Product × Finance × AI: Financial Products in the Age of AIPart Two中文
Part Two

What people want and what they can bearThe second set of constraints: the nature of users

Part One showed what the market can offer. In the same market, different people get very different outcomes. What a person actually gets from a financial product depends first on the job to be done; second on how decisions are made, since decision biases eat into the outcome that could have been had; third on what the person can bear, since plans that exceed it are often abandoned midway; and all of this rests on the premise that there is reason to trust the promise. Part Two takes these four factors apart one by one.

One decomposition

Outcome the user actually gets = outcome the job requires - losses from decision biasesConstraint: within what the user can bear · Premise: the user has reason to trust the promise

The four factors differ in nature. The job sets the goal and is the yardstick for every outcome. Decision biases are losses that can be reduced. The capacity to bear risk is a hard constraint: a plan that exceeds it will not be carried through. Trust is the gate: without it, the job never begins.

What determines the outcome a user actually gets

Start from the outcome the job requires and switch scenarios to see how decision biases, capacity to bear risk, and trust each change the final outcome. The corresponding chapters are below.

JobWhat is to be done, and what counts as done15The jobs users need done
DecisionsHow decisions are made, and where the biases lie16How users make decisions
Capacity to bearHow large a loss and how much volatility can be borne17What users can bear
TrustHow beliefs update with evidence18How users judge and trust
Illustrative numbers; they show only the direction in which each factor acts.

How to read this part

This part deals only with the individual: a person's goals, judgements, capacity to bear risk, and trust. Each chapter starts from a concrete situation and draws on research in economics and psychology to examine people's motives, constraints and outcomes.

Summary of Part Two

Once the user's side is taken apart, the conclusion fits in one sentence: users want to get something done, yet outcomes are shaped by decision biases that run in a consistent direction, a plan can only be carried out at the level of the weakest face of risk tolerance, and all of this depends on well-founded trust; each of the four can be improved by product design, and each can also be exploited by it.

Summary of Part Two: what each term of the decomposition yields

Switch views: each term's conclusion, its implications for financial products, and its implications for AI entering financial products.

Each row corresponds to one term of the decomposition.

These conclusions form the second set of constraints on designing financial products. The third set comes from AI: how much an agent can reliably do on people's behalf.