Talk:AI agent/Archive 1

Archive 1

Redundant article?

Isn't this article redundant as it covers the same topic as Intelligent agent? This lemma also used to redirect to said article. Maxeto0910 (talk) 14:56, 16 February 2025 (UTC)

My concern is that this article is both redundant and trying to "make a definition happen" based on an interpretation and use of terminology that isn't necessarily consensus in the industry. The first source itself says that there's widespread disagreement in use of this terminology, and some of the other sources express the same notion. My suggestion would be that the Intelligent Agents page should reference the ongoing debate around what constitutes an "AI agent", but not attempt to settle that debate by virtue of a definition being accepted here (effectively attempting to create a primary source).

To some extent, this is already covered in the "Alternative definitions and uses" section of that page. BrandonLive (talk) 20:10, 24 March 2025 (UTC)

Future Trends and Research Directions of Agentic AI

Hello, fellow editors.

I've added the "Future Trends and Research Directions" section to include several promising areas where Agentic AI is projected to make substantial advances. This includes future advancements in the combination of quantum computing and agentic AI, the expanding significance of agentic AI in personalized medicine and autonomous transportation, and its impact on creativity and design. Furthermore, I emphasized the significance of ethical governance in ensuring that the deployment of agentic AI is consistent with social norms.

Please evaluate these additions, and I appreciate any thoughts or input to further improve this section. Lavanyasekar11 (talk) 00:30, 5 April 2025 (UTC)

Real-time applications and Ethical considerations

Hello Wikipedians,

After reviewing the article, I felt it would be helpful to include some real-world applications to the applications section where Agentic AI has been successfully implemented. These real-time applications give an overview of how this systems work in real-time situations. I’ve also added relevant web links and news sources that provide further context for each application mentioned.

Additionally, I noticed a significant gap in the article regarding ethical considerations. As we discuss the development and deployment of Agentic AI, it’s equally important to address the ethical implications. So, I have included a separate section on " ethical considerations" wherein few "AI gone wrong" situations has been listed. These examples serve to underscore the importance of integrating ethical safeguards in AI development.

Please take a moment to review my contributions and share your feedback. I’m excited to collaborate and work toward making this article a truly informative and engaging read. Sivapriyagopi27 (talk) 03:24, 5 April 2025 (UTC)

Planned Improvements for Agentic AI Article

Planned Improvements

While the article on Agentic AI presents worthwhile observations about the topic, there are several areas which can be expanded and improved upon. We plan to improve the article by closing the following gaps and including significant references to supplement the content:

1. Explanation of Key Concepts: The article merely touches on how agentic AI works by independently making choices through continuous learning. It does not, however, contain a detailed examination of the specific machine learning models,patterns, and technologies propelling agentic AI. We plan to explore more into the core AI approaches such as reinforcement learning, deep learning, and unsupervised learning that enable AI agents to be independent. This would provide readers with a better understanding of the system dynamics.

2. Comprehensive History and Milestones The history section can be augmented to provide a fuller chronology of the evolution of agentic AI. While it provides a nod to early concepts credited to Alan Turing and Norbert Wiener, it does not address the influence of more recent breakthroughs, including the breakthrough in machine learning, reinforcement learning, and autonomous systems' rise in focused industries. We will add tangible milestones such as integrating AI into autonomous cars, industrial automation and other novel applications.

3. Specific Use Cases and Applications: The paper provides a general description of applications like software development, cybersecurity, and customer service, but it could be enhanced by a more in-depth analysis of these applications. We plan to include case studies of companies or technologies that have successfully implemented agentic AI,such as the ways in which AI agents are used in financial prediction or predictive maintenance in manufacturing. This will not only make the applications section theoretical but based on real-world examples.

4. Ethical Implications and Challenges: An important one lacking now is the ethical implications of agentic AI. With autonomous decision-making, issues of bias, accountability, and transparency come up.We will include a section on how these ethical challenges are being tackled in developing agentic AI.This may involve input from recent literature and existing debates in the field.

5. Future Directions and Trends: Although the article names Agentic AI as an emerging technology to keep an eye on in 2025, it does not explore thoroughly the future direction of the field. We will take a look at the predictions of industry insiders, research centers, and academic reports to develop an impression of the likely trend of the next couple of years for agentic AI. This includes the potential to merge AI with quantum computing, the role of agentic AI in personalized medicine, and more.

Niketha Sabesan (talk) 21:48, 2 April 2025 (UTC)

@Niketha Sabesan: Wikipedia strongly discourages people from sharing accounts. When you say "we", does that mean you using a shared account? Also if you add content, please make sure it's properly sourced and not original research. STEMinfo (talk) 02:12, 23 May 2025 (UTC)

Addition of Frameworks and Standards section

I recently added a new section titled "Frameworks and Standards" to the Agentic AI article. This section introduces the AI Agent Capabilities Periodic Table (AIA CPT), a structured framework developed by the Digital Twin Consortium. It provides a neutral, non-commercial summary of how agent capabilities are classified across six functional domains and five levels of autonomy.

The section also includes two real-world testbeds as supporting examples:

The Automated Negotiation Digital Twins testbed, which uses multi-agent systems to coordinate negotiation across organizational boundaries in supply chain and logistics scenarios.

The Metal 3D Printing and Optimization testbed, which applies autonomous agents in a closed-loop digital twin environment for adaptive quality control in additive manufacturing.

All references are publicly accessible and sourced directly from the Digital Twin Consortium. The content has been written in accordance with Wikipedia’s neutrality and verifiability guidelines. I welcome any suggestions for improvement or sourcing enhancements. — Preceding unsigned comment added by Wooty101 (talk • contribs) 22:51, 30 June 2025 (UTC)

Stop spamming. Count Count (talk) 10:30, 1 July 2025 (UTC)

I appreciate your concern. My intent was not to promote anything, but to provide sourced and neutral documentation of a non-commercial framework and associated academic/industry testbeds relevant to agentic AI implementation.

