USE CASES
What AI agents do with Abacus Labs
Seven things people ask their agents to do with a portfolio, and what Abacus Labs returns for each: options with tradeoffs, then a trade list checked against every rule.
Every page follows the same shape: the problem a language model cannot solve alone, the conversation, the JSON Abacus Labs returns, how the engine handles it, and who uses it. Example accounts and numbers are labelled as examples. Literature numbers link to their sources.
- Tax-loss harvesting for AI agentsFind loss lots, pick wash-sale-safe replacements, return checked trades with the tax effect.
- Portfolio rebalancing for AI agentsThe fewest trades back to target, with tax cost, drift and risk before and after.
- Minimizing portfolio volatility for AI agentsTurn "make it less risky" into a measured target and price each option in tax and turnover.
- Thematic investing for AI agents: turn a view into a portfolioMeasure the exposure the account already has, size the tilt, keep the rest of the risk in place.
- Concentrated stock de-risking for AI agentsTrim with tax, trim using losses, or diversify around the position. Each with lots and risk left.
- Pre-trade checks for AI agents: wash sales, whole shares, cashSend an LLM-drafted trade list. Get pass, or every violation and a repaired list that keeps the intent.
- Portfolio exposure and risk analytics for AI agentsSectors, factors, single-name risk, fund overlap and unrealized gains for any linked account.
Not listed: raising cash for a purchase, investing new cash, swapping one holding for another, and moving an account in kind. They use the same propose and commit calls with a different goal string. See the docs for the request types.