Software teams now ship in hours while a conventional user study still takes four to eight weeks. CB Insights found poor product-market fit behind 43% of the VC-backed startups that shut down since 2023. Vectorial, a San Francisco platform that simulates human behavior, announced a $3 million seed round on October 20 to close that gap, with Smilegate and SenseAI Ventures as lead investors and UC Berkeley and Stanford-linked funds alongside product leaders from Meta, Shopify and Mastercard taking part, so that product, design and marketing teams can test a journey, a campaign or an AI agent against thousands of modeled users in minutes.
A $150 Billion Industry Moving Toward Software
The research market Vectorial sells into is large and changing shape. ESOMAR's 2025 report put global insights industry turnover above $150 billion in 2024 and expects it to pass $160 billion by the end of 2025, with $62 billion of it in research software that grew 11.5%, $56 billion in traditional market research services that grew 4.8% and $35 billion in reporting that grew 8%. The mix matters because software is the segment taking share each year, which means a platform that delivers answers from a model of customers is entering the part of the industry that is expanding fastest.
Researchers themselves are moving the same way. A Qualtrics survey of 3,198 market researchers found that 71% expect most market research to be done with synthetic responses within three years. Its 2025 follow-up of more than 3,000 professionals in 14 countries found 53% using AI regularly, 45% of synthetic-data adopters calling it their most reliable source and 72% of teams using new methods reporting that their organizations used research significantly more than a year earlier. Those numbers describe a profession preparing for a different toolset. They give a startup whose product is a continuously updated model of a customer base an audience that already expects the shift.
Why Product Teams Skip Research
AI has made building and launching products far faster while research has stayed slow, which is why companies skip it at the steps where it would matter. Even with AI-moderated interviews, collecting feedback from 50 to 100 participants takes days. Vectorial's own comparison puts a traditional study at four to eight weeks and $10,000 or more, an AI-moderated round at five to ten days and $5,000 or more and a behavior simulation at hours and about $100. Teams iterating on a launch in hours cannot wait that long for every question, so decisions drift back to instinct and a few experiments with live users, the pattern that leads to products nobody asked for.
The cost of guessing shows up in the failure data. CB Insights reviewed 431 VC-backed startups that shut down since 2023 and could identify reasons for 385, with 70% running out of capital, 43% citing poor product-market fit, 29% pointing to bad timing and 19% reporting unsustainable unit economics. Running out of money is usually the last event rather than the first. A product that does not match a real need is one of the problems that earlier and more frequent customer testing is meant to catch before the cash is gone.
The Founder Thesis
Taranveer Singh, Co-founder and CEO of Vectorial, describes AI as moving along two frontiers, one focused on reasoning and solving complex problems and the other on simulating and decoding human behavior. He sees two missing pieces on the simulation side: the ability to run many scenarios and a way to bring users into the same room as the people building the product. His co-founders cover the rest of the loop. Rahul Garg, Chief Product Officer, launched multiple products at unicorn companies and Jasmine Kaur, Chief Business Officer, led synthetic user work for marketing at Pfizer while Singh has spent a decade modeling human behavior including in defense research and research labs.
The company's forecast is that within four years running a behavior simulation with audience models before building or launching a consumer product or agent will be the norm, in the way simulation already is in aerospace, manufacturing and chip design, even though consumer products that reach billions of people still launch largely on gut instinct. Singh also argues that the two existing options both fall short, since AI-moderated interviews make every new question a multi-day study with a handful of participants while simulated user agents answer instantly with no grounding in real behavior data. Vectorial positions its audiences as grounded in behavior data at scale and built into the product development, launch and build workflow.
