Perspectives
In Perspectives Edition 35, Thoughtworks experts, with hands-on experience in agentic AI, share insights to help you navigate this evolving space, and explain how enterprises can balance the rewards with the risks. Below we share a short summary of this edition. To read the full edition, follow the link below.
The agentic difference
After GenAI gripped the business world, the next big shift looks set to be agentic. The number of companies discussing, evaluating and adopting AI agents has surged as business leaders bet on their transformative potential.
But despite the burst of activity in the space, what constitutes agentic AI, and how enterprises can leverage it to their advantage, isn’t always well defined. In this issue of Perspectives Thoughtworks experts, with track records in agentic AI research and enterprise projects, weigh in on the implications and applications of a technology poised to challenge some standard industry practices.
Understanding the use cases
Early adopters of agentic AI often face barriers when the time comes to move experiments into mass production. As agentic AI is still evolving, Thoughtworks experts advise businesses to apply it to processes that move through a series of well-defined workflows to clear outcomes.
Recent Thoughtworks projects show that when AI is applied to the right use cases, it can significantly improve efficiency and speed up tasks like information retrieval—even if the vision of fully autonomous AI employees is still a way off.
Agentic AI solutions providers are ready for prime-time
Addressing the risks
As agentic AI gains traction, the downsides, from hallucinations to new security vulnerabilities, will become more apparent. Using agents will never be risk-free, making it essential to conduct honest assessments of how the potential negative impacts stack up against prospective gains.
Thoughtworks research aims to support this process by making AI risks more measurable, and by boosting the transparency and explorability of systems that were once ‘black boxes.’ Trust has to be an embedded feature for agents to reach their full potential.
Building the right infrastructure and capabilities
Enterprises experimenting with agentic AI quickly learn a mature data platform and data practices, and dedicated expertise are prerequisites to deploying it at scale. For many this creates a ‘buy vs. build’ dilemma, where investment needs to be made in new products, or specialized technology talent.
The dangers of being locked into a single vendor’s vision or confronting skills shortages mean the best approach is often a hybrid one that includes upskilling, hiring, taking advantage of finished products and finding the right partners. Regardless of the formula, the goal should be the same: to develop systems the enterprise can call its own.
"You can get to a simplistic solution that looks and feels good very quickly. But the challenge lies in taking that into production and making it scale, cost effectively, accurately and reliably."
Prasanna Pendse
Global VP of AI Strategy, Thoughtworks
Preparing teams for augmentation
The main challenges raised by agentic AI have as much to do with management and adoption as they do technology and resources. Advances in AI often spur new controversies and anxieties. These create internal resistance that can only be overcome with strong executive sponsorship and an inclusive approach to AI projects that ensures those whose roles will be affected have a seat at the table.
Rather than dismissing the fears around AI out of hand, Thoughtworks experts advise business leaders to be frank about the technology’s capabilities and limitations, and to speak in terms of benefits systems can generate for the organization and the people within it.
Agents getting bigger, and better
Even if agentic AI is in the early stages, there are reasons for optimism about its prospects. Over the next few years, agentic AI is set to become both more democratized, with a higher margin of safety and reduced barriers to entry, and more tailored to specific users and purposes.
The effects of agentic AI will be more immediately apparent in some industries than others, but virtually every major company is expected to develop an AI practice, as more businesses move to capture the additional value the technology is set to create.
Subscribe to Perspectives to stay ahead of the curve.
Get timely business insights, expert analysis, and industry updates delivered to your inbox when you need them—no noise, just value.
(* Required fields)