Pitch decks often present several of the same ideas with slightly different fonts, illustrations, and numbers.
With misty eyes, many of these VCs are secretly hoping that someone will come along and pitch something completely different – something to that big market opportunity, demographic, or problem that everyone else is ignoring.
I wish these VCs would speak up and say what they really want. luck We’re here to help!
We surveyed nine VCs about where they are in the AI boom and what they’re paying attention to. Here are the nine AI startup ideas VCs want someone to pitch to them.
1: Reform in recruitment
James Currier, founding partner at NFX, said there is a desperate need to use AI in recruiting.
“I’ve been waiting forever for someone to reinvent recruiting on the B2B side,” Currier says. “There are examples of people slowly saying, ‘Let’s do this for resumes, let’s do this for applicant tracking systems (ATS).’ These ATS systems have never been that big. There are a lot of them, but no one company has ever been that dominant.”
Since the early 2000s, numerous recruitment agencies have come and gone trying to solve this very problem, but few have survived.
“Recruiting has always been disruptive,” Currier adds, “but with AI, I think we can reinvent that experience. It’s been over a year and a half, and no one has done it. Let’s throw out everything we know. This is a very nebulous problem, and it’s perfect for AI.”
2: Material discovery
At the intersection of AI and materials discovery, Bison Ventures co-founders Tom Biegala and Ben Hemani would like to see more companies use artificial intelligence and other advanced computational methods to develop new materials with new properties and applications. Think new plastics that are more biodegradable and reusable, new alloys that are lighter and cheaper to manufacture, new materials for batteries, and so on.
“AI has been talked about for a long time as a way to accelerate materials discovery,” says Biegala, a founding partner at Bison Ventures. “I’ve never seen a company do it this well, and for now, I’m continuing to explore the business. It’s a very hard problem and requires a very unique combination of AI expertise, materials and chemistry expertise, high-throughput experimentation, and great engineers who understand the specifics of different use cases and how to bring materials to market.”
But this business will require a highly diverse founding team, with expertise in AI, chemistry, engineering and materials, among others.
“The way AI works is really counterintuitive to someone who isn’t a researcher in the field,” Hemani says, “so using AI to advance the discovery of new materials is very confusing and requires a very high level of expertise and very specific application of these tools. And it can’t just be a black box.”
3. AI Speed Reader
“The miracle of AI is not that it can write for us, but that it can read for us,” says NFX’s Currier.
“So when you think about looking for a job, you look at a lot of job ads. And if you’re a recruiter, you look at a lot of resumes and cover letters. On the consumer side, I think there are opportunities in travel, real estate, dating, and credit cards and other financial products. When you’re looking for a place to travel, when you’re looking for a date, when you’re looking for a house, you’re reading, reading, reading.”
“This is better read than we can read,” Currier added.
AI may soon be able to tell you what to read.
“AI will manage my information consumption — tailoring books, podcasts, and articles to me based on my interests and helping me keep learning and absorbing information seamlessly across platforms, whether I’m commuting or on the go,” Mike Ghaffary, general partner at Canvas Ventures, said in an email.
Four: Open source Nvidia CUDA alternative
Nagraj Kashyap, general partner at Touring Capital, has a very specific idea: build an open-source alternative to NVIDIA’s CUDA software that would enable AI developers to work on a variety of hardware platforms and reduce costs for startups.
“Solving this upstream problem could lead to billions of dollars worth of AI discoveries,” Kashyap said.
Kashyap believes there needs to be an open framework that gives AI developers a uniform development interface regardless of the hardware they’re using, just as Android works across a variety of phone manufacturers. And there are obvious potential partners.
“In my opinion, the best company to work with startups in this space is Meta,” Kashyap says. Meta can’t be a cloud provider because the open-source AI developer ecosystem likely doesn’t trust Google or Microsoft.
“Meta’s philosophy is actually quite different,” Kashyap says. “They’re not selling the cloud. In fact, their philosophy is that they want to lower their own computing costs.”
