Related Posts
- Advertisement -
Agentic AI is one of the latest concepts in artificial intelligence, which now achieves real traction than his early buzz. Current progress in Agentic AI accelerates the development of autonomous business systems, building on the performance of machine learning.
This technology works as an independent ‘agent’ and is equipped to make informed decisions based on the multimodal data and algorithmic logic, and can then ‘learn’ and evolve through experience.
Even more exciting is his ability to act independently. It is this unique ability to adjust, plan and perform complex tasks without human supervision that distinguishes agentic AI from previous generations of AI Tools.
In Supply Chains, for example, AI agents can follow market activity and historical demand trends to predict stock needs and implement measures to prevent shortages, such as by automating parts of the restocking processes. These agents shift their behavior in response to changing market conditions, stimulating efficiency and performance. It is therefore no surprise that 26% of the managers report that their organizations are starting to shape strategic approaches on Agentic AI.
No matter how great it sounds to outsource such tasks to Agentic AI, we must also be incorrect. How can the actions and output of AI agents be completely familiar with all his autonomous strength? If we rely on agent AI to complete advanced tasks in itself, how do we ensure that the decisions are really based on what is happening in the real world, or in the view of the company in the world?
In the same way, our brains use observation and extra input to draw conclusions, AI agents must trust many external sources and signals to improve their reasoning opportunities.
This need can be met by solutions and platforms that gather and present themselves facts In a way that is accessible and collected. Here is how:
The trust challenge in autonomous AI systems
As discussed, what distinguishes Agentic AI from other AI systems, the ability to act autonomously, is not just a linear conversation. The complexity of the tasks that have been completed usually requires that they refer to several dynamic external sources. As a result, the risk that something will go wrong automatically. For example, you can rely on a chatbot To give you an update on the status of a claim or repayment, but would you trust yourself as much as you give your credit card information to book a flight for you when giving an AI agent?
Away from conversation AI Planning and changing on task -based agents actions, depending on the context they have given. They delegate sub -tasks to the different tools that are available through a process that is often called ‘chains’ (the export of one action becomes the entry for the next). This means that questions (or tasks) can be subdivided into smaller tasks, where each needs access to data in real time, processing iteratively to imitate human problem solving.
The chain effect (in which decisions are made) is informed by the environment that is being monitored, ie the data sources. As a result, the collection of explanatory and accurate data is required for two reasons for each step of the chain. Firstly, users must know why the AI agent has landed on a certain decision and have the visibility of the data source on which it is based.
They must be able to trust that the action is in fact the most effective and efficient. Secondly, they must be able to optimize the process to get the best possible result every time, to analyze every phase of the output and learn from any dissatisfied results.
To trust an agent to complete advanced tasks on the basis of multiple collection steps, the value of the data required to support the decision -making process multiplies considerably.
The need to make reliable Enterprise data available to agents is the key. This is why companies are increasingly recognizing the power of the graph database Technology for the wide range of collection strategies it offers, which in turn multiply the value of the data.
How graphing technology AI Reasoning reinforces
Since agent AI stimulates decisions from data, the insights underlying these decisions must be accurate, transparent and explained – benefits that are graphics databases are unique optimized to deliver. Gartner already identifies knowledge charts as an essential capacity for Genai applications, as a burial rag (pick up augmented generation), where the collection path contains a knowledge graph, the accuracy of the outputs can greatly improve.
The unique structure of knowledge graphs, consisting of ‘nodes’ and ‘edges’, is where reactions of higher quality can be derived. Nodes represent existing entities in a graph (such as a person or place), and edges represent the relationship between those entities – ie how they connect. In this type of structure, the larger and more complex the data, the more hidden insights can be revealed earlier. These characteristics are invaluable when presenting the data in a way that makes it easier for AI agents to complete tasks in a more reliable and useful way.
Users have discovered that graphrag -answering is not only more accurate, but also richer, faster, more complete and therefore more useful. For example, an AI agent who is looking for customer service can offer a certain broadband package with a discount based on a complete understanding of the customer, as a result of the use of graphrag to connect various information about that customer. How long has the customer been at the company? Which services are they currently using? Have they previously submitted complaints?
To answer these questions, nodes can be made to each aspect of the passing experience With the company (including previous interactions, service use and location) and edges to show the cheapest or best service for them. A fragmented and distributed image of the data can lead to the agent offering a discounted package if it was not due, which leads to cost implications for the company.
As mentioned by the CEO of Klarna, “will result in feeding the fractionated, fragmented and distributed world of company data in a very confused LLM”. But the result is very different when data is connected in a graph: positive results have been reported by the LinkedIn customer service team, which have shortened the median resolution time per edition by 28.6% since the implementation of graphrag.
Why connected data is crucial for agent AI-cleverness
With every iteration, the LLMS Behind AI agents are improving quickly and agent frameworks make it easier to build complex, multi-step applications. The next essential move is to make company data as rich, connected and contextually possible, so it is fully accessible for these powerful agents.
With this step, companies can unlock the full value of their data, making agents not only more accurate and more efficient, but also easier to understand and explain. This is where the integration of agentic AI and knowledge graphs is transformational. Connected data give agents the context they need to think more clearly, to generate smarter output and have a greater impact.
We have put together a list with the best survey tools.
This article is produced as part of the TechRadarpro expert insight channel, where today we have the best and smartest spirits in the technology industry. The views expressed here are those of the author and are not necessarily those of TechRadarpro or Future PLC. If you are interested in contributing to find out more here: https://www.techradar.com/news/submit-your-story-techradar-pro
- Advertisement -
