{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,14]],"date-time":"2026-03-14T05:01:36Z","timestamp":1773464496867,"version":"3.50.1"},"reference-count":98,"publisher":"Annual Reviews","issue":"1","license":[{"start":{"date-parts":[[2025,5,5]],"date-time":"2025-05-05T00:00:00Z","timestamp":1746403200000},"content-version":"unspecified","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,5,5]]},"abstract":"<jats:p>The development of software agents that can autonomously take actions to achieve goals has been a long-standing foundational objective in the field of AI. Recent advances in generative AI have given rise to a new class of agents. These advances have opened up the possibility of developing agents that can augment knowledge work in finance that primarily involves the cognitive processing of information by skilled humans. In this article, we bring these fields together. We break down the specific challenges in knowledge work in finance, review the current literature on generative AI agents, and identify potential directions for research and development that would help realize the potential of generative AI agents in finance. We conclude by proposing a framework to delineate the levels of autonomy for AI agents in the context of knowledge work, and an architecture for human\u2013AI collaboration that can pave the path for progressively increasing the autonomy of agents.<\/jats:p>","DOI":"10.1146\/annurev-control-022823-033813","type":"journal-article","created":{"date-parts":[[2024,12,16]],"date-time":"2024-12-16T19:40:55Z","timestamp":1734378055000},"page":"189-210","source":"Crossref","is-referenced-by-count":5,"title":["Generative AI Agents for Knowledge Work Augmentation in Finance"],"prefix":"10.1146","volume":"8","author":[{"given":"Sumitra","family":"Ganesh","sequence":"first","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]},{"given":"Leo","family":"Ardon","sequence":"additional","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]},{"given":"Daniel","family":"Borrajo","sequence":"additional","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]},{"given":"Deepeka","family":"Garg","sequence":"additional","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]},{"given":"Udari Madhushani","family":"Sehwag","sequence":"additional","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]},{"given":"Annapoorani Lakshmi","family":"Narayanan","sequence":"additional","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]},{"given":"Giuseppe","family":"Canonaco","sequence":"additional","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]},{"given":"Manuela M.","family":"Veloso","sequence":"additional","affiliation":[{"name":"JPMorgan AI Research, New York, NY, USA; email:\u00a0sumitra.ganesh@jpmorgan.com"}]}],"member":"22","reference":[{"key":"B1","volume-title":"Landmarks of Tomorrow","year":"1959"},{"issue":"2","key":"B2","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1017\/S0269888900008122","article-title":"Intelligent agents: theory and practice","volume":"10","year":"1995","journal-title":"Knowl. Eng. Rev."},{"key":"B3","doi-asserted-by":"crossref","first-page":"345","DOI":"10.1023\/A:1008942012299","article-title":"Multiagent systems: a survey from a machine learning perspective","volume":"8","year":"2000","journal-title":"Auton. Robots"},{"issue":"7","key":"B4","doi-asserted-by":"crossref","first-page":"72","DOI":"10.1145\/176789.176797","article-title":"A softbot-based interface to the Internet","volume":"37","year":"1994","journal-title":"Commun. ACM"},{"key":"B5","article-title":"Sparks of artificial general intelligence: early experiments with GPT-4","year":"2023"},{"key":"B6","article-title":"A survey on multimodal large language models","year":"2023"},{"key":"B7","article-title":"Augmented language models: a survey","year":"2024","journal-title":"Trans. Mach. Learn. Res."},{"key":"B8","first-page":"1877","article-title":"Language models are few-shot learners","volume-title":"Advances in Neural Information Processing Systems 33","year":"2020"},{"issue":"6","key":"B9","doi-asserted-by":"crossref","first-page":"186345","DOI":"10.1007\/s11704-024-40231-1","article-title":"A survey on large language model based autonomous agents","volume":"18","year":"2024","journal-title":"Front. Comput. Sci."},{"key":"B10","article-title":"Large language model based multi-agents: a survey of progress and challenges","year":"2024"},{"key":"B11","article-title":"Rethinking knowledge work: a strategic approach","volume-title":"McKinsey Quarterly, Feb. 1","year":"2011"},{"issue":"6","key":"B12","doi-asserted-by":"crossref","first-page":"553","DOI":"10.1016\/j.ergon.2012.09.001","article-title":"Task complexity: a review and conceptualization framework","volume":"42","year":"2012","journal-title":"Int. J. Ind. Ergon."},{"issue":"6","key":"B13","doi-asserted-by":"crossref","first-page":"633","DOI":"10.1111\/j.1937-5956.2010.01163.x","article-title":"Operations in financial services\u2014an overview","volume":"19","year":"2010","journal-title":"Prod. Oper. Manag."