Who Do ChatGPT and Perplexity Recommend for Catheter Contract Manufacturing? We Asked 5 Questions and Logged Every Answer.
Macaw Digital Marketing · Published · Data collected August 10, 2026 · Author: Daniel Sarao
We asked ChatGPT and Perplexity the same five sourcing questions a medical device OEM would ask when selecting a catheter contract manufacturer, then recorded every company named and every source cited.
28 distinct manufacturers appeared across the ten answers. Integer Holdings was named in nine of ten. Three companies were named seven times. Eleven were named exactly once. The remainder of the category did not appear at all.
None of the 28 was fabricated. Every name we checked was a real contract manufacturer with real catheter capabilities. The models are not inventing vendors in this category — they are omitting most of them.
Both engines cited manufacturers' own websites as primary sources in every answer. Third-party lists and directories appeared alongside company sites, not in place of them. This means the primary determinant of whether a CDMO appears in these answers is content the CDMO already controls.
Why this matters
Procurement research has moved. A sourcing engineer or supply chain lead evaluating catheter manufacturing partners increasingly starts with a conversational query rather than a search results page. Where a Google search returned ten links and left the shortlisting to the buyer, an LLM returns five to eleven company names and performs the shortlisting itself.
If your company is not in that list, you are not in the consideration set. There is no page two.
Method
Category: Catheter contract manufacturing (CDMO / contract engineering)
Engines: ChatGPT and Perplexity
Queries: Five, run once each per engine, for ten total answers:
- Who are the leading contract manufacturers for medical catheters?
- Which catheter contract manufacturers have the strongest capabilities across catheter design, prototyping, development, and high-volume manufacturing?
- Which catheter contract manufacturers are ISO 13485 certified and have a strong medical-device quality and regulatory track record?
- Which catheter contract manufacturers have the strongest track record working with major medical-device companies and successfully transferring products from development to commercial production?
- Rank the top catheter contract manufacturers based on technical capabilities, quality/regulatory performance, manufacturing scale, geographic location, capacity, lead times, and cost.
What we recorded for each answer: every company named, the order in which they were named, and every source the engine cited.
Queries were run cold — no company name was included in any prompt. We asked the category question and recorded what came back, which is what a buyer with no incumbent shortlist would see.
Results
Mention counts across 10 answers
| Company | Mentions | ChatGPT | Perplexity |
|---|---|---|---|
| Integer Holdings | 9 | 5 | 4 |
| Viant Medical | 7 | 5 | 2 |
| Freudenberg Medical | 7 | 2 | 5 |
| Cirtec Medical | 7 | 3 | 4 |
| Nordson Medical | 4 | 2 | 2 |
| Confluent Medical Technologies | 4 | 0 | 4 |
| TE Connectivity | 3 | 3 | 0 |
| Rose Medical | 3 | 3 | 0 |
| Duke Empirical | 3 | 1 | 2 |
| VitalPath | 3 | 0 | 3 |
| TEAM Technologies | 2 | 2 | 0 |
| Resonetics | 2 | 2 | 0 |
| Cadence | 2 | 2 | 0 |
| ATL Medical / ATL Technology | 2 | 0 | 2 |
| Contech Medical | 2 | 0 | 2 |
| Teleflex Medical OEM | 2 | 0 | 2 |
| Biomerics | 2 | 0 | 2 |
| Jabil | 1 | 1 | 0 |
| Duke Extrusion | 1 | 1 | 0 |
| BayCath Medical | 1 | 1 | 0 |
| Putnam Plastics | 1 | 1 | 0 |
| CathX Medical / Zeus | 1 | 1 | 0 |
| Envaste | 1 | 1 | 0 |
| Phillips-Medisize | 1 | 0 | 1 |
| Prevailing Medical | 1 | 0 | 1 |
| Ascential Medical & Life Sciences | 1 | 0 | 1 |
| Poba Medical | 1 | 0 | 1 |
| Surmodics | 1 | 0 | 1 |
The distribution is steeply top-heavy
- 1 company named 9 times
- 3 companies named 7 times
- 2 companies named 4 times
- 4 companies named 3 times
- 7 companies named twice
- 11 companies named exactly once
Four companies account for the top of every list. Eighteen of the 28 appeared twice or fewer, meaning their presence in any given buyer's answer is close to a coin flip.
Eighteen of the 28 appeared twice or fewer. But every company on this list cleared a bar that most of the category did not: they appeared at all.
