Investor Shelf

Who Answers for the Agent? Agentic AI, Delegated Execution, and Accountability in the Corporate Finance Function (Perspectives in Social Sciences)

Research paperPerspectives in Social SciencesNovember 202630 pages45 min readSayed Mahmood Sadat, Hizar khanOriginal source
Key number
USD 400 million
Knight Capital execution failure loss p.2
Largest segment
147 records
Records excluded due to prediction-only AI role p.5
What changed
67 sources
Academic literature sources retained after screening p.4

Summary

  • The paper presents the Algorithmic Financial Accountability Framework (AFAF) anchored on 9 core propositions to address governance in AI delegation.p.1
  • An automated execution failure at Knight Capital on 1 August 2012 resulted in a loss of more than USD 400 million within about 45 minutes.p.2
  • A structured database search conducted in month of January 2026 returned 437 records for screening.p.4
  • Following a two-stage screening process, 67 sources met all inclusion criteria and were retained for synthesis.p.4
  • Out of 370 excluded records, 147 were removed because AI was used solely for prediction or analysis without a decision role.p.5
  • The framework categorizes corporate AI interaction into 3 levels of authority: informational, decisional, and execution autonomy.p.6

Key insights

  1. Agentic AI systems differ from earlier rule-based automation by interpreting general goals, selecting tools at run time, acting non-deterministically, and drifting post-vendor updates.p.6
  2. AFAF differentiates AI permissions into informational autonomy, decisional autonomy, and execution autonomy based on the power to change financial commitments.p.6
  3. Agentic financial autonomy exhibits an inverted-U relationship with financial decision quality, where performance benefits decline beyond an optimal threshold.p.17
  4. Greater agentic financial autonomy directly increases a firm's algorithmic financial risk exposure.p.18
  5. Effective managerial oversight weakens the positive relationship between agentic financial autonomy and algorithmic financial risk exposure.p.18
  6. AI governance maturity mitigates the negative impact of agentic financial autonomy on corporate financial accountability.p.19
  7. Decision irreversibility strengthens risk exposure and lowers the level of autonomy at which financial decision quality is maximized.p.19
  8. Concentration of third-party AI service providers amplifies risk exposure, especially when external providers are difficult to substitute.p.20
  9. A positive autonomy–governance gap, occurring when autonomy deployment outpaces governance reconfiguration capabilities, harms subsequent corporate financial accountability.p.20

Topics and entities

Summary and figures generated from the document and checked automatically: every number shown appears on the cited page. The full document is available from the publisher. Publishers can request removal.