collaborators, or Ambassadors.