Main content

Not just another co-pilot - Asana says its AI provides 'air traffic control' for enterprise teamwork

Phil Wainewright Profile picture for user pwainewright March 20, 2024
Summary:
Asana further expanded its AI feature set this week, which rather than simply offering users a co-pilot for their own work, uses generative AI to help provide what it calls 'air traffic control' for enterprise teamwork.

Air traffic control tower at sunset with aircraft flying towards clouds © eyfoto via Canva.com
(© eyfoto via Canva.com)

While the majority of AI propositions for the enterprise these days promise to make it easier to find and generate various forms of content, work management vendor Asana is taking a different tack. This week it rolled out promised features and announced new ones to expand Asana Intelligence, which it first launched a year ago, focused on organizing and analyzing the work that's being done across the enterprise, rather than simply doing more of it. Paige Costello, Asana’s Head of AI and Co-Head of Product, says that instead of yet another co-pilot, its role is more akin to air traffic control. She explains:

This notion of air traffic control is, there are lots of planes, there's lots of complexity, and you're trying to get everything landed on time and safely. Asana knows what your company objectives are, we know what the key results are. We know who the teams and what the lines of business are. We know what your goals in Salesforce are in terms of your financial results that you're driving. We have the integrations with JIRA to know what's being shipped when, how the launches are tracking. And because Asana is a cross-functional tool, rather than a vertical tool, we have visibility into the performance across teams, and the expectations of delivery. There's a unique opportunity for Asana's AI to be aware of the bigger picture.

When you're doing your work and interacting with your co-pilot, that's very you-centric. It's very narrow and siloed and can help you do what you are trying to achieve. But it rarely takes into account the priorities of the organization and the balance of what's important and urgent. Asana understands key milestones, it understands dependencies, and it understands project duration. So when you hear me get excited about the role of the work graph, in pairing with the the power of new LLM models, that's where we can effectively deliver AI solutions that are not just your typical co-pilot but actually air traffic control for the organization, so that when an organization buys Asana, they can power up the individual's effectiveness, in tandem with the context of the organization's effectiveness.

Fundamental to this role is the work graph, in which Asana maps all the people, projects and other work entities across the organization, along with the relationships between them. This provides the grounding so that when prompts are fed to a Large Language Model (LLM), it is better able to come back with a relevant and accurate answer. Costello says:

When we pair models with the Asana data model, we're actually in a better position to help our customers move their work forward, because we know how they want to work, who's involved, what the work is, and what the next steps are, typically.

New AI features

Costello leads the AI organization at Asana, which includes a group that works closely with both Open AI and Anthropic to fine tune the models they provide and the prompts that Asana designs, while another group focuses on the user experience, such as supporting different languages. She adds:

We don't want customers to have to figure out how to prompt to get high quality value out of Asana. And so we're building prompting behind our workflows and behind our buttons so that customers don't have to think about AI in order to get the value of AI.

For example, one of the new features announced this week to roll out next quarter is called Smart Workflows, where users can describe what they want to achieve in natural language, and the AI will generate automated workflows and rules from those instructions. She elaborates:

We can take the semantic meaning from saying what they're trying to achieve, and turn it into the automations and rules that make their Asana system more robust and work faster for them, and effectively eliminate human steps.

Other new features coming next quarter include AI-assisted goal creation; smart onboarding, which allows admins to customize how people get started with Asana; and smart answers, which provides a conversational dialog in which users can ask Asana questions about work in the system. There are also new integrations to Microsoft 365, including the ability to set up automated workflows with Outlook Calendar, and to create and plan projects directly in Teams. In addition, a new sandbox facility will allow people to try out AI, automations or integrations without first having to turn them on for all users in their team.

Building on several of these capabilities, later this year there will be a workflow console, where administrators will be able to see what workflows and policies are being used in different parts of an organization and deploy new automations. Costello says:

The workflow console will be an area where you can ... see what's being used by other parts of your company and adopt similar ways of working... What's amazing about Asana is, you can say, 'Here's how we work, and apply it to this portfolio.' And then it can set those custom fields, those rules, those automations, all the way down into the projects and tasks, and really cascade this standard. So consistency and control at scale is possible with Asana in a way that's not possible with our competitors.