Five: AI-native apps with AI-native interfaces
For Tomáš Tungus, general partner at Theory Ventures, one thing that AI is currently missing is a fundamentally aesthetic element. Thinking back to when mobile app stores first appeared, “everyone was wrapping their websites and mobile apps. Then Foursquare created the first native location app, and everyone was like, ‘Wow, this is different.'”
But we haven’t had our Foursquare moment yet.
“I don’t think we’ve really seen a native app yet,” he says. “We haven’t figured out what that looks like yet. Apart from ChatGPT, which looks like a text message from your phone, I haven’t seen an app with a completely different UX yet. We’re at the stage of, ‘Let’s add AI here, let’s put a chatbot on it.'”
“I think what’s missing is innovative product managers and product designers who are rethinking important software so that it’s literally something we’ve never seen before,” Tunguz says. “Maybe we’re missing a design paradigm.”
6: AI makes invoicing easy
Most companies selling B2B products still take payments by cheque, according to King Goh, a partner at Headline.
“B2B business owners spend a lot of time manually sending invoices, tracking payments, reconciling checks, and dealing with late payments,” he said in an email. “In fact, 55% of payments are late.”
Goh suggests setting up a company like “AI to Close the Payments Loop,” which combines free cheque digitization services, invoicing and CRM software.
“Invoices can be automatically sent, followed up, reconciled and more, allowing business owners to focus on growing their business rather than just managing their cash flow,” says Goh.
And the way to achieve this is, naturally, through AI.
7: AI helps small businesses understand their customers better
Another B2B idea from Goh: Most B2B businesses outside of professional services, especially small and medium-sized businesses, don’t have a CRM in place, so “they know nothing about their customers beyond hearsay and word of mouth and can’t use technology to grow their business,” Headline’s partners said in an email.
“Imagine a world where independently owned manufacturing companies have AI software that knows all the secure and important information, such as customer payment reliability, growth potential, and industry performance,” Goh muses. “And this information can then automatically act on these insights, ultimately enabling independently owned businesses to grow?”
8: It’s all about the data
If you have lots of data – critical data – spread across different systems, Ken Elefant, partner at Sorenson Capital, might be interested.
In an email, he said he is looking for startups in industries such as “security, especially data centers; construction, especially commercial building; radiology, especially cancer-related issues; oil and gas, especially oil exploration; and meteorology, especially weather forecasting.”
What they have in common, besides the volume of data, is a highly collaborative environment and “the need for highly paid people to make sense of all that information,” Elefant said.
9: AI for AI
AI also needs AI. It sounds like a joke, but in the right applications, AI can help make other AI cheaper, for example.
“We think what’s really going to be important to the future of AI are the types of startups in AI that ultimately make AI much more efficient,” says Bison Ventures’ Biegala. “That means compute efficiency, power efficiency, reduced greenhouse gas emissions, and that can be achieved both on the software side and the hardware side.”
Epilogue: “There are many copycat businesses”
If there was one thing I heard loud and clear, it was that no one I spoke to wanted another LLM.
“In general, we see LLMs as being overfunded and AI infrastructure overbuilt,” said Bennett Siegel, co-founder and general partner at A*. “We see the most promise at the application layer, so that’s where we’re spending most of our time.”
Many of the ideas above are at the application layer.
“If you’re going to do something in this space, I think you should do something completely different than what other people are doing,” says Viegala. “It’s a space where there are a lot of copycat businesses: chatbots, image generators, legal document readers. The last thing you want is to do something just a little bit better than what already exists, because you’ll lose out to the competition.”
At the same time, it is important to know, entrepreneurs, that while there is no substitute for a great idea, completely differentiating yourself from others in the market is your challenge for the future.
“I’m not sure that everyone acknowledges that seed-stage companies are undifferentiated,” Siegel says. “If you have a great team with unique strengths, relationships, and unique distribution, that can give you a leg up to move faster and grow faster. But at the seed stage, there’s nothing defensible, so we look for companies that have unique insights and have put in the work to prove why they can win in this particular market.”