The Digital Twin Consortium is a non-profit standards body, and the testbeds cited include contributions from academic institutions like Rowan University. I agree that Wikipedia must avoid promotional tone, and I’m happy to rephrase or remove anything that reads as advocacy or lacks independent sourcing.

If you or other editors have suggestions on how to improve neutrality or balance the section further, I would welcome that. My goal is to contribute to a more grounded and useful article, and I want to make sure all edits meet Wikipedia’s standards.

Thank you again for flagging this — I take your feedback seriously and would appreciate your input on how to strengthen it Wooty101 (talk) 00:04, 2 July 2025 (UTC)

So you want to add a section about one framework and standard proposed by one organization. supported in the article entirely by a publication from that organization? Are there secondary sources discussing this standard? It seems pretty clearly promotional but I'm happy to be proven wrong. Oblivy (talk) 03:23, 2 July 2025 (UTC)

@Oblivy: Thank you for your question. I understand your concern around referencing only one framework from a single organization. I'm still fairly new to editing on Wikipedia and learning how to improve my contributions, so I appreciate the chance to refine this.

The section was added to reflect the fact that frameworks for Agentic AI are still in early development. The AIA CPT from the Digital Twin Consortium is one of the first attempts to categorize agent capabilities and types in a structured way. It's been applied in academic and industrial testbeds and is publicly maintained, but I agree it shouldn't be the only perspective if others exist.

I’ve been researching additional frameworks that could be added for balance. So far, I’ve found:

• A classification model in a recent arXiv paper distinguishing traditional AI agents from generative, goal-driven systems
• The five-level enterprise agent maturity model from sema4.ai
• Google’s Agent-to-Agent (A2A) protocol for inter-agent communication
• The INTERFACE-EU initiative, which explores autonomy and governance for agent teams
• Anthropic’s Model Context Protocol (MCP), which is aimed at tool and memory integration for LLM-based agent

These examples suggest that the field is evolving quickly and multiple groups are working toward formal definitions. I'm happy to help expand the section with some of these additional efforts and ensure everything is properly sourced. If you have preferences for which directions we should take, I’d be glad to collaborate.

Wooty101 (talk) 00:21, 3 July 2025 (UTC)

Besides expanding to include other frameworks, it would be very helpful (though not strictly necessary, as it's a quickly evolving field) to cite sources that compare the various frameworks. This is what I meant by WP:SECONDARY. Michaelmalak (talk) 12:09, 3 July 2025 (UTC)

I agree that citing secondary sources comparing agentic AI frameworks would improve the section. The field is still early, and those comparisons are limited for now. The intent behind adding the AIA CPT from the Digital Twin Consortium was to document one of the first public attempts to classify agent capabilities in a structured way. I understand the concern that it's from a single organization, and I’m working on expanding the section with additional efforts.

Some frameworks I'm exploring for inclusion:

• Sema4.ai’s Enterprise Agent Maturity Model
• Anthropic’s Model Context Protocol (MCP)
• Google’s A2A Protocol
• The INTERFACE-EU initiative
• Classification approaches like arXiv:2505.10468v4

Where possible, I’ll add secondary sources and clearly attribute each framework’s origin. To better reflect the current state of the field, I'm also open to renaming the section to "Emerging Frameworks for Agentic AI."

To support the article’s development more broadly, I’d also suggest adding:

• A comparison with related agent types
• A section on architectural patterns (e.g., observe–plan–act loops)
• Deployment and safety considerations, especially for industrial or LLM-based systems

Let me know if there's support for these additions, and I’m happy to help draft or source content for them. 2406:3400:41E:37F0:B8A4:EE4A:E469:B250 (talk) 03:41, 4 July 2025 (UTC)

After our discussion I have improved the previous addition of frameworks to include the following...

• • Frameworks section** sources:
• Government frameworks (NIST AI Risk Management Framework, EU AI Act standards)
• International standards bodies (IEEE, ISO/IEC, OECD)
• Multiple industry classification systems (Digital Twin Consortium, Salesforce, enterprise architecture experts)
• Academic research on agentic AI taxonomies
• Comparative analyses from independent sources

• • Industry Analysis section** sources:
• Gartner's 2025 Emerging Tech Impact Radar report
• Industry surveys on adoption challenges
• Market research on enterprise concerns and adoption timelines

This provides balanced coverage of both technical classification efforts and market perspective, helping readers understand the current state of the field from multiple angles. All content is well-sourced from independent, reliable publications.

Please see below for the content, and let me know if there are any issues with adding this contribution: — Preceding unsigned comment added by Wooty101 (talk • contribs) 00:41, 7 July 2025 (UTC)

Emerging Frameworks and Classification Systems

The rapid development of agentic AI has prompted efforts to establish classification systems and frameworks to distinguish these systems from traditional AI agents and guide their implementation.

Academic and Industry Taxonomies

Academic researchers have developed formal taxonomies to clarify agentic AI capabilities and levels of autonomy. Research distinguishes agentic AI systems from traditional single AI agents, highlighting characteristics such as orchestrated multi-agent structures, persistent memory systems, and proactive goal-driven behavior.

Industry Classification Systems

Several organizations have developed comprehensive frameworks for categorizing agentic AI capabilities and maturity levels:

Digital Twin Consortium Framework: The Digital Twin Consortium introduced the AI Agent Capabilities Periodic Table (AIA CPT) in 2025, defining 45 distinct capabilities organized into six core categories: Perception & Knowledge, Cognition & Reasoning, Learning & Adaptation, Action & Execution, Interaction & Collaboration, and Governance & Safety.