How the Platform Works
Vectorial's modeled users are powered by SAPIENS, a behavior model that learns the latent socio-demographic, psychographic and behavioral traits of a target customer and then predicts how that customer will act. The company says it has modeled more than 2.5 million real people and run more than 30,000 simulations, that SAPIENS is 30% more accurate than general-purpose language models at predicting behavior and outperforms academic models such as Be.FM and Centaur, the Nature-published model from Helmholtz Munich trained on more than 10 million human choices and that it is validated against real-user studies. Teams add an audience, ask a question or upload a journey, campaign or agent and watch the modeled customers react before a real user touches it, with the product site adding that every number can be traced to the individuals behind it and that customers can calibrate the population with their own survey, CRM and analytics data.
Early Customer Results
Customers span healthcare, gaming and e-commerce and include public companies and billion-dollar enterprises, among them Midi Health, GoodRx, Suki AI, boAt, Everkind and Concrete Software. They run about 600 simulations a month, the equivalent of 40 user studies and 20 times more research than before. Simulations have reproduced earlier user-study findings with more than 85% accuracy. Jasmine Kaur says enterprises are treating a continuously enriched model of their audience as a strategic investment because product, design and marketing teams all design for the same people.
The customers describe different problems with the same tool. The chief technology officer of Everkind, a publicly traded company in Canada, uses simulation to test how an AI conversation lands with very different people such as a first responder coming off shift or a parent running on no sleep before an update reaches a real user. The chief marketing officer of boAt says consumer research has historically informed only a fraction of the hundreds of decisions in building a brand and sees simulation bringing customer understanding into everyday choices on pricing, positioning and campaigns, while a data services executive at Concrete Software, the studio behind the PGA Tour and NHRA Drag Racing games, uses modeled players during development and live operations to get a first read on features and content in minutes instead of days or weeks.
The Investor Thesis
Capital is moving into the category quickly. Aaru raised a Series A led by Redpoint Ventures in December 2025 at a $1 billion headline valuation, Simile raised a $100 million Series A led by Index Ventures in February and followed it in July with a Series B of more than $200 million at a $2 billion valuation, which means the three largest rounds in behavior simulation total more than $350 million before Vectorial announced its seed. Investors are reading the same demand signal that the Qualtrics survey describes. A seed-stage company with paying customers across several industries and an academic model team gives them an early position in a market whose incumbents are research houses built around panels and fieldwork.
Rahul Agarwalla, Managing Partner at SenseAI Ventures, says software has moved to lightspeed while understanding user needs remains painfully slow and that Vectorial replaces the slow theater of surveys with millions of simulated humans using the product and giving insights, a new class of AI that predicts human behavior and not only the next word the way a language model does, which is the reason his fund invested.
The Round
The $3 million seed is backed by institutional investors Smilegate, the Seongnam-based investment firm, plus SenseAI Ventures, an AI-focused fund that began in 2017 and manages a maiden fund aimed at Indian AI startups, alongside the UC Berkeley network in Berkeley SkyDeck and the Berkeley AI Research Lab, Stanford-led funds in the Stanford GSB fund and Dent Capital, plus product leaders from Meta, Shopify and Mastercard. The investor list mirrors the company's academic backbone across the AI communities of Stanford through StartX, Carnegie Mellon through Venture Bridge and UC Berkeley. The founding team comes from Stanford, CMU, ISB and BAIR.
Vectorial will use the money to expand its behavioral datasets through new data partnerships and to advance SAPIENS 2.0, with the stated goal of becoming the de facto generative simulation infrastructure for building and launching products. More data from more partners should sharpen the audience models. A stronger model widens the set of decisions that teams can test before launch, from a single onboarding screen to a pricing page or an AI agent's tone.
What Comes Next
Vectorial's target is a living model of every company's customers that keeps pace with new behavior, a product that gets better as its datasets and customer base grow. If the Qualtrics respondents are right and synthetic responses dominate research within three years, the companies with audience models already grounded in real behavior data, validated against real studies and wired into product workflows will be positioned to set the standard. The seed round gives Vectorial the resources to build that position before the incumbents adapt.
Don’t forget to like and share the story!