},{"key":"B14","article-title":"Cognitive architectures for language agents","year":"2023"},{"key":"B15","article-title":"Building cooperative embodied agents modularly with large language models","year":"2024","journal-title":"The Twelfth International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent."},{"key":"B16","article-title":"ChatDev: communicative agents for software development","year":"2024"},{"key":"B17","first-page":"8634","article-title":"Reflexion: language agents with verbal reinforcement learning","volume-title":"Advances in Neural Information Processing Systems 36","year":"2024"},{"issue":"1","key":"B18","doi-asserted-by":"crossref","first-page":"216","DOI":"10.1016\/j.ejor.2012.12.018","article-title":"Combining linear programming and automated planning to solve intermodal transportation problems","volume":"227","year":"2013","journal-title":"Eur. J. Oper. Res."},{"issue":"1","key":"B19","doi-asserted-by":"crossref","first-page":"8","DOI":"10.1109\/MIS.2004.1265878","article-title":"MAPGEN: mixed-initiative planning and scheduling for the Mars Exploration Rover mission","volume":"19","year":"2004","journal-title":"IEEE Intell. Syst."},{"issue":"2","key":"B20","doi-asserted-by":"crossref","first-page":"389","DOI":"10.1016\/j.eswa.2006.05.027","article-title":"Integrating planning and scheduling in workflow domains","volume":"33","year":"2007","journal-title":"Expert Syst. Appl."},{"key":"B21","first-page":"1","article-title":"Planning for tourism routes using social networks","volume":"69","year":"2017","journal-title":"Expert Syst. Appl."},{"key":"B22","doi-asserted-by":"crossref","first-page":"61","DOI":"10.1613\/jair.1129","article-title":"PDDL2.1: an extension to PDDL for expressing temporal planning domains","volume":"20","year":"2003","journal-title":"J. Artif. Intell. Res."},{"key":"B23","volume-title":"Automated Planning: Theory and Practice","year":"2004"},{"key":"B24","doi-asserted-by":"crossref","first-page":"24824","DOI":"10.52202\/068431-1800","article-title":"Chain-of-thought prompting elicits reasoning in large language models","volume-title":"Advances in Neural Information Processing Systems 35","year":"2022"},{"key":"B25","doi-asserted-by":"crossref","first-page":"22199","DOI":"10.52202\/068431-1613","article-title":"Large language models are zero-shot reasoners","volume-title":"Advances in Neural Information Processing Systems 35","year":"2022"},{"key":"B26","article-title":"Least-to-most prompting enables complex reasoning in large language models","year":"2023","journal-title":"The Eleventh International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent."},{"key":"B27","article-title":"Self-consistency improves chain of thought reasoning in language models","year":"2023","journal-title":"The Eleventh International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent."},{"key":"B28","article-title":"ReAct: synergizing reasoning and acting in language models","year":"2023","journal-title":"The Eleventh International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent."},{"key":"B29","article-title":"Large language models cannot self-correct reasoning yet","year":"2024","journal-title":"The Twelfth International Conference on Learning Representations"},{"key":"B30","first-page":"1769","article-title":"Inner monologue: embodied reasoning through planning with language models","year":"2022","journal-title":"Proceedings of the 6th Conference on Robot Learning"},{"key":"B31","first-page":"11809","article-title":"Tree of thoughts: deliberate problem solving with large language models","volume-title":"Advances in Neural Information Processing Systems 36","year":"2024"},{"key":"B32","first-page":"8154","article-title":"Reasoning with language model is planning with world model","year":"2023","journal-title":"Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing"},{"key":"B33","article-title":"PlanBench: an extensible benchmark for evaluating large language models on planning and reasoning about change","year":"2023"},{"issue":"18","key":"B34","first-page":"20256","article-title":"Generalized planning in PDDL domains with pretrained large language models","volume":"38","year":"2024","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"B35","article-title":"TRIP-PAL: travel planning with guarantees by combining large language models and automated planners","year":"2024"},{"key":"B36","first-page":"432","article-title":"Large language models as planning domain generators","year":"2024","journal-title":"Proceedings of the Thirty-Fourth International Conference on Automated Planning and Scheduling"},{"key":"B37","article-title":"On the prospects of incorporating large language models (LLMs) in automated planning and scheduling (APS)","year":"2024"},{"key":"B38","doi-asserted-by":"crossref","first-page":"79081","DOI":"10.52202\/075280-3459","article-title":"Leveraging pre-trained large language models to construct and utilize world models for model-based task planning","volume-title":"Advances in Neural Information Processing Systems 36","year":"2023"},{"key":"B39","article-title":"LLM+P: empowering large language models with