The framework establishes a five-level maturity classification:

Type	Description	Example Applications
0	Static Automation – Predefined responses without learning or adaptation	Rule-based control systems
1	Conversational Agents – Basic natural language interaction with simple context handling	Chatbots, voice assistants
2	Procedural Workflow Agents – Execute multi-step tasks using decision logic and tool integration	Workflow automation, RPA
3	Cognitive Autonomous Agents – Plan, reason, and learn from experience to make self-directed decisions	Predictive maintenance, optimization
4	Multi-Agent Generative Systems (MAGS) – Teams of collaborative agents with distributed reasoning and emergent behavior	Industrial orchestration, smart cities

The AIA CPT framework has been applied in industry testbeds, including automated negotiation systems and adaptive manufacturing process control.

Salesforce Maturity Model: Salesforce's Agentic AI Maturity Model outlines progression from rule-based chatbots through single-domain "co-pilots" to multi-agent orchestration systems where heterogeneous agents collaborate autonomously across enterprise technology stacks.

Enterprise Architecture Models: Enterprise architecture experts have proposed five-stage models ranging from "AI as a Tool" (basic deterministic automation) to "AI as an Autonomous Force" (AI-driven processes with near-complete autonomy), emphasizing the need for adaptive governance as organizations progress toward fully agentic AI.

Sema4.ai Enterprise Agent Maturity Model''': Sema4.ai has developed a five-level enterprise agent maturity model that provides a structured approach to evaluating and implementing AI agents within organizational contexts, focusing on enterprise readiness and deployment considerations.

Government and International Standards

NIST AI Risk Management Framework: The U.S. National Institute of Standards and Technology released the AI Risk Management Framework (AI RMF) in 2023, providing a voluntary framework for incorporating trustworthiness considerations into AI system design, development, and deployment. The framework establishes core functions to Govern, Map, Measure, and Manage AI risks while defining characteristics of trustworthy AI systems.

IEEE Standards: The Institute of Electrical and Electronics Engineers is developing multiple standards for autonomous and intelligent systems, including IEEE 7000 series standards for ethically aligned design, IEEE P3428 for Large Language Model Agents in education, and IEEE P2817 for verification of autonomous systems.

ISO/IEC Standards: The International Organization for Standardization has developed ISO/IEC 23053:2022, providing a framework for AI systems using machine learning, and is working on additional standards for AI system governance and risk management.

EU AI Act Standards: The European Union's AI Act implementation is supported by harmonized standards being developed by CEN-CENELEC Joint Technical Committee 21, focusing on AI trustworthiness, risk management, quality management systems, and conformity assessment procedures.

OECD AI Framework: The Organisation for Economic Co-operation and Development has established a framework for classifying AI systems across multiple dimensions including People & Planet, Economic Context, Data & Input, AI Model, and Task & Output, supporting policy-making and governance approaches.

Development Frameworks

Multiple frameworks have emerged for building agentic AI systems, each with distinct architectural approaches. Industry analyses highlight different design philosophies among frameworks such as LangChain's emphasis on modularity, CrewAI's focus on role-based collaboration, and Microsoft AutoGen's enterprise scalability features.

These frameworks provide predefined architectures, communication protocols, and integration tools to streamline agent development, with some specializing in particular domains such as industrial operations or conversational AI applications.

References

— Preceding unsigned comment added by Wooty101 (talk • contribs) 00:41, 7 July 2025 (UTC)

Industry Analysis

Market research firms have begun analyzing agentic AI as an emerging technology with significant disruptive potential. Gartner positioned agentic AI as a disruptive technology in their 2025 Emerging Tech Impact Radar report, placing it in the 6-8 year horizon for widespread enterprise adoption.

According to Gartner's analysis, agentic AI refers to "various architectures, techniques, and frameworks for creating single-agent or collaborative multiagent systems capable of unsupervised task execution." The research identifies enterprise concerns around trusted autonomy as a key factor affecting the adoption timeline.

The report identified several sample vendors working in the agentic AI space, including Adept, Beam, CrewAI, Induced.AI, Meta, Microsoft, Multi-On, Orby.AI, and XMPro, representing diverse approaches to agentic AI implementation across different industries and use cases.

Industry surveys indicate that reliability and oversight remain primary concerns for organizations considering agentic AI adoption, with many pursuing phased implementation approaches that maintain human supervision while gradually increasing agent autonomy.

References

— Preceding unsigned comment added by Wooty101 (talk • contribs) 00:41, 7 July 2025 (UTC)

"Applications" section is suspicious to me/citation issues

The first part of the "Applications" section (first 5 bullet points) is just a brief summary of the relevant paragraph in the article given as a reference. The reason the same source is repeated 5 times is because each point is a short summary of one paragraph of the article. They're even in the same order the article references them in its own text.

I'm not sure if this is a violation of any of Wikipedia's rules, but I feel this is sloppy, personally. To me, it seems like the original user had recently read an article and agreed with its points, so tried to use the article itself as the only citation for a good portion of this page. Currently it is not the only source for the "Applications" section, but it was when the page was created. This section would be improved by adding more sources for these individual use cases or some other manner to reduce dependence on a single source like this.

Regardless of whether citing individual bullet points of an article like this is within Wikipedia's rules, using a direct link to the article has also caused this page to display an inaccurate source. The article was updated between November 2024 and June 2025: this page cites the original version published in November 2024. However, the current article accessible by the link is named "Agentic AI: 9 promising use cases for business" and does not mention Forrester's assessment for 2025. As such, I have updated the citations to be hopefully more accurate/helpful. Augmented Reality Foliage (talk) 15:15, 13 August 2025 (UTC)

Restoring last version by Grayfell -- please discuss here before reverting

I have reinstated Grayfell's edit which were reverted by @Isolatedchimpanzee. Under the bold-revert-discuss framework for editor interactions, I think the revert was inappropriate. Grayfell explained their edit in an edit summary. Although it was certainly open to them to open a talk page comment I don't think it was required. However, once an objection was raised it was incumbent on @Isolatedchimpanzee to respond. They have not done so.

