AI Transformation Starts With Skills, Not Hiring:
Atlassian's Chief Product and AI Officer

Tamar Yehoshua argues that a large company cannot hire only AI-native people the way a small company can. It has to train the people it already employs, and give them time to learn. In her words, "Small companies can say we're only going to hire AI-native people. If you're a large company, you got a lot of great people that you want to train them." Yehoshua is Atlassian's Chief Product and AI Officer.

One of the biggest obstacles Yehoshua hears about is time. The complaint, she told CXOTalk, is "I'm so overworked. I don't have time to learn how to use the tools." Her answer: "You've got to give them that space, and people are so appreciative of that time and space."

Strategy comes after the first business process

Asked by a viewer how a company reimagines entire business processes instead of one at a time, Yehoshua reversed the order: "You have to start by gaining the skills. And so you may start with one business process." Only then come the strategy questions: "What is your North Star? What are your differentiators?"

Host Michael Krigsman objected that this was a matter of organizational dynamics and change, not an AI issue. Yehoshua answered that AI gives a company a tool for that work, which is why the learning comes first. She named one thing that goes wrong: "So one issue I think people do is they go too large with AI. I'm going to transform everything, and those usually fail. That's the boil of the ocean." Her answer ended by agreeing with him: "The bigger prize is just management, leadership."

Atlassian stops product and design work one week a quarter

The training week ends with a project that staff builds and then showcases, something like a demo day. "One of the tools that's been the most effective for us is something we called AI Builders Week," Yehoshua said. The format: "Once a quarter, we take a week out of everything that everyone's doing in product and design, and we train them on a subject. So it might be prototyping or evals or agent building or coding."

Her measure is how much of that work gets used afterward: "70% of what's built during Builders Week actually gets used later." The lesson: "it's really important that you don't just tell people to use AI, but you help them along the path, and you give them the time and the space."

Atlassian runs its AI experiments on itself

Yehoshua says Atlassian, "the zero-day customer" for its own AI, set "goals for each department, for engineering, for finance, for sales." Telling people to "go forth and just use AI" yields prototyping and learning, which she grants has its place.

One team working on Confluence slides was asked: "How can you reimagine all of the workflows?" One result: "They took the designs from Figma and then the code and said, 'Where are the deltas between the designs and the code?' And they used Figma's MCP to automate that gap and coding agents to fix it automatically. So they fixed 14 bugs in an hour, which would have taken days." Her explanation is the goal itself: "There was a task and a goal you gave to a team to achieve something. And then it's amazing what they do."

Atlassian judges token spend by engineering output

Yehoshua measures AI spending by what engineering ships. She does not treat tokens and SaaS licenses as a trade-off. Asked on LinkedIn how she budgets licenses against tokens, she answered, "Value-driven." Most of Atlassian's internal token cost sits in engineering, so the yardsticks are engineering ones: "We look at PRs deployed per engineer. We look at features developed, and we've seen a huge uptick in PRs and features developed."

Her case to the CFO is output: "working with the CFO to say, well, we're delivering more value. Our products are being deployed faster." She does not see licenses and tokens as substitutes: "I don't think of it as tokens versus SaaS." Atlassian keeps Databricks and Salesforce and now uses their AI features too, "because if you have a need for those products, those needs don't go away," at least so far.

The AI gateway controls model cost in Atlassian's products

Atlassian invests significant time in its AI gateway, which Yehoshua says helps optimize costs. In her words, "we do not use one model in our products. We use a gateway that then routes to the model that's the most appropriate model for the job that takes into account quality, latency, and cost." The gateway also keeps Atlassian from being locked into a single model. She adds: "There's tons of things where we don't need the most expensive model by default. We can use a lesser model."

Executives have to use the tools themselves

Yehoshua holds that leaders who do not use AI tools themselves cannot see how their teams will change. "Our CEO, Mike Cannon-Brookes, when OpenClaw came out, got super pilled on OpenClaw and was building agents to run his personal life and still does." He then told his executive team: "You all need to implement OpenClaw just so that you can see what everyone's talking about."

The same bar applies to every executive: "If you're not, then you don't really understand how this is going to change your teams and how they work." Some teams move faster than others, and Atlassian tracks the gap. For now, "we're using the carrot, not the stick."

Yehoshua's larger point is speed. "We are launching so much faster than I've ever seen in my career," she said. The condition is that people can act without waiting: "you have to empower people to do things without approvals and then have a culture where they will escalate when they need." Her counsel to leaders is to aim higher than their teams think possible while understanding the tools themselves, because "these tools can do more than what people are using."

Watch the full conversation with Tamar Yehoshua and read the complete transcript on the episode page.

CXOTalk prepared this article with AI assistance from the verbatim transcript of episode 930. Quotations are unedited from the broadcast.