Other features planned for later on include automatically generated executive summaries, and smart reporting and projects, where the AI creates charts, dashboards and projects in response to natural language instructions. Previously announced features that are now generally available include multi-org deployment, which caters for differential compliance and data residency needs across a deployment, custom onboarding, and smart status, which Costello describes as follows:

That's an example of just taking something that people depend on Asana for — updating progress, keeping an eye on what's blocked, what's completed, what's next — and actually using LLMs to help users write that. It's actually referring to the work underneath it, the people doing that work, what's off track or behind, and it pulls together that first draft of their guidance... Users can even say, 'Here's some stuff that happened, please refer to the work in these three portfolios, ignore this fourth one, and make me my first draft.'

FY2024 results

In Asana's Q4 earnings call earlier this month, Dustin Moskovitz, co-Founder and CEO, spoke about how the company has been using Smart Status internally, and his own use of the upcoming Smart Answers. He said:

We've just wrapped up our objectives and key results process for the year ... Using Smart Status cuts this process in half or more, with higher quality results. And it will only get better over time.

Similarly, we just went through our company-wide 360-degree performance reviews, another famously onerous process. I personally used our Smart Answers product within each of the one-on-one projects I maintain with our executives, summarizing in minutes all the important topics and accomplishments we'd discussed in the past year. At my direction, it even highlighted all the specific times I gave recognition or feedback, so I could easily include them in the reviews.

That's not a special performance review feature we built, to be clear. I simply asked the product to help me use our general Smart Answers feature, suggesting specific categories and questions to consider. Even as an end user, I was empowered and had the flexibility to write a custom workflow in minutes that saved me hours of work across my whole team.

In the earnings call, Moskovitz talked up the impact of its AI products on enterprise adoption and its growth prospects in the year ahead. The full year results for fiscal 2024 came in at total revenues of $652.5 million, up 19% on the previous year. Net loss on a GAAP basis narrowed to $257.0 million from $407.8 million in fiscal 2023, while the non-GAAP equivalent was $45.1 million, down from $207.2 million. Asana ended the year with more than 150,000 paying customers, several with over 10,000 paying seats, and the largest deployment at over 200,000 seats. Customers with annualized spending over $100,000 grew by 100 or 20% year-over-year, with a dollar net retention rate at 115%, generating revenue growth of 29% for the fiscal year and amounting to 27% of total revenue.

For the full fiscal year 2025, revenue is expected to be in a range of $716-722 million, at a growth rate of 10-11% year-over-year. Non-GAAP loss from operations is expected to be in the range $61 million to $55 million, which is an operating margin of negative 8% at the midpoint. The aim is to be free cash flow positive by the end of the current year.

My take

The advent of generative AI has proved challenging for enterprise teamwork vendors, but Asana believes the way that it is adopting the technology is strengthening its pitch to enterprise leaders, particularly CIOs. Costello tells me:

Asana still sees itself as the co-ordination layer for companies. You can think about people's communication tools like email, and Slack and Teams, and their content, like the stuff that they're keeping in their Figma files for design, the work documents in Google, and the spreadsheets in Excel, all these things are still really critical to doing the work.

But the co-ordination of who's doing what, by when and why, is still a perennial problem in organizations of all sizes, but it's a much worse problem, the larger the company becomes. So there's more matrix management, there's more cross-functional work, there are more dependencies, more noise and more complexity. So Asana really is moving up-market and solving that problem for enterprise teams, who are trying to understand and connect strategy to execution and get work done and just have it be less painful.

In line with that move upmarket, the company is increasingly speaking to CIOs, and this week released a 2024 State of the IT Leader research report, in which it highlights the pressure on IT leaders to take charge of AI across their organizations while struggling to adapt existing IT investments and staff to the new challenges it brings. Asana believes it can help relieve that pressure, for example by allowing organizations to consolidate previously separate applications into the platform.

At a time when every enterprise vendor is rushing to refine its own AI proposition, no one is currently giving any thought to how all these different AI assistants talk to each other. There's some appeal therefore in a tool that can potentially provide some oversight of what's going on across the enterprise. Costello sums up:

Customers want more simplicity in their work, and what they're craving is clarity around who is doing what by when. Part of that is going to be clarity around what AI is doing for an organization. We are actually looking at human-in-the-loop for our AI models, but also how to effectively track the work of AI, so that humans are always in the decision-making seat, and humans know what AI has done for them on their behalf.

Loading
A grey colored placeholder image