Founder mindset, police body camera analytics, and NLP applied to trust

Executive overview

Most police misconduct stems from poor communication, not malice — and that's measurable. Anthony Tassone built Truleo to analyze body camera footage using natural language processing, surfacing coaching insights for sergeants.

Good decisions require clarity grounded in reality, not ideology. The quality of questions you ask — of customers, data, or hires — determines the quality of outcomes.

The founder's edge is pattern recognition applied to problems others haven't framed correctly yet.

Accountability and decision-making

• Success correlates to the quality of questions you ask in meetings
• Most founders blame external factors; the person who hired a bad employee is the person who kept them there
• Gut instinct sharpens with experience — especially in hiring, where data only becomes clear once someone is working with you
• Act on poor hires quickly; keeping the wrong person harms the team, investors, customers, and that person
• Political environment is largely irrelevant to operator success — operators adapt to new truths

From Wall Street to NLP

• Quantitative finance forced learning game theory over formulas; herd psychology matters more than math
• Natural language processing (NLP) structures unstructured human communication so machines can analyze sentiment, intent, and patterns
• First company (Green Key) transcribed banker phone calls to surface client insights — sold in 2021
• Key mistake: assumed banks would move data to the cloud; timing mattered more than the technology being correct
• Viral coefficient matters in B2B — banking clients demanded confidentiality, blocking organic referrals; policing is the opposite

Building Truleo: body camera analytics for police

• 650,000–700,000 of ~1 million US cops wear body cameras
• Core insight from data: officers who give more explanation have better outcomes
• Signal vs. noise example: profanity alone doesn't predict bad outcomes — directed insults do
• Tone and cadence carry insufficient signal; language structure and listening behavior do
• Product designed for the sergeant, not the chief — sergeants coach eight or nine officers and can intervene early
• Fear drives command-heavy language in young cops; experience produces negotiation and explanation
• Goal: help young officers speak like experienced officers as quickly as possible

Early results and product direction

• 20 departments live out of 18,000 in the US — very early
• Alameda, CA: 36% drop in use-of-force incidents over 16 months
• Knowing camera data is analyzed changes officer behavior directly
• Roadmap: virtual sergeant bots that automate supervision, coaching, and content creation for departments
• Focus discipline: deliberately excluded tone analysis and image analysis — only work on what can be done well

Crowdfunding as a B2B growth lever

• Crowdfunding creates activated ambassadors, not just capital
• A $100 investor can drive a six-figure department contract by advocating to their city council or mayor
• Relevant for mission-driven B2B companies where end users (civilians) have influence over buyers (municipalities)
• Works as a digital marketing exercise — call to action for engaged stakeholders who want to participate

Founder operating principles

• Define the problem first — everything else follows from problem selection
• Protect the product with "no" — customers in B2B will bolt on requests that dilute core value
• Know precisely what you can do exceptionally well; don't half-ass weaknesses
• Maintain an athlete's discipline: sleep, limit alcohol, treat physical health as a performance input