Organizations are spending billions on AI — as much as $207 billion this year on AI agent software alone , Gartner forecasts. Even so, a recent survey conducted by Atlassian found that most senior leaders hesitate to use it for more than daily tasks, and only 26% say they trust AI for making strategic decisions .

That hesitation has less to do with AI itself than with the fragmented operating models it’s drawing on, says Asha Thurthi, Atlassian’s senior vice president of product for Strategy Collection, the company’s solution for AI-native strategic portfolio management .

“Leaders have every right to be cautious,” Thurthi says. The real issue, she argues, isn’t AI itself but whether an organization has a connected operating model that AI can reason over. Strategic decision-making requires context from across an enterprise — investments, goals, progress and AI use. Most organizations, however, still rely on scattered meeting notes, slide decks and reports rather than a connected system.

Deploying more AI on top of a fragmented operating model only widens the problem, Thurthi says. Even the most sophisticated models can’t deliver reliable guidance without the organizational context to reason effectively. Before spending more on AI, she recommends, organizations must connect their systems first.

The Connectivity Advantage

Better AI outcomes aren’t about picking a smarter model, Atlassian’s research finds. Organizations that are taking full advantage of AI are those whose strategy, investments, work and execution data live in a unified system. It’s that connectivity, not which AI model they use, that is the biggest predictor of whether AI earns leaders’ trust and drives real results.

Linking daily work with strategy also helps teams move faster, the research found. Teams unblock each other and catch risks more quickly. Connectivity was the single biggest factor behind both those improvements.

“The question isn’t, ‘Is AI driving productivity for your organization and for your talent?,’” Thurthi says. “It’s really, ‘Is AI driving the right outcomes for the company?’”

Building the Context AI Needs

Organizations won’t close that trust gap until they connect the fragmented systems undermining AI’s reasoning. “The risk is high,” Thurthi says. “Without fixing the underlying connections, companies don’t get a living signal from a living system. Strategy drifts — and the AI drifts with it. They’ll see false positives. They’ll end up optimizing productivity instead of optimizing outcomes — and that’s my real fear.”

Making the shift to a connected operating model requires the right sequence, Thurthi says. Connect the operating model first, then use AI to amplify it. When goals, work, dependencies and investments are connected, AI has the context it needs to reason reliably about strategy.

“The foundation needs to be connected first,” she says. “If planning and delivery are disconnected, AI can’t understand whether the company is executing its priorities.”

Atlassian’s Strategy Collection is one way to connect strategic goals to daily execution. Powered by Teamwork Graph, the company’s data intelligence layer, Strategy Collection links priorities, portfolios, people, work and investments in a single system. Rovo, Atlassian’s AI, uses that context to help leaders spot issues and make better decisions.

With that foundation, Thurthi says, organizations can see whether their AI spend is tied to the right work and outcomes and adjust if needed. “It unleashes decision-grade intelligence for leaders, which is definitely high signal,” Thurthi says. “I call it decision-grade rather than strategic intelligence. Leaders really understand their entire system in one snapshot.”

Organizations that solve this problem first will have a structural AI advantage that competitors can’t simply buy by spending more money on more powerful models or another AI tool, she says.

Speaking at Atlassian’s Team ’26 conference, Ben Waddle, vice president of strategy, planning, delivery and product marketing at Expedia Group, described his first week on the job. His manager, the chief product officer, had asked him for a product roadmap.

Waddle recalled the “terror” of realizing the company’s product roadmap lived across multiple disconnected systems and hundreds of individual desktops.

At the time, Waddle had no single source of truth. Keeping leadership current meant time-consuming manual work that often produced outdated information.

Rather than layering on another tool, Waddle extended Atlassian’s Jira as the system of record for strategy and product roadmaps, creating a trusted foundation that AI tools, including Rovo, can draw from.

The change required more than technology. “This wasn’t just a tooling shift. It also required alignment on ways of working,” Waddle said. “Change management was absolutely something I underestimated.”

The payoff showed up immediately in how Expedia Group’s leaders planned and made decisions. Quarterly business review preparation, once a multi-week scramble involving more than 50 product managers, now takes about a day and runs entirely from Jira. “I knew it was working when we held a quarterly business review entirely out of an automated dashboard in Jira across our top priorities — no slides, no docs,” Waddle said.

Over eight months, Expedia Group also narrowed thousands of competing initiatives to 10 stack-ranked priorities, each with roughly four measurable objectives and key results (OKRs). An annual replanning cycle that once took more than three months now takes four weeks. By first connecting its planning and delivery systems, Expedia Group created the trusted organizational context that made AI-supported decision-making more effective.

Expedia Group’s experience shows why organizations should connect their operating model before scaling AI, Thurthi says. “The sequence really matters. Connect the operating model first … then use AI to amplify decision speed and outcome management.”