AI has entered pharma naming. Now the real work begins.

A recent EMA acceptance of an AI-supported pharmaceutical brand name is not the end of human naming expertise. It is proof of why that expertise matters.

In April 2026, Brand Institute announced that the European Medicines Agency had approved a pharmaceutical product brand name developed with the support of its proprietary AI-powered naming platform, Brandi™. The name itself will only be disclosed after approval of the corresponding medicinal product. At first glance, this sounds like a simple technology story: AI creates name, regulator accepts name, the future has arrived.

Which is remarkable in itself, given that there are many ways to create names, ranging from individual name creation and discussion groups to name-creation software. This is because the method used to create a name is, in fact, secondary to the ultimate goal in pharma: a name that meets all the criteria and is approved by the relevant regulatory authority.

But for anyone who works seriously in pharmaceutical naming, the more interesting conclusion is different.

AI has not made pharma naming easy. It has made visible just how complex pharma naming already was.

At Readge Pharma Naming, we see AI not as a replacement for naming expertise, but as a useful acceleration layer. It can generate more options, explore more linguistic patterns, compare more data points and help teams move faster through the early creative landscape. But the real challenge in pharma naming has never been a shortage of possible syllables.

The challenge is finding names that are strategically relevant, legally available, linguistically robust, medically safe, internationally usable and acceptable to regulators.

That is a much narrower universe.

The problem was never creativity alone

In consumer branding, a name can often be judged by memorability, distinctiveness, emotional appeal and trademark availability. Those factors matter in pharma too, but they are only the beginning.

A pharmaceutical brand name must function inside a highly sensitive medication-use system. It may be spoken by physicians, read by pharmacists, typed by nurses, searched by patients, scanned on packaging, discussed in multiple languages and used under pressure. A small similarity can become a serious risk. A pleasant sound can become a misleading therapeutic suggestion. A clever construction can conflict with an INN stem, an existing product, a medical term or a regulatory expectation.

The EMA’s guidance on invented names looks at issues such as potential confusion with existing medicinal products, misleading therapeutic or pharmaceutical connotations, promotional implications, INN-related concerns and inappropriate meanings across EU/EEA languages.

FDA guidance similarly treats proprietary name review as a safety issue, assessing look-alike and sound-alike risks, orthographic and phonological similarity, USAN stem conflicts and possible medication errors.

In other words: a pharma name is not just a brand asset. It is part of the safety architecture of a medicine.

What AI can change

AI can be genuinely useful in this environment.

It can rapidly generate broad territories of name candidates. It can test different phonetic routes. It can help identify clusters of similarity. It can support early-stage screening against defined naming rules. It can help teams avoid spending too much time on candidates that are obviously weak, crowded or directionally risky.

That matters, because the pressure on naming teams is increasing. More products, more indications, more biologics, more devices, more combination therapies, more global launch strategies and more regulatory scrutiny all make the naming space tighter. The “good” names are not simply waiting on the shelf.

AI can help create more starting points.

But more starting points are not the same as better decisions.

The risk with AI is that it can produce names that look plausible before they are viable. It can give an impression of completeness. It can create artificial confidence. A generated name may sound polished, modern and pharmaceutical — yet still be unusable because of a hidden regulatory risk, an unfortunate association in another language, a similarity to an existing product, a problematic stem, a weak trademark position or a strategic mismatch with the product profile.

In pharma naming, plausibility is cheap. Defensibility is expensive.

Human judgment becomes more important, not less

The arrival of AI does not remove the need for human expertise. It raises the standard for it.

The value of experienced naming consultants is not simply that they can “come up with names”. It is that they know what to do with names once they appear. They can interpret risk. They can challenge a seemingly attractive candidate. They can balance commercial ambition with regulatory realism. They can understand when a name is distinctive but too strange, safe but too generic, meaningful but too suggestive, elegant but too close to something already on the market.

This is where Readge pharma naming believes the next phase of pharma naming will develop: not as automated naming, but as name engineering.

Name engineering combines creativity, linguistic discipline, trademark strategy, regulatory awareness, medical safety thinking and commercial positioning. AI can support parts of that process. It can widen the funnel. It can improve speed. It can help organise complexity.

But it cannot take responsibility for the final judgment.

A pharmaceutical name is not approved because it exists. It is approved because it survives.

The future is not machine versus human

The interesting question is not whether AI will be used in pharma naming. It will be. The more relevant question is how intelligently it will be used.

Used poorly, AI will flood teams with attractive but fragile candidates.Used well, it will help experts ask better questions earlier.

Does this name create a possible medication-error risk?Could it be confused in handwriting, speech or packaging?Does it suggest an efficacy claim?Does it sit too close to the INN or USAN stem system?Will it travel across European languages?Can it survive trademark screening?Does it support the product’s strategic position without overpromising?

These are not purely technical questions. They require judgment, context and experience.

AI can assist the process. It cannot own the consequence.

A new chapter for pharma naming?

The recent EMA acceptance of an AI-supported name is a meaningful signal for the industry. It shows that AI can participate in a process that is highly regulated, safety-conscious and commercially important. But it should not be read as the moment when machines took over naming.

It should be read as the moment when pharma naming became even more explicitly hybrid.

The winning model will not be AI alone.Nor will it be traditional creativity pretending nothing has changed.

The winning model will be expert-led, AI-supported, regulation-aware and globally disciplined.

At Readge Pharma Naming Consultants, we welcome this development. Not because it makes our work simpler, but because it confirms what we have always believed: a good pharmaceutical name is never just a word. It is a carefully built decision.

And in the age of AI, careful decisions matter more than ever.

Michael Dijkstra Taurel / Readge pharma naming (www.readge.com)