Further edits don't seem to be much better. The citation to Alan Turing strikes me as WP:OR since he was talking about a chatbot not an agent (and does not use the term "agent" at all). The Weiner cite is slightly more apposite but also strikes me as WP:OR and WP:SYNTH since the link to this topic is only supported by further logical steps.

Policies and guidelines encourage/require editors involved in a content dispute to seek WP:consensus on the talk page. I'm taking that step now. Oblivy (talk) 01:14, 27 August 2025 (UTC)

Its good that you brought the issue here. Since Grayfell reverted (and he did not start the discussion) the edit due to lack of WP:IS reference. So I think better I could do was to fix the edit and wait if Grayfell had to say about it if he had to, then I could have brought a discussion. because I am pretty leaned to this Idea: is one of many optional strategies that editors may use to seek consensus. This process is not mandated by Wikipedia policy, but it can be useful for identifying objections, keeping discussion moving forward and helping to break deadlocks. However since we are here, we could talk about this. 1rst revert: refs lack, I fixed it, since new MIT study is a notable research, as it is available in many third party journalism. 2nd edit was on history which had nothing to do with previous revert, it was because of the lack of Article's further explanation on History of Agentic AI that I chose to edit it. So I added it. For your question: I guess you are not confusing the Turing theory from the relation with modern Agentic ai? The sentence-the first idea of agentic AI took a term from the research published by Alan Turing in which he asked the question-"Can Machine Think?". might me misleading but it meant that the Turing theory of "AI intellgence" was a setting stage rather than the direct relation with Agentic ai, because his paper focused on the concept of works (and functions) similar to that today's agentic models perform, though not explicitly related or said there in the paper. That is exactly what later AI textbooks (e.g. Russell & Norvig, Artificial Intelligence: A Modern Approach) formalised as an agent. It was later added: Turing (1950) reframed intelligence as an observable behaviour test. That shifted focus from inner states to inputs, outputs, and interaction. Furthermore, Wiener's theory was also a setting state not a direct correlation with today modern agents. such as this one: His key concepts were closed-loop control where sensors could measure output and the controller were able to correct errors. Wiener proposed some practical artifact such as, thermostats, autopilots, Grey Walter’s “tortoises”, these were simple robots whose behaviour emerged from sensor–actuator feedback. So I think you should not confuse it with the idea that these people (and their own work) were the core reason in today Agentic AI development and how they exist. Isolatedchimpanzee (talk) 01:49, 27 August 2025 (UTC)

You could understand if you have more experience on the "history of Artificial intelligence" or "Machine learning"/ Isolatedchimpanzee (talk) 01:51, 27 August 2025 (UTC)

I don't understand what "pretty leaned to th[e] idea" that discussion is optional means. WP:CONSENSUS says: "According to the consensus policy, all edits should be explained (unless the reason for them is obvious)—either by clear edit summaries, or by discussion on the associated talk page" It appears you never use edit summaries, so perhaps you can start there (see WP:ES). As for the rest of it, I'm unconvinced. Did you even read WP:FORBESCON? What's your response? Have you considered the ideas I mentioned of no-original-research and no-synthesis, with respect to your two historical references?

Obviously you are passionate about improving this article but if you want your edits to stick I recommend you listen to what other editors are trying tell you. If you choose instead to hand-wave away your obligation to seek consensus then this pattern is likely to continue. Oblivy (talk) 02:04, 27 August 2025 (UTC)

Yes I don't (often) use edit summaries or rather chose not but don't say "never" without reading my Global contributions. As for WP:FORBESCON and I would like if you to tell me what do you mean by "even read it". But I guess you mean why I used forbes' new article refs? Then I think along with the Forbes I have added many other WP:IS-respecting sources, which includes MIT own research paper, which may be OR but is pretty relevant there.

Then your thought on historical references does not make any sense. What do you mean we are not supposed to use OR papers to cite them while claiming the proposed sentence if the sentence is related to that research paper? You should definitely read History of Artificial Intelligence. Do you know anything about when and when not to use WP:OR?

Well I think there is nothing to worry about me being passionate or not, because consensus is already happening. That is to say just like you started a discussion in my talk page after reverting my edit at the same place it was Grayfell's job to also notify me about the edit, which he did not do.

Ps. Pretty leaned to idea means that I dont usually start discussion if it not necessory, you should do some research on me before making claims that I never do consensus. Whatever. Isolatedchimpanzee (talk) 02:29, 27 August 2025 (UTC)

Did you read FORBESCON? It's a simple question. Oblivy (talk) 02:32, 27 August 2025 (UTC)

What exactly do you want with that WP page buddy? I have read it, just tell me. JEsus Isolatedchimpanzee (talk) 02:34, 27 August 2025 (UTC)

If you read it, then you know that Forbes Contributor articles like the one you added are considered deprecated sources. And yet you re-added the link even after it was brought to your attention.

As far as I know @Grayfell was under NO obligation to separately notify you of their edit but you can let me know from where you are getting this understanding and perhaps we can learn together. Oblivy (talk) 02:39, 27 August 2025 (UTC)

You know I am here actually thinking about doing a reserve osmosis on you for removing my historical section without having a strong point. Should I? Morever I did not "re-add" the Forbes page (which as per WP page is not reliable excpet if it was written by unless the article was written by a subject-matter expert.), because it must have been readded after revert, maybe I should have removed. I accept that as mistake and it could have been you who would have deleted the specific ref without reverting the whole edit. I usually delete or challenge the specific part of article when editing without bulk deleting whole part in its local.

