Venture capitalist Matt Ocko says that when he talks to the founders of artificial intelligence startups, he wants to hear how they intend to make money, but do so in an equitable way while also solving a massive societal problem.

“If you have a set of tools that can produce superhuman results or provide superhuman enhancement, it should be in the service of something besides gutting the white-collar workforce,” says Ocko, cofounder and managing partner of VC firm DCVC.

It may sound like a lofty goal: ensuring your AI tools help better the planet, and do so while making money. But Ocko isn’t interested in hearing AI founders pontificate about a magical future decades away, or some giant AI model that will take over the planet, or make workers 30% more effective.

“We want to hear from entrepreneurs that you appropriately understand that AI is an incredibly powerful tool that can be combined with deep, specific industry expertise and proprietary data, against a problem in a giant industry or giant societal problem that needs urgent change,” says Ocko.

AI companies pulled in one out of every three dollars invested during 2023, accounting for roughly 20% of deals, according to PitchBook data. The swell of interest in AI wasn’t enough to offset the 30% decline in investor money that flowed into startups last year, falling to $170.6 billion.

Investors and startups that Fortune interviewed agreed that the explosive interest in AI has led to a troublesome dynamic. Some VCs are only taking meetings with startups that have an AI tool or strategy. But at the same time, these same investors fret the companies they do take meetings with are overexaggerating the capability of those tools. The startups are thus incentivized to talk up their AI capabilities to score the meeting and get the cash they need.

At Silicon Valley VC firm IVP, partner Karthik Ramakrishnan has led investments in AI startups DeepL and Jasper. What he wants to hear from AI startups is how they intend to truly reinvent how work is done today, rather than a simple productivity boost.

“Incumbents are moving at a much, much faster pace and have the inherent distribution advantages,” says Ramakrishnan. “We’ve seen from Meta to Adobe, Salesforce to Microsoft, very quickly introduce a lot of AI features into all their major tools, and they’ve actually reached significant adoption.”

DeepL, as an example, is an AI platform that can translate written content ranging from an email to a PDF to a Microsoft Word document. Companies spend billions each year on translation, and Ramakrishnan likes that DeepL can address such a massive market globally.

When hunting deals, Ramakrishnan says the firm also relies heavily on customer feedback. He says they are honest about whether the product works.

Intuit Ventures, the VC arm of TurboTax and QuickBooks provider Intuit, was launched in 2021 to find founders that the company could invest in beyond innovation by Intuit’s workforce and the corporate development team’s acquisitions.

One question Adam Coccari, managing director at Intuit Ventures, asks is, “What do you really mean by an AI company?”

Coccari sees three main categories emerging. The first are the companies like OpenAI, which are building transformative foundational models. These startups are very competitive and capital intensive. Then there are the tools and infrastructure around those models, and finally, applied AI, which takes the tech and applies it to a workflow or industry problem.

Applied AI is what Coccari is attracted to. The questions he asks AI founders don’t differ too much from any other founder. “It’s really looking at their background and understanding, what they see as an unresolved pain point in the industry or workflow, that they think AI can be a 10x better solution,” says Coccari. An extra element, Coccari adds: “How does AI make something truly better?

“What I want to see in the next wave of applied AI is how that gets woven into people’s workflows in a natural way, or a different interface that we really haven’t seen yet,” says Coccari. Like all founders, he also homes in on perceived resilience in the face of challenges, their ability to attract and retain talent, and the market opportunity they intend to exploit to compete in a competitive space.

Soraya Darabi, founder and general partner of early-stage venture firm TMV, says the questions she peppers into diligence conversations with AI founders are everything short of “Who do you want to be when you grow up?

“You want to be getting into business with people that you think have the same values alignment as you in the long run,” says Darabi.

As it applies to AI, the businesses that TMV invests in must align with the firm’s four priority verticals, which include sustainability and logistics—and from TMV’s vantage point are both good for business while driving the greatest returns. A few AI startups that Darabi likes and has invested in include AI medical dictation tool Tali AI, AI scheduler Clockwise, and medical education platform Roon.

Lawrence Watkins, head of local investments for Tulsa-based venture capital firm Atento Capital, says he believes there aren’t too many differences between an AI founder and those that operate in other technological spaces.

“Our theory is that everyone is going to have to be an AI founder at some point in time in the future, because AI is going to literally eat all different industries that are out there,” says Watkins.

With AI startups, Watkins does spend extra time asking questions about the power of data, specifically regarding privacy and bias. He asks AI founders who exactly owns the data, is it public or private, if it infringes on a person’s right to privacy, and how decisions are made with the AI’s algorithms.

At DCVC, Ocko says there are founders that pitch splashes of AI on the surface of their business, built on open-source AI tools. That’s not what he wants to hear from AI founders. Instead, he points to Evolv Technology, the maker of detectors that use advanced sensing and AI to search for guns, knives, and bombs to keep them out of stadiums, schools, and hospitals.

AI, Ocko explains, can give people the promise of an equitable, abundant society if done properly. He sees two potential futures as the technology continues to be deployed: Star Wars, a grim universe with slavery, genocide, and brutality; and the resilient, hopeful future that Star Trek represents.

“We’re investing in a Star Trek future,” says Ocko. “Not a Star Wars future.”