From the Diary of Doug Nissinoff

What the founder of Intelligence Ventures and IV Accelerator asks an AI healthcare company in the first meeting, and what a founder with real data underneath the product gets in return.

“A pitch says the model is 94 percent accurate and a generalist writes down 94,” Doug Nissinoff says. “I want to know 94 percent of what. Retrospective or prospective, how many sites, who labeled the data, and whether the 6 percent it misses are the cases a clinician would lose sleep over. The answer to that one question usually tells me more than the rest of the deck.”

The question isn’t only his. A 2024 analysis in Nature Medicine went through 521 AI medical devices authorized by the FDA and found that 226 of them, about 43 percent, had no published clinical validation data.

Doug is the founder and general partner of Intelligence Ventures, a pre-seed and seed venture fund that invests in AI healthcare companies in the US and Canada, and the founder and CEO of IV Accelerator, the fund’s ten-week program, which invests $25,000 in each company it accepts. The fund has screened more than 900 AI healthcare startups since it started. For the accelerator’s first cohort, more than 100 applications came in during the first two weeks. The fund interviewed 27 companies and made nine offers.

Applications for the second cohort open in October. The founder he wants reading this, he says, “has real clinical or biological data underneath the product and either has not heard of us or is not sure a small fund in Florida is worth an application.”

Leader in Spotlight: Read Adam Dowd Feature in Executives Diary

Where the Model Came From

The pitch he sees most often is “a clinical workflow product with ‘AI-powered’ in the first line of the deck.” There’s money behind that first line: Rock Health counted 54 percent of US digital health venture funding in 2025 going to AI-enabled companies, up from 37 percent the year before.

“The demo is good,” Doug says. “Then I ask where the model came from, and the answer is an off-the-shelf language model with a prompt and some rules on top. The company owns none of the data going in, and under the customer contract none of the data coming out either. Nothing about the product got better because a hundred clinics used it last month. A competitor with the same API key can ship the same thing by Friday.”

“The claim was AI. What was missing was any reason the product would be better in a year than it was at the demo.”

Doug Nissinoff, Founder and General Partner, Intelligence Ventures

His position fits in one sentence: “In AI healthcare, the data is the product.” He goes on: “The model is what you do with it, and the companies we back get better with every case they touch because of data only they have.” So the fund passes, “and I tell the founder exactly why, because it is fixable.”

The other common challenge, he says, is that founders often build products “without knowing if there are any customers that are willing to buy it.” His advice is specific. “Whenever founders struggle to answer the ‘who will buy it’ question, I always advise them to do their customer discovery,” he says. “Target interviewing 25-50 stakeholders in your industry and at least 3 signed letters of intent from prospective paying customers. Even if they don’t convert into sales, it shows that the founder has done their homework and will likely live to fight another day.”

How he reads a letter of intent, or an accuracy figure, comes from the first half of his career.

What a Biotech Background Hears

“I spent the first half of my career in biotech investor relations,” Doug says. “Essentially, my 9-5 involved translating complicated science so that anybody with a generalist business degree could understand the financial implications.” He holds a Master of Science in Biotechnology from Georgetown University, worked in investor relations at the gene editing company Graphite Bio, and before that held two roles at Jakob and Partners, a Danish biotech and pharma consulting firm. Mostly, he says, the background lets him “hear the number correctly.”

The other thing is translation. “Healthcare buyers have their own vocabulary and founders repeat it,” he says. “‘Pilot’ often means a free trial. ‘LOI’ often means an unpaid commitment. ‘Signed’ often means it goes to a committee next month. ‘FDA pathway’ often means nobody has asked the FDA yet. None of it is lying. But if you have worked inside an academic health system you translate automatically, and you can tell the difference between a company that is early and a company that is overstated.”

He has posted a longer glossary on LinkedIn. Two entries from it: “‘Data moat’ = We have a CSV” and “‘Clinical validation’ = We showed it to a doctor.” This, he wrote, is why he spends “an unreasonable amount of time digging into the details. Because sometimes the difference between a $2M company and a $200M company is hidden behind the words. And sometimes it’s just a dashboard.”

The background also changes the first meeting. “When the first question is about the validation cohort instead of the market size slide, the conversation changes, and that is where the trust comes from,” he says. “A generalist can hire that later. I would rather have it in the room on day one.”

“A scientist founder can tell within ten minutes whether you understand what they built.”

Doug, Founder and General Partner, Intelligence Ventures

The accelerator is where he has put that room together.

Ten Weeks, Two Sessions a Week, a Check at Acceptance

IV Accelerator is a ten-week program for pre-seed and seed AI healthcare companies, targeting ten companies per cohort. It invests $25,000 in each company at acceptance, “so every company in the room is a portfolio company from day one and my incentives are the same as the founders’.”

