Online pharmacy sales in Spain rose by about 9% through April 2026, outpacing the 6.3% growth of the overall pharmaceutical market, according to IQVIA. A pace that is compelling the digital pharmacy channel to expand its catalogs faster than product pages can be written.
The generative artificial intelligence fuels that momentum, but in this sector it raises two risks. The first is that all pharmacies end up sounding alike, with dry, technical descriptions that are hard to distinguish. The second is that the model attributes properties to a product it cannot claim. European regulations on cosmetics and dietary supplements limit what can be asserted about them, and a over-the-counter para-pharmacy product without clinical backing cannot be presented as a treatment.
The work begins before the AI generates the text
Product information exists, but it is usually scattered across the manufacturer’s documentation, the catalog, and different teams, and it does not always reach the AI in full or with the necessary guidance. The problem starts here, because when something is missing, the model fills the gaps with its own interpretations. Thus properties the product does not have appear, or descriptions that do not sound like the brand. Preventing this requires defining, before generating the sheet, what is known about the product, what source backs each data point, what can be stated, what information is missing, and what should not be invented. This is the approach OnCatalog, the OnTranslation platform that generates product sheets with AI and adapts them to each market, has developed under the name “Product Truth.”
“In an online pharmacy, the product sheet has to sell by saying exactly what the product does, and nothing more. AI can write thousands of sheets, but someone first has to explain how the brand speaks, what can be asserted and what cannot. That pre-work separates scaling a catalog from scaling errors,” explains Óscar Nogueras, CEO of OnTranslation.
Chachifarma: scaling without losing its own voice
Chachifarma, a Spanish online pharmacy with a warm and direct tone, needed to expand its catalog without blending in with the rest of the sector. Before generating a single sheet, its team and OnCatalog spent several sessions defining that tone, the regulatory limits for each product type, and the structure of each sheet. They started with 30 products, and once the result was validated, rolled out to the rest of the catalog and published each batch in under 24 hours.
The key was in pre-publication control. Each sheet went through an automatic check of its structure and SEO optimization, and then a human review focused on the product’s composition, where an error carries bigger consequences. The pharmacy kept its tone across the entire catalog and, in two months, published 238 new sheets, 16 of which reached Google’s Top 10 without paid campaigns.
Farma2Go: speed to launch without compromising health
Farma2Go, a Cantabria-based online pharmacy, faced the opposite problem this summer. A single product, urgent and delicate: glasses to view the August solar eclipse. It began with only a product photo, and any error in the sheet about its use or warnings could harm the consumer.
Before writing, the OnCatalog team cross-checked the manufacturer documentation with the ISO 12312-2 standard for direct solar viewing filters. With that basis, they generated the sheet, reviewed it, and adapted it to five markets. Two hours later, the product was on sale in Portugal, France, Italy, Germany, and the United Kingdom. The Portuguese page reached the top 3 of Google and became the most visited page in the entire store. The product sold more than 4,000 units across all markets before the eclipse.
Both cases share the same premise: before asking AI to accelerate speed, you must ensure the product information is reliable. Defining what the brand can say, what it cannot say, and what must be adapted for each market allows faster publication without multiplying errors.
