AI-native pilot launch—a new 0 to 1 moment for Pursuit

• Rethink Priors: We've dramatically rethought our staffing and program structures, shifting our “human-instructor”-to-participants ratio from 1:18 to 1:100. We are able to supplement “human-instructor” with “AI-instructor” on the new Pursuit training platform. This represents a 5x increase. The pilot itself is streamlined to seven months, with modular two-month off-ramps, significantly shorter than our traditional 12–18-month programs.
• Days, Not Weeks: Our execution speed has accelerated dramatically. Just 59 days passed from my initial challenge to our team on January 16th to the full pilot launch on March 15th—by far our greatest velocity ever. Historically, major revamps (i.e. from iOS to Android or from online to hybrid) took 9-12 months.
• Build in Public: We’ve launched our Build in Public strategy to share our successes, challenges, and learnings with daily postings across our different social and communication channels. We are also requiring all Builders to actively participate and engage on X as a way of learning as well as potentially leapfrogging the traditional hiring process. NYTimes reporter Steve Lohr is embedded in our activities – he has spent 40+ hours with us so far and is joining 2x a week for the AI-native pilot.

Moreover, we've accomplished this while reducing our overall expense budget by 38% compared to our December Board-approved plan. This is scrappy and focused. We would not be able to do this without embracing an AI-native approach.

• Over 1,200 interested applicants while maintaining mission focus

  • Builders’ demographics: 51% without college; 40% unemployed; 46% on gov’t assistance and 9% public housing residents; $15K average income

• Shortest application cycle since Pursuit’s inception: 2 weeks (vs 6 weeks)

• Shortest recruitment cycle since Pursuit’s inception: 6 weeks (vs 16 weeks)

• This pilot is already demonstrating a faster, cheaper Pursuit. By embracing this approach, faster and cheaper will compound with each component of our work.

• Although it’s early, we’re hopeful that this will also have better results. The shift to group-based, many-to-many facilitated learning, supported by our internally built AI platform, shows enormous potential. Having experienced it firsthand this past weekend as a participant in the pilot, I firmly believe we’re glimpsing our future.

• We won’t get everything right. But that’s expected. And that’s why we are running this pilot. The goal is to apply learnings and refactor “Pursuit Classic” to create a new standard.

• Evolving pedagogy: We are building a new Pursuit training platform that we are already testing in the pilot. Builders are taught by human-instructors, AI-instructors (via LLM and our platform), as well as by volunteers and their peers. The platform captures learning data directly from the Builders to track progress and capture feedback so that we can rapidly iterate. Similarly, we’re building a platform for volunteers to capture feedback and learning, already utilizing this in the pilot.
• Evolving learning: “Vibe-coding” is the future. For 25% of the startups from the latest YCombinator batch, 95% of the code was produced by LLMs. Managing code production, understanding product, and solving business problems is ever more important. “Learn to learn” is the first milestone of the pilot.
• Evolving the participant role: We've intentionally adopted the term "Builders" instead of "Fellows." Our Builders aren't just participants—they’re actively shaping Pursuit’s future by testing new software, enabling data to direct learning, and providing direct feedback on the AI tools we’re developing. They’re collaborators in every sense, literally building alongside us.

• AI-First: Being AI-native goes beyond changing training—it involves embedding AI and leveraging LLMs across the whole org to utilize these tools and improve all of our workflows - from applicant selection to grant writing to volunteer management.

• Everyone Builds: We’ve shifted from zero org-wide software development (2024 and prior) to once-a-month software build sessions for everyone (Jan and Feb) to now-weekly build sessions. Moving forward, every Pursuit staff member is spending at least 1 day a week (20% of our time) either learning how to utilize tools or develop software. The low cost and ease of the tools enables us to do this.

• Record all data: We have a new ability to capture and record everything that Builders learn and also how the organization operates. This dramatically enriches our data environment, and should allow us to better measure, learn, and iterate.

• We recognize that evolving Pursuit to meet this moment won’t be easy. The labor market continues to shift. Technology is moving quickly. Our stakeholders, reasonably, will have differing levels of engagement and perspectives on this transformation. Not everyone will support this. But I see the conversations with our stakeholders change week to week, especially as the federal policy landscape and technology news is absorbed. (For example, Ezra Klein recently published NYTimes: “The Government Knows A.G.I. Is Coming.” He believes AGI will be here in 1-2 years. And he is surprised at how uneven this understanding is. I recommend reading this or listening to the episode. He’s a pretty middle of the road consensus guy—this is an indication that perspectives are starting to shift from fringe to norm).
• Our conviction and optimism compel us forward. I draw confidence directly from our Builders, whose passion for embracing these changes is clear. One applicant perfectly captured this spirit: "Pursuit’s mission aligns perfectly with my personal and professional goals. I see this program as the ideal place to grow, contribute, and be part of an evolving AI-driven world. I want to be part of a program that values people as much as innovation. Pursuit has shown time and again that it’s committed to both—and that’s why I know it’s right for me."

Our entire team is fully committed, and our new Builders are too. Special thanks and congratulations to Ari (newly promoted Chief Product Officer), Joanna (newly promoted Chief Operating Officer), Greg, Stef, Carlos, Laziah, Afiya, Emma, Dave, Caroline, Victoria, and Alexis. I’ve seen their hard work, embrace of this mindset, and their leadership in making this significant cultural and operational shift. This hasn’t been easy, and there will be further challenges ahead. Thank you, thank you!

I'm personally all-in and am actively participating in the initial two-month nights and weekend module to support the team and also tighten the communications loop to our stakeholders (my goal is to meet the pilot standard of 85% persistence/attendance). Just as Dave and I built Pursuit’s first pilot piece-by-piece alongside our volunteers ten years ago, we’re returning to fundamentals.

I want to be direct and candid: There are real, significant challenges facing job training organizations, the social sector, and Pursuit right now. How we win is by turning the AI challenge into a strength. There is new, unbounded opportunity to bring these AI-native learnings back to our existing stakeholders and forge new partnerships to shape the future.

To our Board members – your continued leadership, support, and engagement are now more critical than ever.

Thank you for believing in Pursuit and joining us in shaping this next chapter.

My sincere thanks,

Jukay