Product & AI Strategy

The money went out.
The value hasn’t come back.

Usually an AI program. Sometimes a product line. Either way the first job is finding what’s worth keeping, and the second is making it earn something. Twenty-five years as the product person in medical devices, surgical robotics, and AI. Intuitive Surgical. Caresyntax. St. Jude Medical. Drive.ai.

Book 15 Minutes See the Work
21
Issued Patents
Devices, Robotics & Autonomy
8
Class III FDA Approvals
Portfolio I Owned
2,000+
Companies Evaluated
As a VC Partner
Featured Work

What I get brought in for.

Three cases. A platform built out of three acquired companies, a diligence practice with capital on the line, and a first-of-kind surgical robot with no user to ask. Different scales, the same job.

Caresyntax
2021–2024
VP Product · Surgical data & AI platform

Three acquired companies. One product anyone could sell.

Caresyntax had bought its way to a product line. Three separate companies, three roadmaps, three sets of customers, and a sales team trying to explain all of it at once. Meanwhile the data was already there: video, audio, images, device and operational feeds from thousands of operating rooms.

I owned the product decisions. What consolidated into the enterprise platform and what got retired. Which AI and automation work was worth building on top of the data, and which of it was a demo that would never survive a hospital deployment.

The hard part wasn’t the technology. It was getting three engineering groups and a sales organization to agree on one product.

Outcome
One enterprise platform running in 2,800 operating rooms, supporting 32,000 surgical teams, with sales expanded in the US and internationally.
This is the shape most growth-stage and portfolio-company work takes. Too many product lines, an AI effort with no clear return, and nobody with the standing to say what gets cut.
Mighty Capital
2024–2026
Product Partner · Diligence and AI operations

Two thousand companies, judged with capital on the line.

Product and technical diligence was the deciding input on investment. Is the product real, or a demo with a waitlist. Does the roadmap survive contact with the market it claims. Does the AI claim hold up, or is it a wrapper. What does this team actually have, and what will the path to market really cost.

Two thousand companies later, I know what those answers look like when they are true and what they look like when someone is hoping nobody checks.

I also built the firm’s internal AI operating system: sourcing, briefing, diligence support, and portfolio work, running in production and used daily by the partners.

Outcome
A diligence practice that compressed investment decisions to a one-month consensus, and 50+ production AI workflows a small team ran daily.
The same read works before a fund or a company commits. Technical and product diligence on a target, a partnership, or an internal program about to get funded.
Intuitive Surgical
2012–2015
Senior Product Manager · Flexible robot, stage-one lung cancer

The decision that determined which patient came first.

The platform was being built toward two possible customers. Not two variations of the same one. Two different users, with different workflows, different clinical paths, and different economics. Choosing wrong meant clinical trials pointed in the wrong direction and a platform that reached market late, or not at all.

Resolving it took original research, not a meeting. I trained under a thoracic surgeon and ran 400+ animal and cadaver studies as the stand-in user, because in a stealth program there was no one to ask. Then I went and got the evidence on both users: workflow, clinical path, and the economics of each.

The harder half was bringing the team through the data until they could see it themselves. A decision the team understands is the only kind that survives the next phase review.

Outcome
The platform reached market and became a significant revenue driver for Intuitive Surgical.
That call does not get easier at a larger company. It gets more expensive, because by the time anyone names it, the organization is already building toward both answers.
“The deliverable was rarely the product. The deliverable was the decision. What to build. What to leave on the table. What to do first. The rest follows from that.”
Michael Paris — Qnnovation
The Situations

Where I get called in.

Half the work is naming which of these you actually have. That part is usually free. The other half is going and getting the evidence that settles it.

01
Money spent on AI, nothing to show

Pilots ran. Licenses got bought. A vendor delivered something. None of it is in anyone’s daily work. The job is finding where AI actually fits how you operate, killing what doesn’t, and getting the rest used by people who never asked for it.

02
Two roadmaps, no evidence

R&D can build either one. Marketing wants both. What’s missing isn’t opinion, it’s the research that settles it. Who the user really is, what the clinical and economic path costs for each, and what it costs to be wrong. I go get it, then bring the team through it.

03
About to commit, nobody checked

A diligence read before the money moves, on a medtech or health AI target, a partnership, or an internal program about to get funded. Is the product real, does the AI claim hold up, what does the path to market actually cost. Two thousand companies evaluated with capital on the line.

