Businesses are looking to improve their work with the help of AI, but for this they need their technology to be up to date. Last year, the biggest finding in AI was agentic AI, which is now in tested mode for widely used..
Gary Kotovets from Dun & Bradstreet stated that their company firstly used to AI. which has a more advanced system as compared to the chatbot. They can handle a complete workflow. However,it also combined with new challenges. Unlike older systems that follow strict rules, AI agents always do not provide fixed results . Many companies are struggling to combine their old technology with the advanced AI system . Out of 100,70% developers facing issues to combine their old technology with the advanced AI system .
To get the best result from AI, companies need to advance their old technology. These AI systems require access to multiple sources of data, which makes security and compliance more difficult. Kotovets explains that AI output always depends upon the data , better data is the source for the better result data, having poor data can be the main reason for the worst result.
A survey by Tray.ai found that nearly all businesses agree to combine their own data with AI for the better result . However, most realize that before that they need to upgrade their technology before this process. Ashok Srivastava suggests that—without proper software systems and easy ways for AI to connect, the AI agents won’t work well. Intuit, a company that processes huge amounts of data, needs to modernize their technology to speed up the process . But it is not affordable by all companies.For this reason, Many still use old systems,which are essential for their work but not able to be made for AI . AI agents need access to these systems, but older technology is not able to work in real time.
Companies that have already adopted the new technolgy. But even with modern systems, AI needs to be connected properly. Indicium, a global data company, has had success with AI agents but faced a challenge: AI works in plain English, while business systems use structured formats like XML or APIs. This often means another AI agent is needed just to translate between them.
Nowadays companies are facing the AI generation and they are in two phases either in testing mode or trying to apply it very soon. By 2033, AI agents are expecting that AI will be an integral part of the business by help[ing it in various ways.
Cisco uses traditional software for communication between AI agents and business systems. One major challenge is that AI agents need different sources of information, and managing it in the safest way is a very difficult process.
Another difficulty is the control of AI. Giving AI too much freedom can create risks,so its need to control carefully, Cisco, for example, ensures that AI agents can only access information and take actions based on what the requesting human is allowed to do.
Major concerns are Security and compliance .AI models are sometimes tricked for doing things which they should not so careful monitoring is required by AI agents to do things the correct.A study found that when some confusing instruction is given to AI , it can create security threts. So companies need strict controls over what AI agents can do.
At Dun & Bradstreet, AI agents help customers access business records,but they are strictly prohibited to make changes on it. Similarly, at IT company CDW, AI agents have a boundary and they have to perform within the boundary and their work will be monitored every time , if they perform in any suspicious function then action should be taken against them.
Starting with a small scale can have a high rate of success. Roger Haney from CDW explains that a single and specific work can be manageable and also able to reduce the security issue.Instead of immediately linking multiple AI agents together, companies should first make sure they have clear rules and safety measures in place.
In conclusion AI agents have great potential, but they must be properly set up, controlled, and monitored for effectiveness and safety. Businesses need to balance innovation with caution, making sure their technology, data, and security measures are ready before fully using AI on a large scale.
