Why Knowledge-Based Agents Are The Future Of More Trustworthy AI

Forbes

01-05-2025

Photo by Markus Spiske on Unsplash
There's a common issue with generative AI. I'll explain with an example from a friend. She was drafting her biweekly newsletter. The theme is travel. She created a list of her favorite restaurants in Rome and asked ChatGPT to fill in the corresponding locations. Within seconds, it churned out the list with street addresses for each. She copied and pasted it into her newsletter and hit send.
The problem? Half of the addresses were 100% incorrect.
ChatGPT and other generative AI tools are people pleasers. They would rather say anything, including factually incorrect answers, aka hallucinations, than admit they're unsure of the information. Knowledge-based AI agents are different—and in many ways more reliable. Here's why.
Imagine receiving a research task—explain how to write a screenplay—and having carte blanche on where to look for the answer. You may have unlimited resources at your fingertips, but you also have a wide berth on how to answer the question. You might pull advice from dozens of screenwriting blogs, Reddit threads, and Medium posts. The result could be an answer that sounds confident but blends conflicting guidance and unvetted opinions.
Now, imagine receiving the same task but access to just one book: Syd Field's Screenplay: The Foundations of Screenwriting. Your answer would be narrower, yes—but also traceable, grounded in a known methodology, and more likely to be consistent. That's how knowledge-based AI agents work: they reason from a fixed, deliberately created source instead of improvising based on massive datasets and probabilistic guesses. They use curated knowledge to generate answers and make decisions in real-world situations.
Transparency is one of the key strengths of knowledge-based agents. Because they follow defined rules and rely on sources you select, you can trace their reasoning step by step. A tool like Text Cortex, for instance, shows exactly where it sourced its information—something especially useful for organizations navigating compliance and regulatory demands.
These agents are also adaptable. If a reference is updated or conditions change, they can automatically incorporate new information. And when conflicts arise, they can flag them for human review, streamlining operations without compromising accuracy or trust.
Personally, I love these agents because they help organizations scale decision-making without sacrificing transparency.
As Reid Hoffman and Greg Beato explain in Superagency, sometimes you want to collaborate with AI, like when learning a new skill. Other times, like optimizing energy use based on real-time data, you want it to act autonomously. In both cases, AI boosts your agency by helping you take smarter actions toward your goals.
Getting started with knowledge-based agents is pretty simple. To create the knowledge base, you can add any files, links, websites, and materials that you would like your agent to reference.
If you're wondering what are the best uses for knowledge-based agents, consider any processes that tend to be rule-heavy or regulated. In terms of tasks, these agents excel at searching, learning, and training.
In healthcare, for example, you can use knowledge-based agents to provide quick and accurate medical guidance based on defined medical references. In finance, you can use agents to field customer queries and provide account details. In e-commerce, you can use knowledge-based agents to offer product recommendations.
Take Alibaba, the global commerce juggernaut—it launched Accio, an AI-powered B2B search engine for product sourcing. To ensure the information it provides is accurate and useful, Accio uses retrieval-augmented generation (RAG), which pulls data only from sources trusted by its underlying knowledge base. If the AI encounters uncertainty, it flags it instead of offering potentially inaccurate guesses, according to Alibaba.com. In short, it admits its shortcomings rather than making stuff up.
We're solidly in the era of AI. As reliance on these tools grows, so does skepticism—trust and transparency are more important than ever, especially when it comes to critical business decisions. Knowledge-based agents offer a compelling alternative to generative AI: while the latter can be incredibly useful, it's not always reliable.
If you're looking for a business solution grounded in accuracy and traceability, it's time to start exploring what knowledge-based agents can do. Because sometimes, the best answer isn't the fastest one—it's the one you can trust.