Generic brand prompts fall apart on a biotech company. Ask ChatGPT for a customer persona and you get a 34 year old marketing manager named Sarah. Your buyer is a VP of Translational Research who reads three papers before breakfast and has sat through 200 vendor pitches that all said “innovative”.
We shared our favorite GPT prompts for brand development a while back and it became the most read post on this site. This is the version for science companies. Same disciplines, rewritten for the people who build and buy in life sciences, medtech and research tools.
The same rule applies, only more so
Everything the model gives you is a draft to verify. In biotech the stakes on that are higher. A prompt can hand you a persona that cites a market size from 2019 or a competitor that was acquired last year. Check every claim. Treat the output as a smart intern’s first pass, because that is what it is.
And load context before you prompt. The model needs your pipeline or product stage, your modality or platform, who pays and who uses, and your three closest competitors. Paste that in first, every time. Better still, build it into a project or custom GPT once (we wrote up how to do that).
Prompt 1: Translate the science into a position
The hardest brand problem in biotech is the gap between what the science does and why anyone should care. This prompt forces the translation.
Here is a plain description of our technology: [paste two paragraphs]. Our buyers are [role] at [type of organization]. Write five one-sentence positioning statements, each built on a different benefit: speed, cost, accuracy, risk reduction and something a competitor cannot claim. Then tell me which one a skeptical scientist would believe and which one a skeptical investor would believe, and why they differ.
The last line is the useful part. Scientists and investors want different proof, and most biotech brands pick one and lose the other.
Prompt 2: Pressure-test the name
Biotech naming has traps that consumer naming does not. Names that sound like existing drugs. Names that are a gene symbol. Names that are fine in English and a problem in a market you plan to enter.
Evaluate the name [Name] for a [stage] [type of company] working on [area]. List: 1) existing drug names, gene symbols, or life science companies it could be confused with; 2) how it sounds when spoken on a conference call and whether it needs spelling out; 3) any negative meaning in German, French, Japanese and Mandarin; 4) whether it reads as a platform company, a therapeutics company or a tools company. Finish with the single biggest risk of using this name.
This is not a trademark search. Run the real search after. The prompt catches the embarrassing collisions before you pay a lawyer to find them.
Prompt 3: Build the two personas that matter
You usually have an end user and an economic buyer, and they are rarely the same person. Make the model build both and show the tension.
Create two personas for [product]. Persona A is the scientist or clinician who uses it day to day. Persona B is the person who signs off on the budget. For each: title, what they read, what a bad week looks like, what they have been burned by before, and the one sentence they would need to hear to take a meeting. Then list three places where A and B want different things from us.
The “burned by before” line is where the real messaging lives. Every lab has a freezer full of kits that did not work as advertised.
Prompt 4: Audit the competitors’ visual identity
Most biotech brands look the same: blue, a double helix or a dot pattern, a stock photo of a pipette. The model cannot see your competitors’ sites, but it can structure the audit so you see them properly.
I am going to describe the websites of three competitors: [Company 1: colors, typography, imagery, hero headline]. [Company 2: same]. [Company 3: same]. Identify what all three share. Then propose three visual directions that would stand apart from all of them while still reading as credible to a scientific audience. For each direction, name a brand outside life sciences that does it well.
Asking for a reference outside the sector is how you escape the blue helix.
Prompt 5: Find the brand story investors will repeat
An investor who likes you will describe your company to a partner in two sentences. You want to write those sentences before they do.
Here is our origin: [founder background, the problem they saw, why now]. Write the two sentence version of this story that a VC would repeat to their partners on Monday morning. Then write the version a postdoc would tell a labmate. Then tell me what is missing from our origin story that would make either version stronger.
Prompt 6: Stress-test the message against the room
Before a launch or a raise, run your core message through the objections.
Our core message is: [paste]. Respond as five people in turn: a skeptical KOL, a Series B investor who has seen three companies in this space, a procurement manager, a competitor’s head of marketing, and a science journalist. Each gives their first reaction in two sentences and the question they would ask next.
What the prompts will not do
They will not tell you whether your platform is a brand or a feature. They will not resolve the fight between the CSO who wants the mechanism on the homepage and the CEO who wants the outcome. And they will not design anything. That is strategy and craft, and it is where a brand partner who knows the sector earns their fee.
We have done this work for companies that have raised over $950M combined, including the brand and website behind Ketryx’s $39M raise. If you are heading into a launch or a round and the brand is not ready, start a project. If you want the sector-agnostic prompts too, they are in the original guide.
Want the short version? The 12-prompt brand library has the best of these plus the sector-agnostic set.