JWH – What is the difference between Agentic AI and Agent-based AI?
⚡ Over two decades ago, the foundational elements for successfully utilizing advanced, self-learning algorithms to optimize the procurement process using an agent-based model were well established.
⚡ Somewhere along the line, we got distracted by SaaS and digital transformation, and now we are at risk of being sidetracked by Agentic AI.
⚡ One critical point to avoid having this same discussion about failed or failing initiatives 10 to 15 years from now is not to confuse Agentic (equation-based) and Agent (agent-based) development and implementation models.
What are your thoughts about Agentic AI, e.g., what is it, and how will it help your procurement team?
Koray Köse – Jon W. Hansen – hate to say this – but it’s just too many miles off and wrong.
“Agentic” refers to qualities of autonomy, self-direction, and purposeful behavior, often associated with having agency rather than being defined by modeling types. Hence agentic AI is the future.
“Agent” represents an individual component within a system that can act independently. Hence limited – just as traditional AI and genAI.
Excerpt on Modeling Approaches:
Equation-based modeling describes system behavior at a macro level using mathematical equations, which provide an overarching view of the system as a whole, often summarizing dynamics without focusing on individual components.
Agent-based modeling (ABM) simulates the interactions of individual agents within a system, allowing for the study of emergent behavior that arises from these interactions. ABM is particularly valuable for capturing complex, system-wide effects that result from the interactions of autonomous agents.
You would make a mistake if you say the distinction between agentic and agent is inherently tied to these modeling approaches. It’s not.
Rather, it’s a semantic distinction: agentic is about autonomy, while agent focuses on the independent role within a system.
JWH – Koray Köse, thank you for sharing your thoughts. I respect your opinion but respectfully disagree.
Your description—or interpretation—of agent-based models has been part of mainstream thinking and is too narrow. In short, it doesn’t just simulate interactions. It allows humans to identify and understand the unique operating attributes within seemingly disparate and previously considered unrelated data streams.
Only after you expand your use of an agent-based model beyond theory or simulation can you effectively leverage technology, including AI, GenAI, and Agentic AI.
How will Agentic-AI autonomously or magically acquire and act on insights outside of an agent-based model? How can it gain the autonomy to which you refer without a real-world agent-based understanding? Are you suggesting that Agentic-AI, unlike AI and GenAI, will succeed outside of the real-world insights of an agent-based model?
A point of convergence must take place in the real world, starting with a sound agent-based model. Otherwise, no matter how advanced the technology is, it will result in the same “garbage-in, garbage-out” results, contributing to the generational initiative failure rate of 80%.
This should help – https://youtu.be/49BS-MkGoak.
Koray Köse – I can see that 😅
…though it’s not an opinion – just definitions that are foundational and misused or misunderstood.
This isn’t “free to opinions” or “semantics rule over basics.”
Too much mixing of things and creating a statement rather than building up the argument on existing and accurate definitions.
Agentic AI will do by training and actual purpose driven optimization.
We can compare notes soon – latest in 2030 – and leave it here. 😊
Gotta run 👍
John M Lopes – Difference: In the case of agentic AI, the agents are working on behalf of the primary AI, the thing coordinating all of their efforts and consolidating them for the user.
In agentive technology, the agents are working on behalf of the person, the user. Agentic AI continuously learns from its previous experiences without human intervention to generate a series of actions or decisions. This technology uses machine learning and natural language processing to set goals and make decisions based on available information. Unlike generative AI, agentic AI can make its own decisions.
JWH – John M Lopes, thank you for the added perspective.
Please see my comment regarding Agentic AI and Agent-based modeling in this discussion stream – https://bit.ly/40zCeAJ
You’ll also want to watch the video referenced above, in which I provide a high-level overview of my work with self-learning algorithms in the late 1990s within a nascent AI framework.
The net result is that, outside of an agent-based model, Agentic AI, Generative AI, and AI (or, for that matter – any technology) will not learn and function effectively.
TODAY’S TAKEAWAY
The net result is that, outside of an agent-based model, Agentic AI, Generative AI, and AI (or, for that matter – any technology) will not learn and function effectively.
Unless we understand the above, the generational initiative failure rate of 80% will continue, and we will find ourselves having this same discussion 10 to 15 years from now!
