In AI answers about pharma, trust is contested and Pfizer, J&J and Merck recur

The most common move in this snapshot is a refusal to rank. On trust, pricing and litigation, several answers opened by saying no single list exists and that the result depends on the metric. Only then did they offer names. The answers described the industry as a whole as distrusted, with pricing given as a main reason.

Each question produced its own cast. Trust and patient access pulled in reputation and access indices, with Novartis, Roche and GSK named as leaders in several answers. Pricing pulled in scandal cases such as Turing, Valeant and Mylan. Litigation pulled in Purdue, Johnson & Johnson, GSK and Pfizer, with settlements and opioid cases as the frame. The insurer question centred on Kaiser Permanente.

Pfizer, Johnson & Johnson and Merck are the large drug makers that turn up across questions. They are placed inside other people's rankings and caveats, not presented as category winners.

Pfizer was named in 15 of 25 answers. Johnson & Johnson and Merck were each named in 13, AbbVie in 8 and Eli Lilly in 6. Pfizer was tagged most often under Price & Value and Safety & Compliance. Johnson & Johnson and Merck were tagged most often as Not Substantive, meaning the mention did not carry a specific reason. Pfizer appeared in the pricing answers from ChatGPT, Claude, Gemini, Perplexity and Grok, and in a Perplexity answer that described it as ranked last in one public-reputation ranking.

Coverage differs sharply by assistant. Claude named Johnson & Johnson in none of its five answers. Grok named Eli Lilly in four of its five, while every other assistant named it in at most one. Eli Lilly's presence in this snapshot leans heavily on one assistant.

UnitedHealth Group was named in 4 answers, mainly in the insurer and patient access questions, and was tagged for Customer Experience. In the insurer question, Claude's answer said UnitedHealthcare had dropped to around 12th place in one ranking it cited. Kaiser Permanente carried the insurer answers from the assistants that named plans. CVS Health and Moderna were named in none of the 25 answers.

In the PeakMetrics news index over 30 days, Merck had the most mentions (20,326), followed by Eli Lilly (19,280), Pfizer (18,663) and Johnson & Johnson (15,149). AbbVie had 11,036, Moderna 10,024, UnitedHealth Group 6,182 and CVS Health 5,895.

The recent headlines are mostly investor and market material. They include stock-price moves, holdings disclosures, earnings comparisons between drug makers and analyst targets. Product and pipeline news appears too, such as a Merck EU opinion for Keytruda and AbbVie data on Vraylar and atogepant. Legal items appear at the edges: a headline on Pfizer and 340B protocols, one on AbbVie's Botox classification dispute, one on GLP-1 makers being sued over vision-loss claims, and one asking whether CVS insiders breached fiduciary duties.

The assistants answered questions about trust, price, access and litigation. The headlines mostly concern shares, earnings and deals. Moderna and CVS Health have thousands of monthly mentions and headlines about their stock and shareholders, yet neither was named in any answer.

Eli Lilly is covered at the level of Pfizer and Merck but was named in far fewer answers, and mostly by Grok. Pfizer's larger presence in the answers ties to reputation, pricing and litigation history, not to the recent headlines sampled here.

The names the answers credited as trust and access leaders, including Novartis, Roche and GSK, sit outside the tracked set, so press volume for the tracked brands says little about who the assistants foreground on those questions.

Press volume and AI visibility measure different things. The heaviest-covered brands here were not the most named, and two heavily covered brands were not named at all. Measure both.

The cited domains skew to trade and reference sources. In this snapshot fiercepharma.com was cited 5 times, accesstomedicinefoundation.org and pharmaceutical-technology.com 4 times each, and the source mix was 68% Trade and 32% Other. Third-party assessments such as the Access to Medicine Index, patient-group reputation surveys and settlement compilations feed the answers directly, so those are the documents to monitor and correct.

Because the assistants tend to say a question has no single answer, a brand often appears in a list that cites someone else's ranking. Check which ranking is doing the work in each answer and whether your brand appears in it.

This is a single snapshot of five assistants on 2026-09-28, covering five reputational questions. Live answers vary between runs, so treat it as a baseline, not a ranking.