Usually yes, he was not under any obligation, but it is usually expected from an editor to notify if they reverted or removed any part of other party's edit. Isolatedchimpanzee (talk) 02:53, 27 August 2025 (UTC)

No, that is not expected at all. MrOllie (talk) 02:59, 27 August 2025 (UTC)

As far as I know it is. And we usually notify (sometime) that we have reverted, removed or did something with your edit. Isolatedchimpanzee (talk) 03:00, 27 August 2025 (UTC)

It's the opposite - it is the person who is trying to add content who generally must start the discussion, per WP:BRD and WP:ONUS. You are very new here, please do not presume to tell experienced editors how things are usually done. MrOllie (talk) 03:02, 27 August 2025 (UTC)

Here's the diff showing you re-added the link.

@MrOllie has been known to make mistakes. So has @Grayfell and so have I. But all three of us? There was no obligation/expectation to separately notify you and I suggest you stop insisting there was. Oblivy (talk) 03:07, 27 August 2025 (UTC)

@Oblivy How many times do i have to tell I did not re-add link, it was due to revert. As far as I know I DID not re-add any link. Isolatedchimpanzee (talk) 03:19, 27 August 2025 (UTC)

That is not completely correct. What you telling is for when you are adding something in the article, however it is very loose rule and often optional. You generally don't have to go to talk page and start a discussion before making every edit in the article, that actually against "bold". I was talking about revert discussion policy not otherwise. The Oblivy guy must know that nobody is forcing him to notify, but when I said to do that I meant it is better to notify, however voluntarily. I do not think notifying about edit is anyway wrong idea. Moreover, ONUS is not a strict policy either

P.s. dont start "I have more edits than you ideology", please. Isolatedchimpanzee (talk) 03:15, 27 August 2025 (UTC)

There were a lot of weird issues with the content I removed.

As one example, this source is strange enough to be a red flag, and it was being over-cited and misused in the article in multiple ways. The citation had the wrong title for the story, which is flimsy and very speculative. It wasn't saying these are uses now, it's saying that according to AI executives, they will be used this way in the nebulous future. This is a business puff piece (and WP:CRYSTAL) with no obvious encyclopedic significance.

As another example of these problems, the sources on Albert Bandura do not mention AI at all. This is clear WP:SYNTH and is inappropriate. The info on psychology was not connected to this specific topic. Maybe it could be summarized better at Artificial consciousness or some other article, but it doesn't make sense here. It reads like padding because it probably is padding.

The entire article has WP:TONE problems like this. Unsurprisingly, these are also WP:AITELLS. "...has roots in...", "the core concept...", "Particularly...", "...is essential ...", and so on. The article reads like AI slop because it probably is AI slop. There are other problems here, but the gist is that we need humans to review and summarize better-quality sources which are primarily about "agentic AI". Grayfell (talk) 05:59, 27 August 2025 (UTC)

I think that is a neat critique that can be used for improving the article. Which also includes non AI tools dependency. T Isolatedchimpanzee (talk) 09:53, 27 August 2025 (UTC)

Merge proposal

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.

Partial merge completed. -- Beland (talk) 03:47, 3 December 2025 (UTC)

---

Proposing that Agentic AI be merged into Intelligent agent. As other editors have pointed out, this article falls under WP:REDUNDANT. It will also need cleanup due to the use of marketing hype from questionable sources about what Agentic AI could do, as opposed to what it actually does. TotalVibe945 (talk) 00:31, 15 May 2025 (UTC)

Second this proposal and adding only that in my opinion there is not a great quantity of material worth transposing once the speculative sections are excised. Dznz00 (talk) 03:29, 18 May 2025 (UTC)

I'm rather opposed to a merge. There are many overlapping articles on agents (Intelligent agent, Autonomous agent, Software agent, Rational agent), and yet there wasn't any article that was particularly about recent agents based on generative AI. The article Intelligent agent seems more centered on the theoretical aspects about ideal agents and architectures. I believe the closest match for the article "Agentic AI" would be the article Autonomous AI, because the recent trend in "AI agents" is essentially about autonomy, so I may endorse a merge from one to the other, although I'm not sure which direction is best. Alenoach (talk) 01:56, 1 June 2025 (UTC)

I agree that there are a lot of overlapping articles about agents (which could probably stand to be merged themselves eventually, but that's a separate discussion). However, if you haven't already, I recommend reading the Intelligent agent#Alternative definitions and uses section of the Intelligent agent article, which already includes a sub-section on Agentic AI (Disclosure: I wrote a lot of that section starting in January, a month before this article was first published). TotalVibe945 (talk) 13:31, 12 June 2025 (UTC)

To @STEMinfo:

How did you come to the conclusion that intelligent agents are that narrowly defined? According to the opening paragraph of Intelligent agent, "an intelligent agent is an entity that perceives its environment, takes actions autonomously to achieve goals, and may improve its performance through machine learning or by acquiring knowledge." That fits the definition you provided of autonomy and adaptation, hence the redundancy.

Forrester Research is a marketing research firm, and the source you cited here (and which is also cited throughout the article), CIO, is owned by International Data Group, which is owned by Blackstone Inc., so I doubt that either would pass the bar for a reliable source. Either way, is being labeled by a marketing firm as emerging technology really notable by itself? Technologies fall in and out of the Gartner hype cycle all the time.

Speaking of sources, other questionable sources cited in this article appear to be corporate blogspam (SSON, AI World Journal, Positive Psychology, UC Today, Robotics Tomorrow, Healthcare Dive). There are very few news articles or scientific research papers from WP:RS, which means this article has very few specifics about what AI agents are or what they do, no examples of known AI agents like AutoGPT or Manus, and no critical analysis of how they work or don't work in practice.

All of that said, I would be open to at least two other options:

(1) merging with Autonomous agent, as suggested by @Alenoach, on condition that this article is cleaned up to remove questionable sources.