The program opens with three days in person in Tampa. After that it runs virtually, two sessions a week: a speaker session on Tuesdays and a working session on Thursdays, about four required hours a week, “timed so a founder can run the company around it.” Tuesday speakers come from the fund’s LP base and partner firms, and from the health systems and corporate innovation teams working with the cohort. Thursdays, Doug says, “are built around what each company needs next, so a company selling into hospitals spends that time on procurement and a diagnostics company spends it on validation and regulatory strategy.”

The mentors are the fund’s own investors: “physicians, health system executives, biotech operators, and exited founders, the people who have bought or built what these companies sell.” What a founder gets, concretely, is the check, one-on-one work on their own deck and go-to-market with the mentors and partner firms, introductions to the health systems working with the cohort, sponsor perks “that cover the software and services a seed company pays for anyway,” a room of founders working the same problems, and Demo Day in New York on November 4 in front of the fund’s investor network and health system leaders.

The first cohort is in the middle of it now.

The First Cohort, in Numbers

More than 100 applications in the first two weeks, 27 interviews, nine offers, nine companies. The program opened in Tampa on September 23, Demo Day is November 4 in New York, and the cohort keeps meeting through early December.

When he announced the cohort on LinkedIn, Doug listed what the companies walked in with, including “FDA clearances already in hand,” “NIH grant funding” and “Multi-year hospital contracts.” “Not a slide deck and a dream,” he wrote. “These teams got real before they got funded.” Among the products he described: “AI that predicts brain aneurysm rupture risk from scans hospitals already take” and “Contactless vital sign monitoring for the hospital beds nobody is watching.”

Around the cohort, ten health systems have committed people to teach or to work directly with the companies, alongside innovation teams from pharma and medical device companies and sponsors and partners covering banking, legal, go-to-market, branding, recruiting, cloud credits and accounting. “Every company in this cohort needs a clinical partner,” Doug says, “and the accelerator works because health systems show up for it.” He would like more people inside health systems, biopharma and corporate innovation teams to see the cohort as “a screened set of companies worth an hour of their time.”

Fund I closed in May 2026 and has 25 LPs, and the fund has fourteen portfolio companies including this cohort. Doug adds what he can’t show yet: “The numbers that will matter most are the ones we do not have yet: what these companies sign with health systems, and who they raise from, in the six months after Demo Day.”

Launching the fund behind the program was, in his words, “truly challenging.”

Working on the Fund Until 1 a.m.

Doug’s way into venture came through one person. “Craig Pierson gave me my first true role in venture,” he says. “When he formed AiM Medical Robotics, an MRI-compatible neurosurgical robotics startup, he took a gamble and asked me to consult for them since I had an investor relations background and was local to him in Florida.” It went well, he says, and it led to Intelligence Ventures Group, where he has “consulted for 25+ startups and raised millions of dollars for them.” He continues: “That gave me the confidence to launch Intelligence Ventures Fund I. Thank you for believing in me, Craig.”

“I lacked a track record to lean on, nor did I have the stereotypical VC pedigree,” he says of the launch. “After two years of plugging away, I still couldn’t raise capital. When my first child was born, I had to take a W2 role leading startup venture development at a university to pay the bills (diapers are not cheap).” The university was the University of South Florida, in tech transfer, where he led venture development.

“My daily routine involved working until 5 pm, spending time with the family until 8 pm, hitting the gym until 9, and then working on the fund until 1 am.”

Doug, Founder and General Partner, Intelligence Ventures

“I maintained this schedule for 1.5 years,” he says, “until I secured enough investors and traction to pursue this full-time.”

Where Doug Is Now

Full-time, this year, means one thing first: “Get this cohort to Demo Day on November 4, then do it again in the spring.” Applications for the second cohort open in October, and the founders Doug wants are the ones with clinical or biological data underneath the product, including the ones not sure a small fund in Florida is worth an application. He also wants to hear from people inside health systems, biopharma and corporate innovation teams who would give the cohort an hour. Founders come first: “everything we do is for them.”

Outside work, the count is different. “Three kids under two,” he says. “My son is almost two, and the twins arrived in June, the month after Fund I closed. Nothing in venture has taught me more about triage. Every second of my day AND night is accounted for. Intense, but extremely fulfilling and I wouldn’t have it any other way.” One habit he keeps: “Gym before any major decision. It is the reset button needed for absolute mental clarity.”

For the founder who gets a no from him, and the reason with it, the rest of the answer is short: “Own the data, and come back.”

The Doug Nissinoff Playbook: Reading an AI Healthcare Pitch

Ask where the model came from. If the answer is an off-the-shelf model with a prompt and rules on top, and the company owns none of the data going in or coming out, he passes.
Ask what the 94 percent is a percentage of. Retrospective or prospective, how many sites, who labeled the data, and which cases sit in the 6 percent.
Translate the buyer’s vocabulary. A pilot is often a free trial, an LOI often an unpaid commitment, and “signed” often means a committee next month.
Tell the founder exactly why you passed. He does, because it is fixable.
Write the check at acceptance. $25,000 goes into each company on the way in, so his incentives are the founders’ from day one.