The increasing adoption of AI agents in various industries is a significant development that will impact every professional role. Companies like Thomson Reuters are already leveraging these technologies to enhance their services, and it's only a matter of time before Nigerian businesses follow suit.
Professionals exploring the use of AI agents in their roles will need to consider how to build trustworthy systems that can be relied upon. According to Joel Hron, CTO at Thomson Reuters Labs, measurement, collaboration, and experimentation are key to achieving this goal.
Thomson Reuters uses a combination of in-house models and off-the-shelf tools to power its AI innovations, and Hron emphasizes the importance of exploiting proprietary knowledge and assets. The company's AI-powered legal research tool, Westlaw Advantage, and its Deep Research agent are examples of successful agentic achievements.
Hron has identified four key lessons that professionals can use to build trustworthy agentic AI systems. The first lesson is to measure success, which involves evaluating the performance of AI agents and defining what makes an answer good. This requires a combination of public benchmarks, internal benchmarks, and human assessments.
Thomson Reuters tracks and measures agentic success in several ways, including leveraging public benchmarks and developing internal benchmarks with strong directions for automated evaluations. The company also keeps humans in the loop, ensuring that evaluations go beyond automated assessments and involve human experts.
As AI agents become more prevalent, Nigerian businesses will need to consider how to build trustworthy systems that can be relied upon. By following Hron's lessons, professionals can ensure that their AI agents are effective and reliable.
Thomson Reuters' approach to building trustworthy AI agents is a model for Nigerian businesses to follow. By prioritizing measurement, collaboration, and experimentation, companies can create AI systems that are reliable and effective. This is particularly important in Nigeria, where the adoption of AI technologies is still in its early stages.
