Blok’s AI Personas Test Apps Before Users Ever See Them
Developers can now predict user interactions with their apps before writing any code. Blok, a startup that recently emerged from stealth mode, has created an AI simulation engine that generates virtual user personas to test app features in advance.
This approach tackles a key challenge in product development. While AI coding tools like Cursor, Replit, Claude Code, and Lovable speed up writing code, testing new features often still depends on beta releases or basic simulation software.
Founders and Funding
Founded in 2024 by Tom Charman and Olivia Higgs, Blok benefits from their experience as serial entrepreneurs who worked together on startups in travel and learning sectors. The company has raised $7.5 million in two funding rounds.
MaC Venture Capital led the $5 million seed round, with contributions from employees at Discord, Google, Meta, Apple, Snapchat, and Pinterest. Earlier support came from Protagonist, Rackhouse, Ryan Hoover’s Weekend Fund, and Blank Ventures.
How Blok Differs from Existing Tools
Marlon Nichols, managing GP at MaC Venture Capital, explains why Blok stands apart. Unlike tools such as Optimizely and Amplitude that analyze user behavior after features launch, Blok predicts user actions before any code is written.
“Teams ship faster than ever but still rely on A/B tests and gut instinct for critical decisions. Blok flips that model by simulating user behavior ahead of development,” Nichols said.
Addressing Growing Testing Needs
Modern interfaces are becoming more complex, increasing the need for better testing. Olivia Higgs interviewed over 100 product engineers to identify key pain points. She found that as users interact through chat and voice, introducing new visual UI elements requires careful testing to avoid workflow friction.
Testing challenges vary by company size. Small companies lack large user groups for feedback, while larger companies want to prevent cluttering apps with poorly tested features. “Our goal is to eliminate the need for experimental feature releases that require weeks or months for results,” Charman added.
How Blok’s Platform Works
- Companies upload event log data from platforms like Amplitude, Mixpanel, or Segment.
- Blok uses behavioral modeling to create user personas representing most of the app's users.
- Development teams submit Figma designs and experiment details, including hypotheses and user goals.
- The AI personas run multiple simulations, providing insights on feature usage and improvement suggestions.
- Teams receive detailed experiment reports with persona-specific analysis and recommendations.
- A chatbot interface allows direct querying of experiment data.
Current Focus and Business Model
Blok currently operates behind a waitlist and works primarily with customers in finance and healthcare, sectors that cannot risk public experiments with poor user experiences.
The startup uses a SaaS model and optimizes compute costs. It aims to hit mid-single-digit millions in revenue this year while expanding beyond the initial waitlist.
For product teams looking to improve feature testing efficiency, solutions like Blok’s AI simulation engine represent a shift from reactive to predictive testing methods.
Your membership also unlocks:
