What are AI Agents?
You might have already seen Satya Nadella’s Clip from Podcast with BG2 where he predicted the AI Agents going to replace SAAS Based Applications going forward. This is what he said “In Agentic Era, the business applications that are essentially CRUD(Create, Read, Update and Delete) databases with a bunch of business logic, and this business logic will be handled by AI agents or will be in AI tier in future.”
and also, Mark Zuckerberg predicting that there going to be more AI agents in the world than the people, and every small business will be having an AI agent similar to an email or a social media account.
AI Agents are going to bring a seismic change in the way SAAS or Business applications are going to be build, used and function in the future. Be it CRM, Finance and Operations, or any field, whatever the backend be, AI agents will be replacing the traditional software as a service.
What are AI Agents?
AI Agents are software programs that performs goal based actions by interacting with environment, collecting data and performing certain tasks based on the collected data.
How AI Agents are different?
Ability to reason
Ability to act (via tools like search, calculate, or an API)
Access memory
Think Slow
How do they work?
AI Agents are rational and they analyze data and take informed decisions. They store the new data in their memory and use it to smartly make the decision when they face a similar situation the next time.
They use machine learning to process huge amount of data and arrive at optimal solutions for the goal or task at hand.
They take inputs from the user in terms of goals, and break the goals into smaller actionable tasks and then they acquire information from database or LLM or any data source and finally they act on it through reasoning and perception in accomplishing the tasks. They iterate over or think slow accessing their memory and reasoning untill they find an optimum solution.
What are the benefits?
Improved productivity
Reduced costs
Informed decision making
Improved customer experience
Handling complex use cases in less time
Some real life examples
You have a business where you store inventory and the inventory is huge. For example in traditional software setup, you can write a software code or business logic where you get an email alert or a notification when inventory is emptied or is below certain level. In this case, you can have an AI Agent where you can set a goal where it intelligently captures data when inventory is down and autonomously raise purchase orders to replenish the inventory based on the demand forecast and past sales data.
You can use AI agent in Source Code repository like Guthub in analysing any software related best standards or help you in writing unit test cases for your code or in generating more optimum software code for the logic in hand or in migrating your code from one version to other version by handling library changes or any version related changes between the source and target.
In DevOps, you can use an AI Agent which can analyze application logs and real time monitoring data to figure out any application related errors and infra related bottle necks and the agent can analyze the error or anomaly to guide you how to fix it or it can itself interact with infra and make related changes.