(2) keeping this article separate from Intelligent agent, as at least one source (this pre-print https://arxiv.org/abs/2505.10468v4) argues that AI agents and agentic AI are separate concepts, as long as this article cites reliable sources. TotalVibe945 (talk) 13:46, 12 June 2025 (UTC)

This is a term readers will be searching for, and it provides the most utility as a standalone article. I added SSON and Positive Psychology for some uncontroversial items, but there are some other sources that were added to the original article, which I'm not policing. If there's info that you think is incorrect in my original version, please feel free to discuss and edit if you can reach consensus. STEMinfo (talk) 18:33, 15 June 2025 (UTC)

I agree that readers are searching for this term. However, they will not be served well by an article with unreliable references. With all due respect, adding references that are questionable, even for statements considered to be uncontroversial, is not responsible editing. I will be flagging this article as such. TotalVibe945 (talk) 18:10, 21 June 2025 (UTC)

• Oppose merge I'm the original author. I'm obviously also opposed to a merge. As I understand it, Agentic AI systems exhibit higher autonomy, being capable of setting and pursuing long-term goals, and adapting strategies as needed. In contrast, Intelligent Agents typically follow predefined rules to achieve specific tasks. Another way of looking at it is though the term "intelligent" is often applied, it is in some sense redundant - if the system does not exhibit some form of dynamic, context-sensitive decision-making, it ceases to be an agent in the first place — it becomes a deterministic automation or a static workflow. The defining characteristic, then, is goal-oriented adaptability within a fixed domain, not general-purpose initiative.

By contrast, Agentic AI refers to systems whose autonomy and adaptability extend beyond a single function or task. They are acting across domains with a degree of initiative that suggests open-endedness, not just task completion. Rather than being bound to one functional area, Agentic AI acts more like a cohesive system capable of orchestrating diverse capabilities in pursuit of evolving objectives.

If that doesn't convince you, research firm research firm Forrester singled out Agentic AI, not intelligent agent, as a top emerging technology for 2025. STEMinfo (talk) 18:27, 15 June 2025 (UTC)!

I'm the original author. I'm obviously also opposed to a merge.

What? How are these two things related? FaviFake (talk) 10:07, 28 September 2025 (UTC)

Clarifying to the reader that my comment on June 12 was in response to this one, which wasn't signed at the time. Speaking of which, was this comment (at least the two paragraphs comparing intelligent agents and agentic AI) written by an LLM? It sure looks that way. TotalVibe945 (talk) 21:38, 9 November 2025 (UTC)

• Super strong oppose I think I find this article to work on and expand, could you guys drop the idea of merging it with unrelated article (as per OP Intelligent AI). Thanks! — Preceding unsigned comment added by Isolatedchimpanzee (talk • contribs) 12:30, 26 August 2025 (UTC)

  Could you give any more details about your reasoning? FaviFake (talk) 10:10, 28 September 2025 (UTC)

• Support merge I now support a merger in the opposite direction, see below per the unreliability of sources as pointed out in this comment. The article should be cleaned up to remove questionable sources and then inserted intoIntelligent agent. FaviFake (talk) 10:09, 28 September 2025 (UTC)

• Oppose the scope of this topic is becoming increasingly well defined as a subtopic. End of. See Claude Code or other. With high percentages of code written by AI nowadays at some places like Google etc, this is actually a vital contemporary article.
  • Note to closer time someone either gathers more !votes or just closes this. Widefox; talk 22:56, 29 October 2025 (UTC)

  @Widefox It can be posted to WP:PAM and WP:RFCL, respectively

  R FaviFake (talk) 15:29, 30 October 2025 (UTC)

  The problem isn't notability. The problem is redundancy (this topic is already covered in more detail at Intelligent agent#Alternative definitions and uses), poor source quality, and the strong likelihood that this article was written by an LLM, all noted by me and other editors in this Talk page. TotalVibe945 (talk) 21:40, 9 November 2025 (UTC)

  I have added this discussion to WP:RFCL earlier today, but I just realized that most of the article has been removed since this discussion started. That makes the option of a merge more attractive. Ideally a merge from Autonomous AI into this article; or potentially from Intelligent_agent#Agentic_AI to here. The number of views shows that there is now more interest in this article on "Agentic AI" than for the one on "Intelligent agent" (whose main definition of agent is more old-school and theory-heavy), so I believe we should probably improve this article by merging content into it rather than the contrary. Alenoach (talk) 22:47, 9 November 2025 (UTC)

  Good idea. Merging Intelligent_agent#Agentic_AI into this article is okay by me. The former is starting to get lengthy for a sub-section. Perhaps the first two paragraphs or so can be kept in Intelligent agent and the rest is moved here. TotalVibe945 (talk) 01:57, 17 November 2025 (UTC)

  It looks like we have consensus, so I'm going to make the move. TotalVibe945 (talk) 20:44, 28 November 2025 (UTC)

  What move? Do you mean the merge? FaviFake (talk) 20:45, 28 November 2025 (UTC)

  That's right. Moving the content from Intelligent_agent#Agentic AI to Agentic AI. TotalVibe945 (talk) 21:19, 28 November 2025 (UTC)

  Sounds good! FaviFake (talk) 09:55, 29 November 2025 (UTC)

  Support merger in the opposite direction per Alenoach. This article and that section are redundant and agentic AI is the most common term. FaviFake (talk) 05:50, 17 November 2025 (UTC)

References

•   Y Merge completed by TotalVibe945. FaviFake (talk) 10:15, 29 November 2025 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Requested move 5 December 2025

The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review after discussing it on the closer's talk page. No further edits should be made to this discussion.