optimal planning proficiency","year":"2023"},{"key":"B40","article-title":"Automatically correcting large language models: surveying the landscape of diverse self-correction strategies","year":"2023"},{"key":"B41","article-title":"CRITIC: Large language models can self-correct with tool-interactive critiquing","year":"2024"},{"key":"B42","article-title":"Improving code generation by training with natural language feedback","year":"2024"},{"key":"B43","first-page":"622","article-title":"OPEx: a component-wise analysis of LLM-centric agents in embodied instruction following","year":"2024","journal-title":"Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics"},{"key":"B44","article-title":"Finetuned language models are zero-shot learners","year":"2022","journal-title":"The Tenth International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent."},{"key":"B45","first-page":"2986","article-title":"LLM-Planner: few-shot grounded planning for embodied agents with large language models","year":"2022","journal-title":"2023 IEEE\/CVF International Conference on Computer Vision"},{"key":"B46","article-title":"O3D: offline data-driven discovery and distillation for sequential decision-making with large language models","year":"2024","journal-title":"First Conference on Language Modeling. N.p.: COLM Organ. Comm"},{"key":"B47","article-title":"Self-refine: iterative refinement with self-feedback","year":"2023"},{"key":"B48","article-title":"WorldCoder, a model-based LLM agent: building world models by writing code and interacting with the environment","year":"2024"},{"issue":"17","key":"B49","first-page":"19632","article-title":"Expel: LLM agents are experiential learners","volume":"38","year":"2024","journal-title":"Proc. AAAI Conf. Artif. Intell."},{"key":"B50","article-title":"Voyager: an open-ended embodied agent with large language models.","year":"2023"},{"key":"B51","article-title":"OS-Copilot: towards generalist computer agents with self-improvement","year":"2024"},{"key":"B52","article-title":"When not to trust language models: investigating effectiveness of parametric and non-parametric memories","year":"2023"},{"key":"B53","article-title":"MetaTool benchmark for large language models: deciding whether to use tools and which to use","year":"2024"},{"key":"B54","article-title":"On the tool manipulation capability of open-source large language models","year":"2023"},{"key":"B55","article-title":"Check your facts and try again: improving large language models with external knowledge and automated feedback","year":"2023"},{"key":"B56","article-title":"Tool learning with foundation models","year":"2024"},{"key":"B57","first-page":"43447","article-title":"Chameleon: plug-and-play compositional reasoning with large language models","volume-title":"Advances in Neural Information Processing Systems 36","year":"2024"},{"key":"B58","doi-asserted-by":"crossref","first-page":"0063","DOI":"10.34133\/icomputing.0063","article-title":"TaskMatrix.AI: completing tasks by connecting foundation models with millions of APIs","volume":"3","year":"2024","journal-title":"Intell. Comput."},{"key":"B59","article-title":"Large language models for information retrieval: a survey","year":"2024"},{"issue":"4","key":"B60","first-page":"1413","article-title":"A survey of knowledge enhanced pre-trained language models","volume":"36","year":"2023","journal-title":"IEEE Trans. Knowledge Data Eng."},{"key":"B61","article-title":"CRAFT: customizing LLMs by creating and retrieving from specialized toolsets","year":"2024"},{"key":"B62","article-title":"ToolAlpaca: generalized tool learning for language models with 3000 simulated cases","year":"2023"},{"key":"B63","article-title":"ART: automatic multi-step reasoning and tool-use for large language models","year":"2023"},{"key":"B64","article-title":"Chain-of-thoughts prompting with language models for accurate math problem-solving","year":"2023","journal-title":"2023 IEEE MIT Undergraduate Research Technology Conference"},{"key":"B65","doi-asserted-by":"crossref","first-page":"525","DOI":"10.1038\/s42256-024-00832-8","article-title":"Augmenting large language models with chemistry tools","volume":"6","year":"2024","journal-title":"Nat. Mach. Intell."},{"key":"B66","article-title":"WebGPT: browser-assisted question-answering with human feedback","year":"2022"},{"key":"B67","article-title":"Social-LLM: modeling user behavior at scale using language models and social network data","year":"2023"},{"key":"B68","article-title":"User modeling in the era of large language models: current research and future directions","year":"2023"},{"key":"B69","article-title":"Visual ChatGPT: talking, drawing and editing with visual foundation models","year":"2023"},{"key":"B70","article-title":"RT-2: vision-language-action models transfer web knowledge to robotic control","year":"2023"},{"key":"B71","first-page":"9118","article-title":"Language models as zero-shot planners: extracting actionable knowledge for embodied agents","year":"2022","journal-title":"Proceedings of the 39th International Conference on Machine