Those AI copilots that tech companies have dispatched to join the office are getting a promotion.
Agents—autonomous systems that can perform tasks beyond the scope of a chatbot—are fast replacing copilots as the buzzword du jour in the race to outfit generative AI for the workplace.
Salesforce has retooled its Einstein Copilot into a new product called Agentforce, which became generally available last week. Microsoft Copilot isn’t going anywhere, but the tech giant recently added a studio to create agents, as well as a cast of 10 pre-built agents to serve functions like sales, finance, and supply chain. Plenty of others, from Palantir to Asana, have thronged to the build-your-own-agent space as well.
“In contrast to now-outdated copilots and chatbots that rely on human requests and struggle with complex or multi-step tasks, Agentforce offers a new level of sophistication by operating autonomously, retrieving the right data on demand, building action plans for any task, and executing these plans without requiring human intervention,” Salesforce touted in its announcement.
(“Outdated copilots” could be read as a swipe at Microsoft; Salesforce CEO Marc Benioff derided his rival’s Copilot product and its foray into agents in an X missive last month. Microsoft and Salesforce did not immediately respond to requests for comment.)
A new role: Agents essentially aim to shift the AI from the copilot’s chair to the driver’s seat. Rather than just answering questions or spitting out bits of copy at a human’s behest, they can perform multi-step tasks on their own.
Anthropic’s new “computer use” feature for Claude Sonnet 3.5, in which the bot can commandeer your computer—cursoring, clicking, and typing as it goes—is one illustration of the difference.
In enterprise contexts, agents might be assigned to research and prioritize sales leads and reach out to customers, as one of Microsoft’s new offerings is advertised, or whip up a campaign brief, target audience segment, and content for a marketing campaign, as one of Salesforce’s prefab agents can supposedly do.
Keep up with the innovative tech transforming business
Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.
Amalgam Insights CEO and principal analyst Hyoun Park told Tech Brew that agents work best for jobs with structured processes, but open-ended roles, like fielding customer service calls, could prove trickier.
“The real value of these agents is coming from structured areas where it is easy to take multiple steps, such as coding or accounting or certain types of manufacturing operations where there is a discrete next step that can be taken based on a first step,” Park said.
Trusted agent: But the switch to agents also comes at a time when many businesses are still figuring out how to best implement generative AI at a scale beyond prototypes. In that environment, can companies trust AI with self-directed decision-making over key operations?
Park said AI companies should be upfront about how agents work and what their strengths and weaknesses are.
“It helps to be able to show some of the lineage and logic associated with designing the agents,” Park said. “Agents need to at least provide some guidance that they may not know what they’re doing, so that they can show users where to show some additional caution.”
These kinds of agents are still very nascent, though. While there are certain simple tasks that agents can easily field right now, Park said, more complex and varied workplace processes could take years to get right.
“As we get closer and closer to the more human aspects of work, it will be more of a challenge to create those agents, and we should expect that those will take several years to polish, despite the fact that Microsoft and Salesforce want to sell them to us right now,” Park said.