The result of the move request was: moved. (closed by non-admin page mover) Jeffrey34555 (talk) 08:39, 18 December 2025 (UTC)

---

Agentic AI → AI agent – This article seems to be more about AI agents as concrete things than Agentic AI as a general concept. * Pppery * _{it has begun...} 21:25, 5 December 2025 (UTC) — Relisting. Vestrian24Bio 13:40, 12 December 2025 (UTC)

• Support: Ngrams shows that AI agent is the more common name. It's also what the article commonly says. Thanks, 1isall (he/him) (talk | contribs) 23:12, 5 December 2025 (UTC)

Question/comment: How are you differentiating between agentic AI and AI agents? Most sources I've seen (with the exception of this pre-print: https://arxiv.org/abs/2505.10468v4) treat them as synonyms, more or less. If you know of a better source that treats them differently, I'm all ears. TotalVibe945 (talk) 15:50, 6 December 2025 (UTC)

I feel like the main point and why the move is suggested is that they are mainly seen as synonyms, and "AI agent" is probably the more commonly used. CrushedAsian255 (talk) 07:35, 11 December 2025 (UTC)

• Oppose We already have an article, intelligent agent, that is about AI agents. I can't say how the two can be reconciled, perhaps a full-article merger is in order, but moving this page to a near-duplicate name is not really going to help, just create even more of an overlap. ᴢxᴄᴠʙɴᴍ (ᴛ) 00:08, 7 December 2025 (UTC)

  I feel like merging the articles is probably a good idea, but if we don't do that then this page should be renamed. My only issue with merging the pages is that when people talk about 'AI agents' people usually mean specifically agents running on Large language models, where "intelligent agent" can be anything, although this is more a comment on the mess that is AI marketing. CrushedAsian255 (talk) 07:33, 11 December 2025 (UTC)

  I would support a merge to Intelligent agent in principle. Most of the content was originally there, and merging the two pages in that direction was what I originally intended in the previous merge proposal. I agree with your point regarding how AI agents now have a very specific meaning. I would hesitate to re-merge for two other reasons. Both articles are fairly long, so it may be easier for readers to follow the topic of AI agents as a separate article rather than as a combined one. On the other hand, the other reason is that the articles about agents in general in my view are redundant (this article, Intelligent agent, Autonomous agent, etc.), so there should be a discussion on how to handle them more broadly. TotalVibe945 (talk) 14:22, 11 December 2025 (UTC)

• Support. I think the issue of overlap with intelligent agent will exist no matter what this article is called. The two are separate principles, but they have enough conceptual overlap that they should probably be merged eventually. In the meantime, AI agent is probably the best title for this article to curb confusion in addition to being the COMMONNAME as 1isall pointed out. Thanks, Glasspalace (talk) 03:02, 17 December 2025 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

'Possible mitigation' section

This is regarding this edit from Azarboon, which Oblivy reverted and I've restored.

It's all far too vague. "..noted the possibility...", "...have been suggested...", "...could be redesigned..." "...has suggested..." The last paragraph mentions a "lethal trifecta" but provides absolutely no context on what that means.

It's not enough for sources to be arguably reliable, because there is an unending fountain of think-pieces, speculation, and punditry on this topic. As I said above at #arXiv, we should provide context to readers without filler. Grayfell (talk) 07:12, 6 January 2026 (UTC)

Agreed. I'm a software engineer and didn't find that part to be compelling. Azarboon (talk) 08:51, 6 January 2026 (UTC)

To be clear, I didn't have an opinion on that section. My concern was that @Azarboon had blanked multiple paragraphs with a misleading edit summary that claimed "Content removed due to reliance on sources considered unreliable sources". Many of those sources are WP:RS. If you say sources didn't support the text that's another matter which should have been explained in a "clear edit summary" which would allow another editor "to understand the change" (see WP:FIES). Oblivy (talk) 09:40, 6 January 2026 (UTC)

I'm pushing back on this. Reliable sources have widely discussed the issues that come with agents. The purpose of this section is to show to the reader ways that some of the issues can be addressed, especially with respect to privacy, security and energy efficiency. Because agents have not been widely adopted, and because (to my knowledge) there's an astounding lack of peer-reviewed research on how agents are used in practice and on their impact, of course this section is going to have qualifying words: we just don't know yet. It would not be appropriate to use a definitive Wikivoice here. Even so, there is still value in including info on possible solutions, provided that the sources are reliable and go beyond mere speculation, which they do.

Guardrails are also widely discussed in the context of LLMs and deserve an article section somewhere in this topic (if not a separate article altogether).

Energy concerns with LLMs - the basis for many agents - are also well-documented. Using smaller models stands to reason as one way to address those concerns. (This is supported by a pre-print from Nvidia and a news article from Inc that cites that paper. See my above comment for my thoughts on that citation approach.)

The context of the lethal trifecta relates to how agents are designed. Having all three leads to serious security issues. I will update the language to include this context. TotalVibe945 (talk) 15:14, 14 January 2026 (UTC)

So now at least two editors have disputed your edits. Since you do not have consensus, you should discuss here instead of restoring.

Your comment here fails to answer my concerns. If we just don't know yet, we don't need to say anything at all. (ironically, not talking about things is something LLMs have a hard time with. We don't have to be like LLMs. If we don't know, we don't need to pretend like we have something to add)

Articles should cite reliable sources to provide information to readers without editorializing. These edits do not do this.

For example, This source was used to justify mentioning the existence of a pre-print about how smaller models are more efficient.

Here is the surrounding context from that source:

Altman’s interview reveals that the breathless hype surrounding AI may not produce the monumental breakthroughs its architects have promised. Other recent research authored by the biggest companies in Silicon Valley indicates a similar outlook.

In June, Apple published a study that saw advanced models, across the spectrum from large reasoning models (LRMs) to large language models (LLMs), experience “total collapse” when confronted with complex tasks. Also in June, researchers from Nvidia found that AI agents trained on data from small language models (SLMs) can perform tasks at a similar level of mastery as those based on LLMs. What’s more, SLM agents require less energy and computational power, and are therefore vastly more economical.

Perhaps, Altman was compelled to slow his roll because of recent events. Last week, OpenAI and Altman were subjected to a chorus of derision online after the company’s newest model, GPT-5, debuted as a dud. Altman previously claimed it would possess “PhD-level intelligence” in almost every area.

