Agentic AI is a 'force multiplier' for small businesses - AWS - Rahul Pathak — Amazon Web Services’ Rahul Pathak

Start-ups and small businesses may be best placed to take advantage of new ways of working made possible by agentic AI systems.

The agentic artificial intelligence wave is set to reinvent the way businesses design and execute their workflows, allowing them to use the power of generative AI to supercharge productivity.

Like other cutting-edge technologies, enterprises – because they are able to dedicate vast amounts of resources to exploration and implementation – are at the forefront of agentic AI. However, small and medium businesses – as well as newly minted start-ups – may have the most to gain from using agents to deliver upsized value to their customers.

“I do think customers of all sizes see the opportunity, and in the enterprise world they have scale, resources and vast amounts of data,” said Rahul Pathak, vice president for data and AI GTM (go to market) at Amazon Web Services, in an interview with TechCentral at the company’s Seattle headquarters in the US.

“I think this idea of ‘lean in or be left behind’ applies at all levels, not just enterprises. I would say one driver of adoption for AI systems at the medium and small scale is that AI creates leverage. With AI and with agents, smaller companies can actually do much more than they might otherwise be able to do. It’s a huge force multiplier.”

According to Pathak, agentic systems represent a paradigm shift in the way humans and AI systems interact. He explained that interactions with generative AI applications, including OpenAI’s ChatGPT, Google’s Gemini and Microsoft’s Copilot, are based on a request-response model where users get answers to their queries almost immediately.

Agentic AI systems augment this paradigm by allowing for more complex interactions, enabling users to give the agent a complex task to perform over longer periods of time, where the answer could take, say, a week to produce. By analogy, prompting an AI agent is like instructing a human to perform the same task.

Autonomy

“From an agentic perspective, what we see is customers looking end-to-end across the workflow and having a system that can intelligently adapt to various decision points that happen along the way. In complex processes, autonomy is a big part of what agents do,” said Pathak.

The power agentic systems gives to their users can be used to augment existing processes with AI to drive productivity gains. But there is a greater opportunity lurking there, which requires that businesses look at their workflows creatively to see how they can be reinvented with agents as components in the process, said Pathak.

Compared to large enterprises with a long history of execution and traditional ways of doing things, start-ups can implement novel workflows without having to give credence to legacy. Small and medium enterprises have a similar advantage in that they tend to be nimbler, allowing for the implementation of new processes quickly.

Read: OpenAI bets $38-billion on AWS in cloud power grab

“Start-ups are a hugely important segment, and they are deeply embracing AI, both from the perspective of building it, but also from the perspective of employing it. We have small companies doing massive things because they are heavily leaning into agents for software development and business productivity.”

Implementing workflows that make use of AI agents does come with its own set of challenges. For users – who tend to be the employees that would, in a pre-agent paradigm, be the ones executing the processes themselves – a change in perspective is required. The employee is no longer the executor of the process but an orchestrator using multiple agents to coordinate its various parts until the whole is complete.

Another challenge hampering rapid uptake is an organisation’s level of digitisation. Digital businesses with strong data governance procedures are in a good position to implement AI tools since the data AI systems require is easily available and easy to ringfence.

However, according to Pathak, there are opportunities for organisations at all stages of their digitisation journeys to take advantage of agentic AI.

“The nice thing about the AI world is that the product offerings that we have are able to meet customers where they are. If you can use AI as a force multiplier that allows each of your employees to do more, that allows you to have much more impact, regardless of scale,” he said.