AI Safety Quest would like scale up their successful Navigation Calls program to increase completed calls from 150 to 750 calls per year by recruiting more coaches, improving software systems, and increasing marketing.

What's the problem you're trying to solve?

Imagine an altruistic, ambitious, truth-seeking individual trying to find their place in the AI safety ecosystem. In the ideal scenario, they quickly land on the resources, communities, and projects that kickstart them into productive work with AI safety impacts. Alternatively, they might get mired in doubts about their capabilities and the likelihood they can reach their goals. Others might get lost in questions about how to have impact and how to execute their plans. One likely outcome of unproductive and unsatisfying efforts is to stop working on AI safety and select another problem area.

A quick (free) call with a coach could make a huge difference to keeping these people on the path to contribution.

Who is AI Safety Quest and what are Navigation Calls?

AI Safety Quest's Navigation Calls provide free, personalized 30 to 45 minute calls connecting newcomers with volunteer coaches who help them find direction, understand the landscape, and take concrete next steps. The program has existed since March 2023. We have advised hundreds of callers at various stages of career experience and AI safety involvement. The goal is simple: identify ways the callers can deepen their involvement with AI safety, whether that means learning, projects, volunteering, networking, or job hunting.

Our structure as a volunteer-based organization allows us to increase bandwidth (i.e. take more calls) without the overhead costs of hiring employees. We currently operate with a core team of 3 operations volunteers (Chris Lonsberry, Mick Zijdel, and Ricky Holtz) who keep the software systems (applications, scheduling, surveys) running smoothly and recruit volunteer coaches.

How do you plan to scale up?

We can frame this project as a test of several hypotheses. Funding us at the 25k level allows us to run this test for 6 months (and 50k would allow us to run the test for 12 months). We would like to test the following hypotheses.

1. There is significant unmet demand for coaching calls from people who are likely to contribute to AI safety outcomes. Our mid-term impact survey, which we began rolling out in July 2026, will allow us to gather better evidence of effectiveness than the immediate post-call survey.
2. A paid contractor can recruit new coaches to maintain an active coach roster of 6-8 coaches (under the assumption that each coach takes 2-3 calls/week). Faster execution on inbound volunteer interest leads to dramatically higher volunteer retention.
3. Human calls will retain their value as AI advising tools come online and improve (both Lens Academy and 80,000 Hours now have tools in this category).

If any of the above hypotheses fail during the testing period, the likely result is that AI Safety Quest winds down their program and publishes their retrospective.

How will you spend the money?

6 months of contractor wages at 20 hours per week + travel and living expenses to attend EAG or other relevant conferences. (20 h/wk x 45 $/h x 25 wks = $22,500, leaving $2,500 for conference expenses.)

Rather than attempting to find and outside hire during this critical testing period, we will promote one of the existing core volunteers to a paid employee; most likely Chris. If the testing period proves successful, we will aim to hire at that time with the hope we can recruit from our pool of proven volunteers.

What will your contractor do?

Operations, recruiting, and marketing. We expect the main task is to fix the current problems with onboarding new coaches. More coaches translates directly to more calls.

1. In the past, very modest amounts of marketing have been sufficient to fill the call pipeline, so we're not worried about that. In fact, we've explicitly avoiding marketing our services for fear of creating long wait times for callers.
2. We have existing ideas for recruiting coaches that we haven't acted on due to lack of time for onboarding. We also have a new avenue for recruiting coaches that didn't exist last year: as Lens Academy scales up, they'll be producing large numbers of facilitators for their courses and we intend to actively recruit from that pool.
3. Under the current structure, we have not been good at helping new coaches complete the onboarding program. Volunteer interest has a short half-life. If you don't respond quickly and assign meaningful tasks, potential volunteers lose interest and move onto other projects. Having an employee will ensure we're checking the inbox daily and following up immediately and aggressively with promising Navigator candidates.

Don't other organizations already do this kind of coaching?

Yes and we are hypothesizing significant unmet demand for calls that still prove impactful. We also think our calls are unique in important ways.

• Our application is short and easy. Our bar for accepting calls is easy to clear: the applicant need only demonstrate basic AI safety knowledge.
• We are fast. Our review time for incoming applications is typically only 48 hours and following approval, we almost always have call slots available within less than 1 week.
• Even after hiring an employee, AI Safety Quest's cost to deliver calls will remain a small fraction of what it costs other organizations with paid coaches.