Two years ago, a sharp technical team, thoughtful customer interviews and a compelling pitch deck could secure a seed round. Today, that’s merely table stakes and no longer enough to get you funded.
In today’s AI era, the rules have fundamentally changed. For founders building in the application layer, the message from investors is deceptively simple: just build and get customers.
Venture capital firms still publish their opportunity lists and Y Combinator continues to release quarterly Requests for Startups. Yet, one piece of advice remains consistent, if somewhat overstated: build, ship and acquire customers before raising money.
Two factors drive this shift. First, AI has demolished development barriers, letting founders ship initial versions at unprecedented speed.
Meanwhile, the proliferation of AI startups, alongside major companies building their own solutions, has raised the bar significantly. VCs now expect fully functional applications with real users before they’ll write checks.
This standard holds even as AI startups flood social media with eye-catching growth metrics, anointing a new “fastest-growing SaaS in history” seemingly every few weeks.
The Zero-to-Traction Imperative
The new normal looks radically different from even two years ago.
“You’re now seeing companies that get really far with almost no funding,” Kory Jeffrey, a partner at Inovia Capital, said in an interaction with AIM.
Inovia most recently led a $500 million round for the enterprise tech startup Cohere.
“Unless companies are building really novel technology today, there really isn’t a reason to raise $10 million before you have customers to show for it,” he added.
The scepticism that once surrounded AI products has evaporated. People are now ready to pay for AI-powered tools through monthly or yearly subscriptions.
The rise of vibe-coding, using AI to build complete applications, has compressed development timelines so dramatically that having paying customers before fundraising has become the baseline expectation.
“The burden of proof on the founder has gone from like market research, insight and a good tech team that will build it to people who are using it,” Jeffrey added.
Consider Supermemory AI’s trajectory. Founder Dhravya Shah spent two years building before raising a $3 million pre-seed round. During that time, he accumulated over 50,000 users and 10,000 GitHub stars before methodically fixing scaling issues and attracting companies seeking AI memory infrastructure.
By the time Shah approached investors, Supermemory had become, according to a company blog post, “extremely scalable while providing some of the best latencies in the market”.
His customer list included AI video editor Montra, AI search engine Scira AI, real estate tech firm Rets, a16z-backed desktop assistant Cluely, and a robotics company.
This isn’t an isolated case. Base44 grew to 2.5 lakh users while bootstrapped, ultimately selling for $80 million.
Lovable accumulated 27,000 waitlist signups and 52,000 GitHub stars before raising its €6.8 million pre-seed round in October 2024.
And there’s a growing list of such examples.
The Product-Market Fit Paradox
Yet, having customers and traction doesn’t necessarily mean founders have achieved product-market fit (PMF) and that distinction matters when seeking funding.
Bessemer Venture Partners’ recent playbook on PMF in the AI era offers practical guidance: brilliant AI ideas mean nothing unless they solve a real problem for real users.
“One of the things that’s unique in this moment is that market and buyer preferences and needs are changing at the same time that founders are trying to find PMF,” Lauri Moore, a partner at Bessemer, said in the report.
“It’s like buying shoes for a growing kid. One day they fit perfectly, and the next they are too snug or their favourite colour has changed.”
Founders must focus on delivering an MVP that shows immediate, clear, undeniable value, even if scrappy.
This involves measuring actual user behaviour, time-to-value and usage metrics, not just collecting positive feedback.
The success of any product demonstration requires showcasing the “wow moment” during the sales process. If prospects don’t see value during a trial, they won’t trust the solution after purchase.
Companies that allow prospects to quickly see how their product adds value to existing workflows, even without deep integration, will drive faster adoption. Once the value is proven, founders must design for repeatable use by embedding these features into daily workflows.
What Investors Still Value
While traction matters, it’s not everything. The intangibles still play a crucial role.
Jeffrey emphasises that assessing founding teams remains essential at early stages.
“We look for a lot of evidence of exceptional ability—whether from previous startups, corporate work, academics or any knowledge that comes up in a conversation.”
He’s particularly drawn to founders solving their own problems. He stated two examples. Shopify exists because Toby Lutke struggled to sell snowboarding gear online. Spellbook emerged when founder Scott Stevenson, frustrated by exorbitant legal fees, decided to build his own legal tech solution.
“You want people who are just so determined that they can’t picture a world that doesn’t work the way they want,” Jeffrey added.
The second founder profile he seeks: domain experts who built solutions long before AI became ubiquitous. “These companies know exactly where and how, and more importantly, how not to apply AI,” he said.
For instance, Replit spent years perfecting browser-based collaborative coding and IDE features. When AI arrived, it integrated Replit Agent for natural-language application building and crossed $100 million in ARR in 2025. Their deep domain knowledge made the difference.
“We really value the hard-earned knowledge that startups have gained through their efforts in the trenches,” he said.
Chemistry Matters
Beyond traction and technical chops, Jeffrey underscored another critical yet often overlooked factor: chemistry.
Securing investment involves a multi-year partnership, especially with lead investors who join your board. It’s a significant relationship that founders often underestimate.
“People whom you don’t get along with don’t get better after the money goes in,” Jeffrey warned.
VCs aren’t just evaluating business fundamentals; they’re asking themselves if they want to spend the next decade working with you. The personal dynamic matters because you’ll navigate both victories and crises together.
“You are in the trenches, you will have good days and really bad days…The people that you go through it with are really going to matter,” he said.
The Expanding Opportunity
Despite—or perhaps because of—these higher barriers, the AI investment market continues to grow rapidly.
According to CB Insights, 72% of AI deals as of Q2 2025 went to early-stage startups.
The opportunity lies in market structure.
While incumbents and tech giants continue to build horizontal platforms, they consistently miss niche use cases. This allows startups to have a space of their own, despite the frenzy of announcements from the big companies.
These overlooked gaps also create acquisition opportunities, as larger players buy innovative startups addressing targeted problems.
Bessemer captured this dynamic in its report titled ‘The State of AI 2025’. “SaaS giants are buying their way into AI. Build technical and data moats. Be M&A-ready, but operate like you’ll own the category.”
The post How AI is Changing the Way VCs Fund at the Seed Stage appeared first on Analytics India Magazine.
