24-06-2025
How AI Agents Can Take Your Business Analytics To Another Level
Alon Goren, CEO and Cofounder of AnswerRocket, transforming your analytics with AI.
In my last FTC piece, I provided a primer on the capabilities of agentic AI, the value of the tech and how it can be tested and tweaked to improve accuracy. AI agents can make a major impact in many ways, such as cybersecurity, robotic process automation and customer support. But there's one use case where I've seen them really shine: business analytics.
Back in November 2023, I discussed how generative AI is accelerating enterprise analytics. LLMs allow normal business users to tap into their organization's data and uncover critical new insights. Agentic AI is taking things to an exciting new level here, and I believe they will eventually make LLMs obsolete when it comes to driving value from business data.
Agentic AI Analytics: Fulfilling The Role Of An Expert Analyst
Traditional GenAI analytics is powerful, making enterprise data more accessible and yielding far more insights versus plain business intelligence (BI) analytics. It works within clearly defined guardrails, learns as you interact with it and provides precise answers. LLM analytics ultimately fulfills the role of a junior analyst for organizations.
Agentic AI plays the role of a manager or expert analyst. It teaches itself new things, researches things on its own and delivers insights autonomously.
The key differentiator for agentic AI analytics is its proactive nature—it delivers valuable insights without needing explicit requests or prompts. For instance, consider a consumer goods company specializing in beverages. An AI agent could proactively alert business users that sales of a seasonal product line, such as flavored seltzers, are projected to decline significantly over the next quarter due to shifting consumer preferences. At the same time, AI could highlight emerging trends, such as rising interest in non-alcoholic spirits, recommending that the company explore opportunities in this growing market segment within the upcoming year.
As is always the case with AI and analytics, the important thing is that insights support meaningful actions. In the first example, the liquor company might want to consider pivoting away early from the declining category before sales tank. In the second example, they would want to think about launching a new product to get ahead of their competitors.
Here are the features that define early-stage generative AI analytics solutions:
• Rule-Based: Performs only the tasks it's explicitly programmed to do
• Opaque: Offers answers without explaining how it reached them
• Tool-Limited: Can only operate within a fixed set of preloaded tools
• Inflexible: Needs manual corrections or instructions to adapt
• Requires Oversight: Relies heavily on expert oversight to function properly
Here's how agentic AI analytics contrasts in the same categories:
• Autonomous Decision Making: Weighs options and makes choices independently
• Explainable: Clearly shows how it reached its conclusions
• Tool-Agnostic: Can choose and use tools on its own as needed
• Self-Adaptive: Adjusts behavior in real time without external input
• Self-Monitoring: Performs built-in checks to stay compliant and accurate
Don't Fall For Regular GenAI Posing As Agentic AI
The AI market is evolving rapidly. It can be difficult for enterprises to make heads or tails of all the various moving parts. Complicating things further—and this is always the case with the rise of significant new technologies—there are a lot of vendors that cling to buzzwords even when they don't fit their offerings. Organizations looking to leverage agentic AI to accelerate their analytics efforts need to be careful not to fall for plain generative AI that rebrands itself as agentic.
This will become less of a problem as the agentic AI market matures and winners and losers emerge within the next two to three years. In the near term, organizations will just have to do a little research. The best place to start is with this checklist, reflecting the points I hit above.
Agentic AI analytics should:
1. Make decisions independently.
2. Explain reasoning.
3. Use tools autonomously.
4. Self-correct and adapt on its own.
5. Be overseen by verifiers to ensure optimal accuracy.
A Step Further: Multi-Agent Networks
Looking even further ahead, agentic AI gets even more groundbreaking. Eventually, singular AI agents will evolve into multi-agent networks. Here, several AI agents will connect into a network with broader access to enterprise datasets, tools, models and domain context. These agents will be highly goal-driven and capable of completing more complex tasks that span multiple systems within a business.
AI Agents: Transforming Enterprise Analytics In 2026 And Beyond
AI continues to develop at a breakneck pace. It wasn't long ago that LLMs were a brand new, cutting-edge way to support analytics.
It should be repeated that traditional GenAI is still a fantastic, powerful method to improve analytics workflows and uncover more insights. However, AI agents are going to raise the bar. The tech is still in its nascent stage, though the market will start to take shape in a year or so, delivering insights that will prove transformative for organizations across the spectrum.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
