Skip to main content

Finance Players

Who will form the circle of trust?

Tightness Diagnostic — 3/5

Incumbent roles are clear. Disruptor roles are clear. The new player class — the AI agent acting as principal rather than tool — has no regulatory home, no driver-governance pattern, and no settled accountability map.

  • Why 3/5 — Banks, VCs, wealth managers, hedge funds, exchanges, custodians: all unambiguous. DAOs, DeFi protocols, AI-native funds: unambiguous. Agent-as-principal: unmapped.
  • What is broken — Every regulated entity knows what it is. An autonomous agent operating its own treasury does not fit any existing category. The driver pattern is the answer, but it is not yet standard.
  • What to teach — Map the player class to the accountable human. If no human, the position is not yet investable.

Ecosystem Map

PlayerRoleAI disruption signal
Investment banksCapital raising, M&A, advisoryJunior analyst roles automated; pitch book generation compressed
VC fundsRisk capital for high-growth companiesAI compresses deal sourcing, DD, and cap table analysis
Wealth managersHNW preservation and jurisdictional planningAdvisory model under pressure; AI surfaces what advisors used to gate
Hedge fundsActive strategies across public marketsEdge eroding as retail gets same research tools
DAOsOn-chain treasury managementPermissionless capital formation; no custodian required
DeFi protocolsLending, trading, yield on-chainReplacing correspondent banking rails for settlement
AI agentsResearch, monitoring, execution — and increasingly principalMoving from tool to participant; own wallets, own treasuries, own positions
Retail investorsPreviously passive; now active with AI toolsResearch capacity 10 stocks/week → 50+ with AI plugins

Three Tribes

The industry sorts every participant into one of three tribes. The ratio between them predicts how fast the agentic shift converts to revenue.

TribeWho in financeCurrent stateImbalance signal
ExplorersCrypto-native funds, AI research startups, agentic-trading desks, DAO treasuries, fintech VCsCapital-rich, capability-rich, shipping in weeks not yearsOver-indexed — pitches outrun proofs
AutomatorsQuant funds, fintech infrastructure, neobanks, DeFi protocols, stablecoin issuers, agent-payment buildersOperationalising explorer findings; throughput risingHealthy — but the bottleneck is now governance, not throughput
ValidatorsRegulators (SEC, FCA, MAS), auditors, rating agencies, compliance officers, custodian banksSlow, fragmented, no framework for agent-as-principalUnder-resourced — every new regime stacks on the last

The current explorer-to-validator capital ratio is roughly eight to one. Validators control gating for one hundred percent of institutional capital deployment. The arithmetic produces pilot purgatory by default — and a consulting opportunity for whoever helps incumbents cross the validator chokepoint with agent governance intact.

Agent Roles Emerging

The player map is expanding. AI agents are not just tools — they are becoming participants with their own wallets, treasuries, and accountability surface.

  • Research agents — Synthesise earnings calls, monitor news, flag guidance changes
  • Execution agents — Place orders on rules (price threshold, volatility trigger, earnings)
  • Portfolio agents — Maintain allocation targets, rebalance on drift signals
  • Compliance agents — Monitor transactions for AML/KYC flags; surface source-of-funds risk
  • Capital-formation agents — Issue tokens, manage treasury, allocate to compute or yield positions
  • Driver agents (governance) — Watch other agents, escalate when a peer commits outside its envelope

The question for every incumbent: is the agent your tool, or your competitor?

Agent-as-Principal — the new player class

When the agent holds the wallet, signs the transaction, and reconciles the position, the existing player map breaks. The agent is not the bank. It is not the broker. It is not the custodian. It is acting in their place.

Three patterns are emerging to absorb this:

  1. Named human driver — every production agent has an accountable human; the agent's actions clear through the driver's authority. The simplest answer. Most defensible. Most current.
  2. Entity-wrapped agent — the agent is encapsulated in a legal entity (LLC, foundation, DAO with members). The entity is the principal; the agent is its operating arm. Works for treasury and trading agents.
  3. Protocol-bounded agent — the agent operates within a smart contract that defines its envelope. The contract is the rule; the agent is the operator. Works for narrow, well-specified jobs.

None of these is settled. The firm that publishes a credible governance pattern first sets the standard.

Buyer Problem Map

Three buyers, three buying jobs, three hidden objections. The hidden objection is what makes the sale.

JTBDQuant at a macro fundWealth advisor at a private bankDAO treasurer / agent driver
The jobShip a strategy idea into productionJustify the advisory fee as AI commoditises adviceAllocate across on-chain and off-chain at agent speed
Current solutionMemo → 4 meetings → dev queue → weeksBespoke planning + tax structuringSpreadsheets + group chats + opportunistic votes
TriggerIdea decays before it shipsClient compares the fee to a free AI advisorYield-arb window closes inside four hours
Hidden objection (Sutherland)"If I make agents easy, my seat disappears""If I admit AI is better, I was overpaid for twenty years""If we automate, the multi-sig signers lose status"
Progress they are hiring forVelocity and status — shipping fast is winningStay relevant in the client's mind through the transitionDefensible alpha + safety at machine speed

The hidden objection is the work. Every agent-enablement sale into finance is — underneath the deck — a sale that reassures the buyer their seat survives the transition. Sell the seat first; then sell the agent.

STAR Pages

These pages go deeper on specific players:

  • Venture Capital — How AI and crypto reshape who gets funded and on what terms
  • Wealth Management — Preservation strategies; HNW planning across jurisdictions
  • Participatory Capital — DePIN ownership, tokenised participation, permissionless investment vehicles

Context

Questions

Which player in this map is most likely to capture the AI-augmented research edge first — and which is most likely to deny it the longest?

  • At what point does an AI agent managing a portfolio become a regulated fund manager?
  • Which DeFi protocol is closest to replacing a traditional correspondent banking relationship?
  • If "agent-as-principal" became the standard, which of today's entity types would absorb it — and which would dissolve?
  • What is the hidden objection your firm has not yet named about handing the workflow to an agent?