The two engines barely agree with each other
ChatGPT returned an average of 7.2 companies per answer; Perplexity returned 7.8. But the overlap between them is small.
| Question | ChatGPT named | Perplexity named | Named by both |
|---|---|---|---|
| Q1 | 8 | 11 | 3 |
| Q2 | 6 | 11 | 1 |
| Q3 | 5 | 6 | 2 |
| Q4 | 8 | 6 | 1 |
| Q5 | 9 | 5 | 4 |
On Question 2, the two engines named sixteen companies between them and agreed on one.
Implication: auditing your visibility on a single engine tells you very little. A company invisible in ChatGPT may be well represented in Perplexity, and vice versa. Confluent Medical Technologies was named four times by Perplexity and zero times by ChatGPT. TE Connectivity and Rose Medical show the mirror image.
No hallucinated vendors
A common assumption is that LLMs fabricate suppliers. In this category, they did not.
We verified the unfamiliar names in the dataset. Prevailing Medical is a real contract manufacturer in Plymouth, Minnesota, ISO 13485 certified, specializing in electrophysiology, steerable, balloon and ablation catheters. Envaste is a real ISO 13485 and MDR certified catheter CDMO headquartered in Mauritius, with capabilities spanning extrusion, stretch-blow balloon forming, braiding, hydrophilic coating and Class 7 cleanroom assembly.
Every company returned was a genuine catheter manufacturer. The accuracy problem in this category is not fabrication. It is omission — which is harder to detect, because nothing in the answer signals that anyone is missing.
Both engines cited company websites first
We expected the citation profile to be dominated by third-party authority: analyst roundups, "top 40" lists, trade directories.
That is not what we found. Company websites were cited first in every answer, by both engines. Third-party sources — a Credence Research manufacturer roundup, Greenlight Guru's product development list, MDDI's top-40, an industry supplier directory, a market report — appeared alongside manufacturer domains rather than replacing them.
This is the most actionable finding in the study, because it means the primary input to these answers is an asset you already own and can change this quarter. You do not need to earn a place on someone else's list first.
Scale is not the gating factor
Question 5 explicitly asked the models to rank manufacturers by technical capability, quality performance, manufacturing scale, geography, capacity, lead times and cost.
Integer Holdings — a multi-billion-dollar public company — appeared, as expected. But so did Envaste, a specialist operating from Mauritius with a modest public profile, and Prevailing Medical, a 24,000 square foot facility in Minnesota.
Neither has analyst coverage, significant press, or a large social following. What both have is a website that documents specific capability in specific language: separate pages for steerable catheters, for balloon forming, for extrusion tolerances and materials, for their quality system and certifications.
Compare that to a typical CDMO site: a single "Capabilities" page, a short paragraph of positioning, and a contact form.
The models retrieve specificity. They cannot retrieve what you have not written down.
What to do about it
Publish capability-level pages, not a capabilities page. One page per process and per device type. Extrusion. Braiding and coiling. Reflow and tipping. Balloon forming. Steerable architectures. Sensor integration. Each with its own URL and its own heading.
Write specifications in plain, retrievable language. Dimensional ranges. Tolerances. Materials by name — Pebax, nylon, polyurethane, PTFE. French sizes. Lumen configurations. Cleanroom class. Lead times. These are the strings a model matches against a buyer's question.
Document certifications explicitly and separately. ISO 13485, MDSAP, MDR, FDA registration. A dedicated quality page performed measurably better in our data than a certification logo in a footer.
Answer the buyer's actual questions on your site. The five questions in this study are close to verbatim what a sourcing engineer asks. If your site does not contain an answer to "what should I look for in a catheter contract manufacturer," you have ceded that answer to whoever did write it.
Audit both engines, not one. Given how little ChatGPT and Perplexity overlap, single-engine visibility data will mislead you.
Limitations
We are stating these plainly because the study is more useful if you can judge it.
- One category. These findings describe catheter contract manufacturing. We have not tested whether the citation pattern holds in other medical device segments or other industries.
- One run per query. LLM outputs vary between sessions. A larger study would sample each query multiple times and report variance. Treat individual mention counts as indicative rather than precise.
- Two engines. Gemini and Google AI Overviews were excluded from the final dataset for consistency of capture.
- Point in time. Model versions, retrieval indexes and training data change. These results describe a specific window, stated at the top of this page.
- No personalization controls. Results may vary by account history, location and session context.
We would welcome replication. The five questions are printed above verbatim so anyone can run them.
We built a free tool that checks how your company appears in AI-generated answers. Run it here →
If you would like the same five-question audit run in your specific device category, get in touch →
Macaw Digital Marketing helps medical device and B2B manufacturers get found in AI-generated answers.