30
Procurement Insights
An important and timely discussion about Agentic AI
Posted on November 5, 2024
0
JWH – What is the difference between Agentic AI and Agent-based AI?
⚡ Over two decades ago, the foundational elements for successfully utilizing advanced, self-learning algorithms to optimize the procurement process using an agent-based model were well established.
⚡ Somewhere along the line, we got distracted by SaaS and digital transformation, and now we are at risk of being sidetracked by Agentic AI.
⚡ One critical point to avoid having this same discussion about failed or failing initiatives 10 to 15 years from now is not to confuse Agentic (equation-based) and Agent (agent-based) development and implementation models.
What are your thoughts about Agentic AI, e.g., what is it, and how will it help your procurement team?
Koray Köse – Jon W. Hansen – hate to say this – but it’s just too many miles off and wrong.
“Agentic” refers to qualities of autonomy, self-direction, and purposeful behavior, often associated with having agency rather than being defined by modeling types. Hence agentic AI is the future.
“Agent” represents an individual component within a system that can act independently. Hence limited – just as traditional AI and genAI.
Excerpt on Modeling Approaches:
Equation-based modeling describes system behavior at a macro level using mathematical equations, which provide an overarching view of the system as a whole, often summarizing dynamics without focusing on individual components.
Agent-based modeling (ABM) simulates the interactions of individual agents within a system, allowing for the study of emergent behavior that arises from these interactions. ABM is particularly valuable for capturing complex, system-wide effects that result from the interactions of autonomous agents.
You would make a mistake if you say the distinction between agentic and agent is inherently tied to these modeling approaches. It’s not.
Rather, it’s a semantic distinction: agentic is about autonomy, while agent focuses on the independent role within a system.
JWH – Koray Köse, thank you for sharing your thoughts. I respect your opinion but respectfully disagree.
Your description—or interpretation—of agent-based models has been part of mainstream thinking and is too narrow. In short, it doesn’t just simulate interactions. It allows humans to identify and understand the unique operating attributes within seemingly disparate and previously considered unrelated data streams.
Only after you expand your use of an agent-based model beyond theory or simulation can you effectively leverage technology, including AI, GenAI, and Agentic AI.
How will Agentic-AI autonomously or magically acquire and act on insights outside of an agent-based model? How can it gain the autonomy to which you refer without a real-world agent-based understanding? Are you suggesting that Agentic-AI, unlike AI and GenAI, will succeed outside of the real-world insights of an agent-based model?
A point of convergence must take place in the real world, starting with a sound agent-based model. Otherwise, no matter how advanced the technology is, it will result in the same “garbage-in, garbage-out” results, contributing to the generational initiative failure rate of 80%.
This should help – https://youtu.be/49BS-MkGoak.
Koray Köse – I can see that 😅
…though it’s not an opinion – just definitions that are foundational and misused or misunderstood.
This isn’t “free to opinions” or “semantics rule over basics.”
Too much mixing of things and creating a statement rather than building up the argument on existing and accurate definitions.
Agentic AI will do by training and actual purpose driven optimization.
We can compare notes soon – latest in 2030 – and leave it here. 😊
Gotta run 👍
John M Lopes – Difference: In the case of agentic AI, the agents are working on behalf of the primary AI, the thing coordinating all of their efforts and consolidating them for the user.
In agentive technology, the agents are working on behalf of the person, the user. Agentic AI continuously learns from its previous experiences without human intervention to generate a series of actions or decisions. This technology uses machine learning and natural language processing to set goals and make decisions based on available information. Unlike generative AI, agentic AI can make its own decisions.
JWH – John M Lopes, thank you for the added perspective.
Please see my comment regarding Agentic AI and Agent-based modeling in this discussion stream – https://bit.ly/40zCeAJ
You’ll also want to watch the video referenced above, in which I provide a high-level overview of my work with self-learning algorithms in the late 1990s within a nascent AI framework.
The net result is that, outside of an agent-based model, Agentic AI, Generative AI, and AI (or, for that matter – any technology) will not learn and function effectively.
TODAY’S TAKEAWAY
The net result is that, outside of an agent-based model, Agentic AI, Generative AI, and AI (or, for that matter – any technology) will not learn and function effectively.
Unless we understand the above, the generational initiative failure rate of 80% will continue, and we will find ourselves having this same discussion 10 to 15 years from now!
30
Share this:
Related