04
Nobody senior owns product

The CEO built it and got pulled into fundraising. Or the CTO is covering it without being a product person. Roadmap goes to whoever asked most recently, and anything that needs a year of runway never gets started.

Product
Product strategy and roadmap. 0-to-1 product definition and MVP scoping. AI feasibility. Upstream marketing. User needs and design inputs in stealth programs. Phase gate. KOL development. Product organization design.
Robotics & Perception
Surgical robotics. Autonomous vehicles. Computer vision products. Translating perception into product requirements. Preclinical validation design. OR data and workflow. Medical imaging.
Clinical Development
Clinical evidence strategy. Clinical risk assessment. First-in-human. Preclinical study design. Working knowledge of design controls and the FDA and EU approval processes.
Therapeutic Areas
Surgical robotics. Thoracic oncology. Perioperative and OR workflow. Cardiac rhythm management. Heart failure. Structural heart and TAVR. Medical imaging. Neuromodulation. Sleep apnea. Behavioral health.
How This Works

Two shapes.

Both start the same way. Find the decision that’s actually stuck, resolve it with real work, and say plainly what to build, what to skip, and what has to be true for each.

01 — Device & Health Technology Companies

A Scoped Read on a Stalled Decision

One decision, a few weeks, fixed fee, defined deliverable. I do the work inside your program, talk to the people on both sides of it, and hand back a resolved call with the reasoning and the ranked list of what follows. You get an outside read from someone who has owned a Class III portfolio and a first-of-kind surgical robot, not a framework.

Start here
02 — Funded Startups

Fractional Chief Product & AI Officer

A fixed-term retainer for a company where nobody senior owns product. Usually the CEO built it and got pulled away, or the CTO is covering it without being a product person. Roadmap, prioritization, and the decisions that have been waiting on someone with the scar tissue to make them. Three-month initial term.

Start here
Michael Paris
About Michael

I build products out of technology
that doesn’t have a market yet.

Head of Product at Drive.ai when deep learning perception was a research result, not a product. Before that, product lead on Intuitive Surgical’s first flexible robot, where no user could be asked what they needed, so I trained under a thoracic surgeon and ran 400+ animal and cadaver studies to find out. Novel technology with thin data, going into a world that has to learn to accept it. That’s the problem I keep coming back to, and it’s the problem most AI initiatives actually are.

Underneath that, twenty-five years of medical devices as the product person, with 21 issued patents along the way. Implantable defibrillators at St. Jude Medical through eight Class III approvals. Transcatheter valves through first-in-human at two companies. Automated breast ultrasound at GE Healthcare. An OR video and data platform at Caresyntax running in 2,800 operating rooms.

Then two years as a Partner at Mighty Capital. Two thousand companies evaluated, and the internal AI systems the firm ran on, designed and deployed by me. At Products That Count, an AI talent-matching system that has run for eighteen months with a full-time operator hired to keep it going. That’s what I mean by AI credentials. Systems in production that someone else now owns.

The deliverable is rarely the product. It’s the decision. What to build. What to leave on the table. What to do first. The rest follows from that.

21
Issued patents in devices,
robotics & autonomy
8
Class III FDA approvals
on a portfolio I owned
michael@qnnovation.com LinkedIn
From People I’ve Worked With

The vouches.

A few recent recommendations from people I’ve built, shipped, and figured things out with. More on LinkedIn.

“
Michael has the wisdom to know that saying “no” to one-off feature requests can be just as important as saying “yes” to the initiatives conveyed in the roadmap.
Elizabeth Bacon
Founder & Principal — Devise Consulting
“
A creative powerhouse meets execution excellence. Michael is a force to be reckoned with.
Fabien Blanc-Paques
Product Lead, Gemini — Google
“
I can attest to the thoughtfulness Michael puts into everything he does and especially have been appreciative of his strategic thinking.
Victoria Gambardella
Director of Research, Generative AI — Meta
“
His approach and the way his mind works is highly creative. He gets things done and is so talented.
Carol Reiley
Co-founder & CEO — Drive.ai
See all recommendations on LinkedIn →
Selected Engagements

Where this work has happened.

Twenty-five years across medical devices, surgical robotics, autonomous vehicles, health AI, and venture capital. The featured cases above are examples, not the full list.

Get Started

Bring the decision
that keeps getting deferred.

Fifteen minutes. No deck, no pitch. You describe where it’s stuck and who owns which half of it. I tell you what I see. Pick a time that works.

Book 15 Minutes See the Work First