Asana AI Studio now offers AI agent creation for workflow management

The number of platforms being released to help enterprises integrate AI agents into their technology stack is not slowing down as the year winds down. 

Work management platform Asana already has an AI agent service, but with its new AI Studio feature, the company wants its customers to think about AI agents as part of their larger workflow.

AI Studio lets users build workflows on Asana and then deploy multiple custom AI agents directly on the workflow. Customers can create the agents without code and allow them to “take on the busywor,” including handling project coordination. 

“The difference here is that we’ve opened up the toolkit to create agents to the folks who build workflows at companies, the folks that orchestrate a large body of work,” said Alex Hood, Asana’s chief product officer, in an interview with VentureBeat. “We can now bring agents to all the places where teams show up to hand off work, inserting an AI agent to take work off people’s plates.” 

Hood cited a recent Asana 2024 State of Work Innovation Report that showed 53% of an employee’s time is spent on “busy work,” with unproductive meetings doubling since 2019. By bringing in agents, Hood said, teams are freed from doing tasks crucial to a workflow, such as intake for some marketing teams, to focus on other important work. 

AI Studio, which will be a tool integrated into Asana but with an additional fee for access, is built on Asana’s Work Graph data mode, which tracks cross-functional work in an organization. 

“There are other platforms who are building AI agents and are doing it on top of places where they have specialties,” Hood said. “For us, our specialty is the Work Graph, the place where work happens and the workflows powering that work leverages the right data.”

Customers saw an improvement

One of Asana’s first customers to use AI Studio is the financial data company Morningstar. 

Hood said Morningstar used AI Studio to centralize IT project requests to streamline the workflow to evaluate new projects.

Belinda Hardman, director of Program Management at Morningstar, said in a press release that the new workflow helped the company “eliminate time spent on manual back-and-forth because Asana AI identifies and captures the information we need right off the bat.”

Islands of AI agents

Agents have become the hot topic in AI this year, with several companies announcing either a platform to customize agents or to access a library of ready-made agents.  To name a few: Microsoft announced it will release a suite of AI agents for its Dynamics 365 service this week. Salesforce released Agentforce last month, and ServiceNow launched its agent library on its Now Assist platform. Asana’s earlier released agentic system joins Agentforce, and other ready-made agents from other service providers will be integrated into Slack. 

“The things we might have dreamed of and were talking about a couple of years ago are playing out now because the models are getting that much better,” Hood said. “But the models can’t create great agents on their own. They need to be hooked into software and we, as builders, have gotten good at figuring out how to best integrate deeply AI capabilities.”

However, many of these agents — even those embedded in third-party applications like Slack — still function as individual islands of agents talking to other agents built on the same platform. The next frontier for agents coming from workflow systems or other enterprise-focused software will be the ability to communicate with other agents elsewhere. 

We’re not there yet, but as more enterprises become comfortable with AI agents and begin deploying these into their organizations, that future may come soon enough.