■ CRITICAL RISK ■ Finance
For execution, the replacement already happened — equities trade algorithmically and the floor is a museum. What survives isn't trading as most people picture it, but a thin layer of humans supervising machines and negotiating the trades too big or weird for them.
“Algorithms trade in microseconds. Your gut feeling takes minutes.”
Our AI replacement risk score — how we score jobs
Start with what's already gone: the shouting floor, the bank of phones, the human market-maker quoting spreads. Equity execution is overwhelmingly electronic and has been for years — algorithms slice institutional orders into thousands of micro-trades to minimize market impact, market-making firms quote continuously in microseconds, and a human simply cannot compete on the timescale where modern liquidity lives. Trading desks that employed rows of execution traders now run a fraction of the headcount supervising algorithms, and each downturn ratchets the number lower.
Today's human trader is better described as an algorithm operator and exception handler: selecting execution strategies, monitoring fills, intervening when markets dislocate, and handling blocks so large they need discretion and relationships to move without spooking the market. Less-electronic corners — parts of credit, munis, exotic derivatives, distressed debt — still run on negotiation and dealer relationships, though electronification is grinding through fixed income too, corner by corner. Systematic funds pushed the same logic upstream: strategy itself, not just execution, is increasingly model-driven, with machine learning mining signals no human would find or trust.
What resists is judgment at the edges: crisis moments when models trained on normal markets meet abnormal ones, block trades where counterparty trust is the product, structuring bespoke instruments, and the accountability regulators want attached to a person. Sales-trading survives where clients want a human explaining what's happening to their order. But these are senior, scarce seats — the apprenticeship rungs below them were the jobs execution algorithms deleted. Our risk score of 87 describes an occupation where the machines aren't coming for the job; they took it, and the survivors are the ones holding the leash.
Automatability: our editorial assessment of current and near-term AI capability
Mostly past tense: electronic execution conquered equities years ago and headcount fell with it; fixed income and derivatives are electronifying now, corner by corner, through the late 2020s. The remaining human seats — algo supervision, blocks, illiquid products — keep thinning as machine learning moves upstream into strategy. New entrants should assume the trading career of finance lore no longer exists to enter.
Yes, but the job title misleads. Modern desk humans supervise execution algorithms, handle block trades needing discretion, and cover clients — they rarely execute routine orders themselves, because software does it faster and cheaper. The armies of execution traders from the 1990s are gone, and the surviving seats are senior, quantitative, or relationship-driven. The middle of the old career ladder is missing.
Not at speed, and mostly not at signal-finding either — systematic funds process more data than any discretionary trader can. Where humans retain an edge: illiquid markets where prices are negotiated rather than quoted, regime breaks where models trained on normal conditions misfire, and situations where trust and accountability are part of the trade. That edge employs hundreds, not tens of thousands.
The traditional path — join a desk, learn to execute, work up to risk-taking — has largely evaporated. What's hiring is adjacent: quantitative research, algo development, execution consulting, and risk management. If you want markets exposure, build quantitative and programming skills; the desks that remain hire people who can read both a market and a codebase. Pure gut-feel trading is a nostalgia product.
The classic migrations: sales and client coverage, risk management, fintech and trading-technology firms, portfolio management, and — for a fortunate cohort — retirement on the winnings. Banks' execution desks shrank over years of attrition rather than one dramatic purge, which is why the transformation got less attention than factory automation despite being more complete. It's the white-collar automation case study nobody staged a documentary about.