Your Trial Shows Up In AI. Now What?

By Ross Jackson, Ross Jackson Consulting

In part one of this series, I argued that clinical trial visibility is no longer simply a question of whether a study can be found online. As patients, physicians, site teams, sponsors, and CROs increasingly use AI tools to interpret clinical trial information, the more important question is whether AI-generated answers help the right stakeholder trust the trial, understand it, and take an appropriate next step.

That raises the follow-up question: If an AI visibility audit shows that a trial is being misunderstood, undersold, poorly sourced, or presented without a clear route forward, what should sponsors and their partners do about it?

The answer is not to game AI systems. In clinical research, that would be inappropriate and unlikely to work for long. The answer is to improve the information environment AI tools are drawing from — not only the sponsor’s trial page, but registry listings, patient-facing materials, site pages, CRO and recruitment partner content, referrer assets, publications, advocacy sources, and educational content that may shape how the study is found, interpreted, compared, and acted upon.

Start With What The Audit Shows

The most useful audit findings are rarely dramatic hallucinations or obvious factual errors. More often, the gaps are subtler.

In a recent audit, as referenced in part one of this series, the study did not appear to have a credibility problem. AI tools could generally explain the sponsor, asset, trial design, investigational status, and scientific rationale when prompted with the right names or source material. The issue was that the public information environment did not consistently help a patient, caregiver, or referrer move from awareness to reassurance, verification, and appropriate inquiry.

That is the difference between being visible and being actionable.

The audit found stronger performance for branded searches involving the study, asset, sponsor, or registry ID, but weaker performance for unbranded patient queries such as condition-specific trial searches or "new treatment study near me." It also found that actionability was the weakest area, largely because there was no clear, low-friction patient inquiry route, site finder, pre-screener, or expectation of what would happen after inquiry.

A trial may therefore be visible, scientifically credible, and technically explainable, yet still fail to provide the practical bridge that patients, physicians, and sites need.

Fix The Source Material First

When AI tools provide weak or incomplete answers about a trial, the problem often starts with the source material.

That source material may include a registry entry, press release, investigator deck, scientific poster, patient recruitment page, sponsor pipeline description, protocol synopsis, or investor presentation. The issue is not always absence. It is whether the available information is accessible, structured, current, and written for the questions different stakeholders are likely to ask.

A registry listing may contain the necessary scientific and operational detail but still be difficult for a patient to understand. A sponsor pipeline page may communicate asset value to investors but say little about eligibility, burden, geography, or referral pathways. A poster may contain useful evidence but not provide a clear patient or physician-facing explanation.

The first fix, therefore, is not necessarily creating more content but making existing information more useful.

Sponsors should review whether their public trial information clearly explains:

• who the trial is for
• what problem the study is trying to address
• what participation may involve
• where the study is available
• what is known and not yet known about the investigational treatment and the study
• how the trial compares with other relevant trials or standard of care options
• what patients, physicians, or sites should do next
• where the most authoritative current information can be found.

If the public source material is thin, fragmented, outdated, or written for the wrong audience, AI tools will either struggle to interpret it or rely on someone else’s explanation instead.

The relevant trial registry remains the foundation for core study facts. Public trial pages, recruitment materials, site content, and partner communications should therefore point back to the current registry record and use consistent descriptions of status, locations, eligibility, and contacts.

This also creates an ownership issue. Someone needs to know which source is authoritative, who is responsible for updating each public page, and how changes to recruitment status, site availability, eligibility, or study contacts are reflected across the wider information ecosystem. Otherwise, a sponsor may improve one page while an older or inconsistent source continues to shape the AI answer.

Make Important Information Easier To Retrieve And Cite

A common misconception is that AI tools cannot read PDFs, slide decks, posters, or other non-HTML content. Increasingly, they can.

But "can read" is not the same as "will reliably retrieve, cite, and interpret correctly."

A scientific poster, investor deck, or webinar may contain useful trial context, but I would not want those to be the main sources an AI tool relies on when trying to explain the study to a patient, physician, site, or referrer.

Sponsors should not assume that because trial-relevant information exists somewhere, AI systems will find it, prioritize it correctly, and present it to the right person in the right way.

A practical fix is to connect those assets to clear, current, structured web pages. That may include an HTML trial overview page, a plain English patient summary, a physician or referrer FAQ, a site-facing study overview, a page summarizing key publications or posters, transcripts and summaries for important video content, updated links to registry entries, and clearly marked dates showing when content was last reviewed or updated.

The aim is not to replace scientific publications, posters, or registry records. The aim is to make the most important information easier for both humans and AI systems to understand in context.

Structure Content For Extraction, Not Just Reading

Sponsors should also think about how trial information is structured on the page. AI tools may be able to interpret long, unstructured content, but they are more likely to retrieve and summarize information accurately when the page provides clear signals about what each section means.

Practical improvements may include:

• clear titles, headings, and plain English summaries
• structured eligibility, burden, location, status, and contact information
• FAQ sections answering patient, physician, site, and referrer questions
• internal links to registries, publications, site pages, and relevant audience pages
• transcripts/summaries for key video, poster, or deck content
• appropriate structured data where relevant, secondary to clear human-readable information.

These elements are not a substitute for accurate content. A poorly written page does not become useful because it has schema markup. But clear structure helps both humans and AI systems understand what the content is about, how current it is, which audience it serves, and how it connects to the wider evidence trail.

Once the source material is clearer and easier to interpret, the next question is whether that visibility leads a patient, physician, site, or sponsor to take an appropriate next step. More on that in the final article

About The Author:

Ross Jackson is a patient recruitment specialist and author of the books The Patient Recruitment Conundrum and Patient Recruitment for Clinical Trials using Facebook Ads.

Having started out with digital marketing in 1998, Ross quickly developed a specialty in the healthcare niche, evolving into a focus on clinical trials and the problems of patient recruitment and retention.

Over the years Ross branched out from the purely digital and now operates in an advisory capacity helping sponsors, CROs, sites, solutions providers, and others in the industry to improve their patient recruitment and retention capabilities — having advised and consulted on over 100 successful projects.