Anthony Pompliano’s Bet on a Specialized AI Financial Assistant

Anthony Pompliano says an extraordinary life means different things to different people. For him, it combines working hard to solve problems with doing so where his children can see and learn from that effort. He prefers blending professional and family life.

Rather than follow a fixed ten-year plan, Pompliano says he has pursued opportunities at the intersection of business, technology and finance. After starting businesses, he joined Facebook to learn scaling and growth; afterward, he explored investing partly to learn from other founders.

Silvia began with a problem in his own life: a spreadsheet tracking his finances needed frequent manual updates. He and a technical colleague built a dashboard for him and then considered whether other people wanted it.

At Facebook, Pompliano worked on audience growth and helped executives understand how posts performed. He later reasoned that an audience could help him meet founders and gain access to investment opportunities. He began deliberately building one in 2016.

His first pre-seed fund was about $3.5 million and did not charge a management fee, he says. He and his partner wanted to test whether they enjoyed investing and were good at it. Meeting founders and seeing emerging technology trends made learning itself valuable to him.

Yang says working alone gives him flexibility but he misses collaborating with people. Pompliano argues that assembling capable people who can each operate independently can create a stronger team than isolated individuals working alone.

Pompliano says his hiring approach changed over time. At first he hired people willing to join; later he recruited people with strong large-company résumés. He found that some struggled with a small organization’s speed and lack of resources, even though they performed well in their previous environments.

His conclusion is that the person and the company need to fit each other. He looks for initiative, ambition and willingness to work in an uncertain setting, and says a culture that suits one person may not suit another.

His companies now use paid trials of two to four weeks for many roles. Candidates join the team and work on real tasks at market pay. He evaluates output, communication, response to feedback and how the person works with colleagues; the candidate can also decide whether the environment suits them.

For example, an engineering candidate was asked to build a demo account for Silvia in two weeks. The feature would let prospective users try the product before connecting financial accounts. Pompliano says the candidate shipped it, giving both sides evidence about the work and producing a useful feature.

The trial can end early if either side sees a mismatch, he says, and direct feedback is part of the process. Yang asks how he thinks about hiring costs; Pompliano says he starts by asking how much additional revenue a hire could help generate, rather than looking only at the immediate effect on margin.

He keeps support functions small and uses technology and outside vendors, while looking for people comfortable operating as a one- or two-person team. He recognizes that some candidates want to manage a large team and says the mismatch should be identified in hiring.

Yang turns to Silvia, describing his own difficulty finding time to manage investments and taxes. Pompliano says the product started with a real-time net-worth dashboard and an AI model supplied with the user’s financial context. Customer conversations led the company toward independent investors with more complex assets.

He gives the example of a personal financial statement requested by a private bank every quarter. His company built a way to generate one quickly. He argues that firsthand knowledge of the user group helps identify such problems, alongside technical work and trust.

Pompliano says the company moved from a general model wrapped in software toward a custom system with model routing, tools, agents and its own models. He claims its specialized models perform better on certain personal-finance questions and cost less; he presents these as company results.

As an example of contextual tax questions, he describes a user whose income and business situation led the system to suggest considering a defined-benefit plan. He says the potential tax savings were substantial if implemented. The example depends on that user’s circumstances and is his account of the case.

Pompliano also describes people recording why they own assets and what family members should know if they take over. In one example, a friend wanted to explain why paying off a low-rate mortgage immediately might not match his own plan. Pompliano sees such context and memory as useful for long-term planning.

He suggests asking the assistant to interview a user about goals, risk tolerance, dependents, location and preferences. He reports that a quarter of monthly active users pay and that paying users ask more than twenty questions a week. He attributes some value to people feeling comfortable asking an AI questions they might hesitate to ask a human.

Yang raises privacy concerns about connected financial accounts. Pompliano describes encryption, certification, anonymization and data-retention arrangements as measures his company uses. He says users who do not want to connect an account can enter assets manually instead.

A large AI company entering personal finance creates competition for a small team. Pompliano argues that routing among models, specialized workflows and the company’s own financial system could provide an advantage. He says his engineering team has fewer than ten people and acknowledges that users will decide whether the approach is better.

His proposed ultimate measure is whether the product helps a user increase net worth relative to not using it, through investment decisions, taxes or planning. Yang adds that saving time also matters.

Asked about Bitcoin, Pompliano says the users he observes hold around ten percent of connected assets in Bitcoin and other crypto. He regards that as between traditional and maximalist positions, and says he personally sees Bitcoin as an asset to hold over time. He connects his view to rising prices and monetary policy.

The speakers compare the rising prices they have seen for vehicles and education. Pompliano attributes much of that change to money creation; Yang says that as he gets older, time can matter more to him than additional money.

Pompliano returns to his four young children and the desire to blend work and family. He talks with them about what he does and asks about their days. His goal is to set an example while spending time together.

He tells his children that they solve problems, regardless of whether they ultimately become founders, athletes, entertainers or work in construction. Even in disputes between children, he asks them to consider alternative ways to solve the problem. Yang summarizes the idea as being useful.