It costs $1,799 a month to rent a model of your audience. You feed it a small survey, and it spits out predictions for new questions in about 30 seconds. That is the initial product from Semilattice, a London-based AI startup founded in 2023. Its long-term goal is far more ambitious: building a real-time model of human organizations, markets, communities, and countries [PERPLEXITY SONAR PRO BRIEF].
For now, the wedge is faster, cheaper research. The company says its models can simulate how defined audiences respond to questions, product changes, and policy decisions [PERPLEXITY SONAR PRO BRIEF]. The bet is that product managers and marketers will prefer querying an AI representation over fielding a new survey every time a question arises.
The wedge into complex systems
The company's name hints at its technical ambition. A semilattice is a mathematical structure describing partial order, a concept used in the study of complex adaptive systems [Complex adaptive system - Wikipedia]. The founders are applying this thinking to human behavior, aiming to develop minimal models consistent with observed real-world dynamics, such as those in financial markets [Computational Modeling of Collective Human Behavior: The Example of Financial Markets | Springer Nature Link].
The initial application is pragmatic. The core product creates a custom audience model from a small amount of survey data. It then predicts how that audience would answer new multiple-choice questions [PERPLEXITY SONAR PRO BRIEF]. A second product surface, called 'Projects,' uses an agent to simulate users and assess the consequences of proposed product or specification changes [PERPLEXITY SONAR PRO BRIEF]. The price tag for a custom model is publicly listed at $1,799 per month [PERPLEXITY SONAR PRO BRIEF].
The operator pedigree
The founders bring a combined three decades of experience from some of tech's most respected product and engineering shops. This pedigree is the company's first line of defense against skepticism about its grand vision.
| Founder | Role | Prior Experience |
|---|---|---|
| Joseph Wright | Co-Founder, Product & Engineering | Google, Stripe, Figma [PERPLEXITY SONAR PRO BRIEF] |
| Fabian Teichmueller | Co-Founder, Commercial & Operations | Google, Kaluza [PERPLEXITY SONAR PRO BRIEF] |
Wright's background spans product and engineering at Google, payments infrastructure at Stripe, and design tools at Figma. Teichmueller adds commercial and operational experience from Google and Kaluza, an energy-tech software company. The team claims more than 30 years of combined experience across product, engineering, marketing, and commercial functions [PERPLEXITY SONAR PRO BRIEF]. This operator-heavy founding team is a classic pattern for startups aiming to bridge deep tech with enterprise adoption.
The pre-seed backing
Semilattice identifies as a pre-seed startup [PERPLEXITY SONAR PRO BRIEF]. While no priced round has been publicly announced, the company's investor roster includes several early-stage funds and angels known for backing technical founders in Europe.
The list includes Concept Ventures, golden square, P2 SPV LLP, and the Tiny Supercomputer Investment Company [PERPLEXITY SONAR PRO BRIEF]. These are not household names, but they represent a cohort of specialist investors who often get the first look at ambitious AI research commercializing out of London. The absence of a loud funding announcement is not unusual for very early-stage teams focused on product before scaling sales.
Where the simulation could stall
The ambition is vast, but the path is narrow. Semilattice's success hinges on convincing businesses that its simulations are reliable enough to inform real decisions. This introduces several concrete risks that the company must navigate.
- The accuracy ceiling. The core value proposition depends on predictive accuracy. If simulated responses consistently diverge from real-world survey results, trust erodes. The company's technical papers suggest a foundation in complex systems theory, but commercial tolerance for error in market research is low [Computational Modeling of Collective Human Behavior: The Example of Financial Markets | Springer Nature Link].
- The data dependency. The model quality is inherently tied to the initial survey data provided by the customer. Garbage in, garbage out remains a fundamental law of AI. The product may struggle in markets where high-quality, representative survey data is expensive or difficult to obtain.
- The incumbent workflow. Traditional market research firms have decades of relationship capital and methodological validation. Displacing them requires more than speed; it requires proving that a faster, cheaper method does not sacrifice insight depth. Semilattice's answer appears to be a focus on rapid, iterative questioning that traditional methods cannot economically support [PERPLEXITY SONAR PRO BRIEF].
The company's most plausible near-term path is not to replace all market research, but to own a specific slot: the rapid, iterative question that arises between formal, large-scale studies. Its API platform for simulating user opinions and behaviors suggests a build-for-developers approach that could embed the tool into product development workflows [Tracxn, 2026].
The next validation milestone
For a company with such a long-range goal, the immediate milestones are practical. The next twelve months will likely focus on proving the initial wedge. That means converting early users into referenceable customers, demonstrating repeatable use cases, and likely raising a seed round to scale the team beyond its founding engineers.
The founders were recruiting for founding full-stack, product, and machine learning roles in late 2024, indicating a build-phase focus [PERPLEXITY SONAR PRO BRIEF]. The company's current site states it is not hiring, suggesting a possible team consolidation or a pivot to a focused launch phase [PERPLEXITY SONAR PRO BRIEF]. The key metric to watch will be the growth of its $1,799-a-month custom model subscriptions. A handful of committed enterprise contracts at that price point would provide the traction needed for a compelling seed story.
Backed by Concept Ventures and golden square, the company has secured the initial capital to test its hypothesis [PERPLEXITY SONAR PRO BRIEF]. The question for the market is straightforward: when you need to know what your users think, will you wait for a survey, or will you ask the simulation?
Sources
- [Tracxn, 2026] Semilattice - 2026 Company Profile, Competitors & Financials | https://tracxn.com/d/companies/semilattice/__dY9iUdshvyXMCkqlLQoIItBf6Sxr2TQo_KXK0PG49Ac
- [Computational Modeling of Collective Human Behavior: The Example of Financial Markets | Springer Nature Link] Computational Modeling of Collective Human Behavior | https://escholarship.org/uc/hcs
- [Complex adaptive system - Wikipedia] Complex adaptive system | https://en.wikipedia.org/wiki/Complex_adaptive_system