Learning"},{"key":"B72","first-page":"28091","article-title":"Mind2Web: towards a generalist agent for the web","volume-title":"Advances in Neural Information Processing Systems 36","year":"2024"},{"key":"B73","first-page":"4952","article-title":"SheetCopilot: bringing software productivity to the next level through large language models","volume-title":"Advances in Neural Information Processing Systems 36","year":"2024"},{"key":"B74","article-title":"SheetAgent: a generalist agent for spreadsheet reasoning and manipulation via large language models","year":"2024"},{"key":"B75","article-title":"Data-Copilot: bridging billions of data and humans with autonomous workflow","year":"2024"},{"key":"B76","first-page":"10764","article-title":"PAL: program-aided language models","year":"2023","journal-title":"Proceedings of the 40th International Conference on Machine Learning"},{"key":"B77","article-title":"Executable code actions elicit better LLM agents","year":"2024"},{"key":"B78","article-title":"MetaGPT: meta programming for multi-agent collaborative framework","year":"2024","journal-title":"The Twelfth International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent"},{"key":"B79","article-title":"AgentVerse: facilitating multi-agent collaboration and exploring emergent behaviors","year":"2024","journal-title":"The Twelfth International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent"},{"key":"B80","doi-asserted-by":"crossref","first-page":"51991","DOI":"10.52202\/075280-2264","article-title":"CAMEL: communicative agents for \u201cmind\u201d exploration of large language model society","volume-title":"Advances in Neural Information Processing Systems 36","year":"2023"},{"key":"B81","article-title":"MAC-SQL: a multi-agent collaborative framework for text-to-SQL","year":"2024"},{"key":"B82","article-title":"MAGIC: generating self-correction guideline for in-context text-to-SQL","year":"2024"},{"key":"B83","article-title":"RoCo: dialectic multi-robot collaboration with large language models","year":"2023"},{"key":"B84","article-title":"Embodied LLM agents learn to cooperate in organized teams","year":"2024"},{"key":"B85","article-title":"Improving factuality and reasoning in language models through multiagent debate","year":"2023"},{"key":"B86","article-title":"Encouraging divergent thinking in large language models through multi-agent debate","year":"2024"},{"key":"B87","article-title":"Unleashing cognitive synergy in large language models: a task-solving agent through multi-persona self-collaboration","year":"2023"},{"key":"B88","article-title":"ChatEval: towards better LLM-based evaluators through multi-agent debate","year":"2024","journal-title":"The Twelfth International Conference on Learning Representations. La Jolla, CA: Int. Conf. Learn. Represent"},{"key":"B89","article-title":"Generative agents: interactive simulacra of human behavior","year":"2023","journal-title":"Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology"},{"key":"B90","article-title":"Is this the real life? Is this just fantasy? The misleading success of simulating social interactions with LLMs","year":"2024"},{"key":"B91","article-title":"SOTOPIA: interactive evaluation for social intelligence in language agents","year":"2024"},{"key":"B92","article-title":"Sotopia-\u03c0: interactive learning of socially intelligent language agents","year":"2024"},{"key":"B93","article-title":"Can LLMs keep a secret? Testing privacy implications of language models via contextual integrity theory","year":"2024"},{"key":"B94","article-title":"Deal or no deal?","volume-title":"End-to-end learning for negotiation dialogues","year":"2017"},{"key":"B95","article-title":"How well can LLMs negotiate? NegotiationArena platform and analysis","year":"2024"},{"key":"B96","article-title":"Rehearsal: simulating conflict to teach conflict resolution","year":"2024"},{"key":"B97","first-page":"279","article-title":"Multi-agent team formation: diversity beats strength?","year":"2013","journal-title":"Proceedings of the Twenty-Third International Joint Conference on Artificial Intelligence"},{"key":"B98","first-page":"136","article-title":"Teamwork in cyberspace: using TEAMCORE to make agents team-ready","year":"1999","journal-title":"AAAI Spring Symposium on Agents in Cyberspace"}],"container-title":["Annual Review of Control, Robotics, and Autonomous Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.annualreviews.org\/content\/journals\/10.1146\/annurev-control-022823-033813?crawler=true&mimetype=application\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,14]],"date-time":"2026-03-14T04:16:39Z","timestamp":1773461799000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.annualreviews.org\/content\/journals\/10.1146\/annurev-control-022823-033813"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,5,5]]},"references-count":98,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2025,5,5]]}},"URL":"https:\/\/doi.org\/10.1146\/annurev-control-022823-033813","relation":{},"ISSN":["2573-5144"],"issn-type":[{"value":"2573-5144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,5,5]]}}}