2 days Ago By Jacob Clemente
It’s rare to come across a case study that addresses as many of Charter’s current AI and work obsessions as one we recently learned about from Asana, the work-management software company. It has elements of employee-led , peer-to-peer learning and collaboration, and even early hints of . It started in January 2023, when , an Asana sales development representative at the time, started experimenting with ChatGPT to see if it could help him personalize his customer interactions, while also saving him time.
Underwhelmed with the results, he decided to give it another try months later with OpenAI’s newest model, GPT-4. “I was like, ‘Oh my God, it works’..
.That led me to get more excited about experimentation.” This type of exploration was made possible by Asana’s approach to adoption.
“What we have tried to do is allow for very broad-based experimentation with AI,” explains the company’s head of engineering, , who adds that any employee who wants ChatGPT Plus can purchase it through their learning and development budget. “Obviously we paired that with guidelines [on]..
.what data is okay to put in which AI accounts. There are some [situations] when we are using the enterprise-tier accounts, where we've signed contractual agreements which protect the data in certain more ways,” says Pandey.
To encourage employees to share what they’re working on, the company has a Slack channel called “AI Inspiration and Adventures.” “People just post about everything from ‘Here's a cool image I made’..
.[to] ‘look how I automated this workflow using what Asana has built for AI but also some of our workflow features,’” says Pandey. The company has also hosted a few show-and-tell sessions where people can present to their colleagues the different uses they’ve developed.
The sharing element is particularly useful for disseminating applications beyond the company’s most active users. As Pandey told us, a small share of the company’s workers is actively experimenting and developing new uses—but, he adds, many more will adopt the examples they see. DeWaal was an active member of the Slack channel, where he learned how to prompt from the company’s AI engineering team.
“Through that community of learning, some of the more cutting-edge stuff that the engineering team was working on got transferred to me, and I was able to translate some of that into go-to-market [uses],” he says. Still determined to use AI to help him with different parts of his job, DeWaal started creating his own bots using the OpenAI API on top of the Asana platform. The AI community within Asana helped DeWaal by answering questions whenever he hit a wall, he says.
After developing those bots, he wanted a better way to share them with his colleagues. “People were copy and pasting in a super inefficient way,” he explains. During a company-wide hackathon, he partnered with an API engineer and an infrastructure engineer to create an internal app store, which he describes as a place “where people who might not have interest in ever learning how to prompt-engineer could go.
..and have more of a prepackaged experience with AI.
” Employees can search the app store and pick the bot that’s best for their specific task. DeWaal tells us that there’s been “significant adoption” of bots in different departments within Asana. As a result of those efforts and others, DeWaal became Asana’s AI go-to-market program manager, where he helps drive adoption within the company’s business organization.
He’s also the owner of one of the company’s AI goals to integrate similar bots to the ones he built into a sales workflow. Pandey explains: Here’s how Pandey describes the current vision for the workflow: To continue reading, subscribe to the Charter Briefing newsletter to stay ahead on the future of work. Continue reading by subscribing to about the future of work.
If you’re already a Charter newsletter subscriber, this will only verify your email address..