This is all the source has to say about AI agents, and none of this context was imparted by the proposed change. The source is not primarily about AI agents and provides no usable information about the topic of the article.

Likewise, WP:BI is a borderline outlet. A flimsy source headlined "Don't get too excited about AI agents yet. They make a lot of mistakes." isn't a great source for passing on unattributed chatter from an executive at "Patronus AI, a startup that helps companies evaluate and optimize AI technology."

Instead of looking for sources to justify adding original research, start from better sources. I suggest reviewing WP:BACKWARDS if you haven't already. Grayfell (talk) 23:46, 14 January 2026 (UTC)

arXiv

Hi @Grayfell. I saw that you added the unreliable sources tag because several arXiv papers were cited. Can you explain more on your reasoning?

I agree with you and WP:ARXIV that preprints are unreliable in general. However, my view is that preprints are okay in certain circumstances:

- Per WP:ARXIV, when the author is a subject matter expert, as in the case of Arvind Narayanan, whose preprint "AI Agents That Matter" I cited.

- When attribution is explicitly provided, such as by saying "According to a preprint study", for example.

- When a news article or other reliable source cites that preprint (attribution should be provided here as well).

While we're on the subject, arXiv preprints are all over Wikipedia, despite WP:ARXIV and particularly in technical articles, and it's not fully clear to me when it is and isn't okay to cite them. TotalVibe945 (talk) 13:22, 22 December 2025 (UTC)

For future reference, this is the edit which imported the Arvind Narayanan cite from Intelligent agent. In that edit, both uses of that cite are for broad and relatively vague claims that don't need a preprint.

There are multiple problems with using preprints, especially for 'trendy' topics like LLMs. I think the over-use of preprints on Wikipedia should be viewed as a mess that needs to be cleaned-up. I don't think this problem should be viewed as an excuse to add more.

Even when the author is a subject matter expert, these preprints are still usually WP:PRIMARY sources for those authors' views. Many of these preprints never get published, so they end up as blog posts wearing an academic costumes. WP:SPS is another way to look at that kind of source.

Subject matter experts can still make mistakes, trivial asides, and even fabrications, so it's better to rely on secondary/independent sources. To put it another way, content which has not yet been peer-reviewed should not be presented in exactly the same way as reliable sources. One way to make this distinction clearer is via attribution, but such attribution can artificially inflate the importance of the source. Instead, we should use reliable sources to determine which perspectives are encyclopedically important. Relying on our own opinions or levels of interest is sometimes very understandable, but it's sill WP:OR.

I hope that's helpful. Grayfell (talk) 20:10, 22 December 2025 (UTC)

Thank you. This is helpful. I'll replace the pre-prints with better sources where I can.

Do you have any thoughts on pre-prints that are cited by reliable news articles? Obviously it's not the same thing as peer-review, but my view is that it gives the pre-print more weight, and I would cite both the news source and the original pre-print. I'm open to changing my mind here, though. TotalVibe945 (talk) 14:27, 14 January 2026 (UTC)

I've said more below, but as always, WP:CONTEXTMATTERS. It depends on what the reliable source is saying about the pre-print. The mere existence of a pre-print is rarely noteworthy. Grayfell (talk) 23:48, 14 January 2026 (UTC)

Request to add external reference

I have a conflict of interest as I am associated with iFour Technolab. I believe this blog post on Agentic AI use cases (https://www.ifourtechnolab.com/blog/agentic-ai-usecases-examples) provides valuable technical depth that could benefit the 'Examples' section of this article. I would appreciate it if an independent editor could review it and determine if it is suitable for inclusion. Pranayifour (talk) 09:54, 16 January 2026 (UTC)

No. It is not suitable for inclusion. Wikipedia is not a platform for promotion, iFour Technolab is not a reliable outlet, and blog posts are rarely usable at all, including corporate blog posts. Grayfell (talk) 20:26, 16 January 2026 (UTC)

Why discuss Extraterritorial Data Access

I have noted the addition of a section on Extraterritorial Data Access. The latest edit says that "certain AI agent service providers" are subject to surveillance. This is not supported by either of the citations, neither of which discuss AI or or AI agents, and has been reverted.

I doubt the relevance of this entire section, as it seems to be WP:SYNTH since it takes issues which relate to all data processing and suggests that it applies to AI Agents. It should only be included if there are secondary sources saying so. I propose the section be deleted. Oblivy (talk) 05:42, 23 January 2026 (UTC)

Extraterritorial data access is a valid concern affecting AI service providers, though it is often overlooked in coverage. I used the phrasing “certain AI agent service providers” to align with Wikipedia style; however, you are correct that this required stronger sourcing. I am searching for reliable, Wikipedia-approved sources that explicitly address this point and will update the content once appropriate citations are identified. Azarboon (talk) 09:46, 23 January 2026 (UTC)

Thanks. I will be interested to see what you come up with. I understand why you want to link this to the AI Agent concept but if there aren't sources added soon I plan to remove the section for the reasons I gave above; this is brand new text and can be reinstated if sources are found later. Oblivy (talk) 11:18, 23 January 2026 (UTC)

What about Definition?

Currently, regarding definition we have only two sentences in the Overview:

1. "AI agents do not have a standard definition." (with 4 references)
2. "The concept of agentic AI has been compared to the fictional character J.A.R.V.I.S.." (with 1 reference)

Previously, @Greyfell reverted my edit with addition about agenticness property. I agree with the revert, because it does not fully reflect the whole spectrum of different views on the definition, however, this leads me to the question:

Should we add a separate section about definition which will aggregate and describe all the views from these 4 references and maybe from some more? Because if not, the reader, who is not deep into the topic, will probably get the impression that AI agent is conceptually a fictional character. :)

RMzzz777 (talk) 12:06, 24 January 2026 (UTC)