Investors at GE Ventures, Sanofi Ventures, and Nexus Venture Partners turned up at Health 2.0 in Santa Clara, California this week to deliver a reality check about some of the innovations coming down the pipe in health tech.

“If you’re an early stage company looking for funding, make sure your deck says you do AI,” Iana Dimkova, director of healthcare investing at GE Ventures, joked. “It doesn’t matter what you do, just make sure it’s on there and it automatically results in a 20 percent bump in valuation.”

But despite investors’ love of artificial intelligence and machine learning, Dimkova says that in practice, the data often doesn’t exist in a format that an algorithm can work with.

“So for all of these companies saying they can actually implement artificial intelligence and machine learning, you have to understand how bad the data problem actually is,” she said. “The inability to aggregate, to draw from the EHR and to, in any meaningful way, push back into the EHR [is problematic]. Interoperability efforts have largely failed. I was really hopeful about HL7 solving some of those problems, but I don’t think it’s happening. I think we have a long way to go before the very innovative solutions being worked on by really brilliant people actually impact the day-to-day care of the patients.”

Another area that Dimkova poured some cold water on was consumer-facing healthcare products.

“While I do believe the consumerization of healthcare will happen, we’re not really looking at clinical products that are going directly to the consumer. Unfortunately, a lot of money has been spent and so far there’s not really a business model to emulate,” she said. “It turns out most people really don’t want to pay for their healthcare out of pocket. If they’re sick, they still go to the doctor. We don’t really see value because even if your solution is good, you have to spend so much to really market to the consumer. I believe we’ll get there but it’s not one we’re looking at very quickly.”

Abhishek Sharma, an investor at Nexus Venture Partners, said that he’s avoiding the whole large group of startups that sells to hospitals or health plans.

“Instead I’m looking for tenacious, ambitious entrepreneurs that want to stand up to mimic one of those stakeholders by becoming a new-age PBM or a new-age payer,” he said. “It’s a harder life choice, but a lot of value can be created inside closed loop systems if you become a stakeholder yourself. Otherwise, more and more we are seeing that the sales cycles are long, there’s a spiral of pilots after pilots, and incumbents keep giving entrepreneurs false hope of light at the end of a tunnel, when the reality is 500 companies are knocking on their door and it’s really hard for incumbents to evaluate those products.”

Ruchita Sinha, director of investments at Sanofi Ventures, pointed out a couple of examples of companies going that route.

“I am a big fan of things that merge the traditional boundaries in healthcare,” she said. “Healthcare for a big part has very fundamental structure problems. That to the most part has led us to all the issues we have today. As an entrepreneur you have the freedom to not be bound by those. So you get Oscar, a new age payer becoming a provider. 23andMe is becoming a pharmaceutical company. So they’re breaking down those traditional barriers.”

One more area where the technology hasn’t quite caught up with its own hype cycle is virtual reality. The problem there, Sinha and Dimkova said, is that it’s too established for hardware plays to be worth investing in, but not ubiquitous enough for software plays to be a good bet, making it hard to find smart VR investments.

Interestingly, there’s another area of investment that used to have that problem, but has now matured to a point where the prospects are rosier: Voice.

“With Alexas and Google Homes being in most houses, there are so many data plays you can have,” Sinha said. “We’re looking at very interesting applications from a health perspective, in neurodegenerative diseases [for instance] where they can basically build biomarkers based on how you interact with Alexa at home. So that is an example of a platform that is now mature and is mass media, and now you can start building on it.”

Besides ambitious startups and Amazon Alexa skills, investors are looking for companies that address major pain points, Sinha said, especially the ones no one’s tackling.

“Anywhere where you do not understand how things work is an area ripe for disruption,” Sinha said, pointing to examples like how risk pools are calculated and how drug formularies are created for pharmacy benefit managers. “There are so many areas with a complete lack of transparency a complete lack of understanding of how they work.”