The AI world is abuzz with talk of agents – Not the ones which would have Ian Fleming or LeCarre fans in a tizzy, but autonomous AI-based software systems that perceive their environment, and your preferences to make decisions, and take actions to achieve specific goals without direct human intervention. Agents or Agentic AI first caught my attention through a prescient Bill Gates blog (https://bit.ly/3tSMNkB ), where he declared: “In the next five years…you won’t have to use different apps for different tasks. You’ll simply tell your device, in everyday language, what you want to do. And depending on how much information you choose to share with it, the software will be able to respond personally because it will have a rich understanding of your life. This type of software—something that responds to natural language and can accomplish many different tasks based on its knowledge of the user—is called an agent.” He later went on to claim: “Agents are not only going to change how everyone interacts with computers. They’re also going to upend the software industry.” Agents, he declared, are the new platform.

So how are agents different from apps, and why are they creating the new AI gold rush? As Bain Capital’s Sarah Hinkfuss elegantly describes (https://bit.ly/3Ye4reU) it: “we are used to ‘pulling’ information from computers; we need AI to ‘push’ finished work to us instead”. Instead of tapping apps multiple times or ask ChatGPT a series of questions to work out a complex and frustrating travel itinerary, what if, say, a Booking.com agent would select a hotel and airline which we usually like, design a day-to-day schedule based on our interests, and go ahead and book the tickets and restaurants, since we have given it the permission and the agency to. So, this autonomous AI based system makes decisions and actions to achieve our goals basis the travel environment and what it knows about our travel history and preferences.

Agents have started proliferating around us, as thousands of startups amp up their innovation agents to build on top of the Large Language Models. Genie by Gather helps parents manage their time, Minday scours the Internet and mines your preferences to find the best restaurant or shop around you. Relevance AI automates prospect meetings for harried sales reps and Greenlite creates digital compliance ‘workers’ for heavily regulated banks and financial institutions. Klarna, a European fintech, made waves when its CEO announced that customer service agents built on OpenAI platforms have ‘replaced’ 700 human agents, that they resolve queries in a fifth of the time it took humans, and this would translate to a thousand less human agents plus a $40mn positive bottom-line impact!

Unsurprisingly, BigTech is all over agents, and the last three weeks have been abuzz with agentic AI announcements. At Salesforce mega summit Dreamforce, Marc Beinhoff launched Agentforce, its agentic AI suite. Oracle CloudWorld announced 50+ AI agents created with its cloud applications. The eight-hundred-pound AI gorilla Microsoft released Copilot Agents in the Wave 2 of its 365 copilot suite. Pioneer OpenAI announced a partnership wit T-mobile to create Ai agents for their customer service platform. In fact, OpenAI’s started the whole wave with its Custom GPT store, where about 4mn of them jostled for space. These Custom GPTs were nothing but proto-agents’

AI agents would have a major impact on work, business, society, and even humanity. I have written earlier about how the end of the app store is nigh, with agent platforms replacing them. The mighty smartphone will not be spared too, and the next evolution of it will be agent and LLM

driven, rather than app and OS driven (like the Rabbit R1). AI will be the new UI, where we will ‘talk’ to agents on our devices to get work done, not tap or click on them. In the enterprise, a new SaaS – Service as a Software, will replace the traditional SaaS (Software as a service), with agents bundling up services and workflows to offer them as software. Workplaces will need to expand their definition of DEI, to create inclusive workplace environments for non-human actors.

The ethical impact of this is immense, and I will write this in more detail in a subsequent article. However much the intention is to have humans as a pilot, and the agents as a co-pilot, giving them agency would mean putting them on autopilot mode. We would need to take creation of guardrails and regulations to a new level, as AI agents proliferate among humans. Well known AI author Max Tegmark is not wrong when he says that 2024 will be remembered as the year of AI agents and “they will be more of a new species than a new technology.” How we work and live with this powerful agentic species we have created ourselves will define humanity in the near future.