Apr 30, 2026 · 55m · in-depth
AI killed the product roadmap | Diya Jolly (CPO & CTO of Xero)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this in-depth interview, Diya Jolly, CPO and CTO of Xero, explores the evolving landscape of product leadership, detailing how generative AI transforms roadmaps into autonomous workflows, strategies for navigating high-stakes ambiguity, and methods for fostering ambitious, high-performing cultures.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Brett holds 13.4% of the talking time here. How this is scored →
Brett as informed peer 3.4 Guest teaching 5.3 Guest disagreement 2.1 Brett pushing back 1.2
speaking balance: gold is Brett, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 28:20 Forceful agreement on raising ambition
Diya immediately and forcefully interrupts Brett to state that raising organizational ambition is literally her number one job.
Hardest push from Brett ▶ 1:50 Probing founder-led vs executive CPO nuanceBrett challenges Diya's initial generalized CPO definition by pressing whether the role fundamentally shifts depending on founder or hired CEO leadership.
Biggest teaching moment ▶ 35:00 Decision-making framework masterclassDiya gives a masterclass on bucketing operational decisions versus reserving 50% of executive cognitive bandwidth for unstructured strategic ambiguity.
Brett holds their own ▶ 5:10 Framing the founder alignment tensionBrett demonstrates sharp executive insight by diagnosing the nuanced tension CPOs face between being overly deferential or insufficiently challenging with founder CEOs.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Brett as informed peer | Guest teaching | Guest disagreement | Brett pushing back | Why |
|---|---|---|---|---|---|---|
| Defining the Role of an Excellent Chief Product Officer | 4 | 6 | 3 | 2 | Brett prompts Diya on the role of a CPO in founder-led versus executive-led companies. Diya reframes the traditional definition of a CPO shipping features to setting vision, understanding customers deeply, and dynamic resource allocation. | |
| Customer Value Creation and AI Roadmaps | 3 | 6 | 3 | 1 | Diya explains why traditional rule-of-thumb roadmap allocations break down in the AI era. She illustrates customer psychology with the automated bank reconciliation bet at Xero. | |
| Structuring Greenfield Innovation and AI Bets | 4 | 5 | 2 | 2 | Brett asks how to structure high-variance bets in greenfield innovation. Diya explains the necessity of sheltering small teams from existing org gravity and legacy constraints. | |
| Redefining Product Management and Workflows for AI | 3 | 6 | 2 | 1 | Diya outlines how AI shifts PM focus from feature-level capability delivery to holistic outcome orchestration across customer workflows. | |
| Deep Technical Architecture and Strategic Altitude | 4 | 5 | 2 | 2 | Brett and Diya explore technical altitude and organizational politics. Diya breaks down politics into misaligned incentives and gives a practical example of giving up product marketing to GTM. | |
| High Demands, Psychological Safety, and Embracing Failure | 3 | 5 | 2 | 1 | Diya articulates how high expectations must be paired with psychological safety and OKR target batting averages to avoid sandbagged roadmaps. | |
| Raising Ambition and Lessons in Executive Leadership | 4 | 5 | 3 | 2 | Brett probes whether raising ambition is a core CPO task. Diya forcefully confirms it is her number one job and describes using macro benchmarks to push past internal pushback. | |
| High-Leverage Ambiguity and Three Tiers of Decision-Making | 3 | 6 | 2 | 1 | Diya shares her three-tier decision-making framework, detailing how low-impact and well-defined tasks are delegated while ambiguous strategic pivots receive personal focus. | |
| Calendar Discipline and Deep Thinking Frameworks | 4 | 5 | 2 | 1 | Brett and Diya discuss calendar management and deep thinking tactics. Diya describes dedicating meeting-free days and switching browsers to prevent tactical distraction. | |
| Executing Strategic Bets: Melio Acquisition and FinSites AI | 3 | 6 | 2 | 1 | Diya reviews two core strategic decisions: acquiring Melio to own early small business cash flow and developing FinSites to lead conversational AI financial scenario planning. | |
| Horizon Planning and Multi-Hypothesis AI Scenarios | 3 | 6 | 2 | 1 | Diya contrasts historical multi-year planning with modern 18-to-24-month horizon planning, emphasizing holding two to four active hypotheses rather than a single bet. | |
| Navigating the Executive Suite and Enterprise Problem Solving | 4 | 6 | 2 | 1 | Diya walks through the requirements of the C-suite, managing star PMs, delivering performance management, and finding unconventional ways to protect thinking time. | |
| Building Lasting Impact and Concluding Reflections | 2 | 2 | 0 | 0 | A brief, cordial outro where Diya defines the lasting impact she hopes to achieve in her career. |
Statements from this episode (25)
CPOs must set strategic vision instead of micromanaging product features
“The key job of a great chief product officer is essentially setting the direction, not worrying about every single feature, not worrying about even big, even reasonably, reasonable sized products down to every feature level. But setting the vision and the dire…”
Org design and resource allocation must be revisited quarterly
“People think resource allocation or organization is a one-time thing, but it's not. The world around us these days changes every quarter. So these are decisions you have to revisit almost every quarter.”
CPOs under founder CEOs must surrender product autonomy to succeed
“The difference is you have to be willing to give up some level of autonomy when thinking about the product. In a founder-like company. And you have to be willing to, you have to know how to influence really well, and you have to be able to understand what infl…”
Product vision is synthesized from collaborative customer insights, not invented solo
“The other thing everybody thinks is like PMs or CPOs are supposed to come up with The vision, like, honestly, like, a vision is a collection of people coming together and understanding customers deeply enough and stitch, and then, yes, usually a CPO will stitc…”
An inability to monetize proves a product failed to deliver customer value
“I am a firm believer that if you add value to the customer, you will find a way to capture the value eventually. And yes, I almost think of it as if you cannot capture value, then you are fooling yourself that you added value to the customer, right? Or enough …”
Xero's AI bank reconciliation hits 97% accuracy, saving businesses 22 hours monthly
“So we built automatic bank reconciliation. What this means is your transactions in an accounting system are matched to your cash in your bank. Our customers are like, no, no, no, no, no, no. You can't do this. You can't do this with the machine. There's so muc…”
Users distrust AI automation without a transparent, verifiable audit trail
“And then the thing we learned and we iterated our way into was people are uncomfortable till they can see what you did. The second you can show them exactly what you did, People are happy.”
Product teams building solely for current customer requests will fall behind
“If all you're doing is what customers are asking for right now, I think you'll get left behind.”
Jolly: Small teams can now validate product directions quickly
“Resource allocation used to be a problem. Today, you can get a lot of certainty whether you're in the right direction or not with a very small team reasonably fast.”
Internal business model pivots fail without structurally isolated tiger teams
“SaaS versus a license-based model versus if you wanted to build a transaction-based model. Super hard to do it in the current system. Existing customers will revolt. Billing systems won't work. Sales incentives won't work, but if you put someone on the side, t…”
Modern product design shifted from delivering capabilities to collapsing workflows
“It is more now, the way you need to think is, how do I make their workflows easier, right? It's shifted from, here's the capability, do they do utilization? To how do I take what exists in front of me and collapse the workflows and make it easier so that in a …”
AI transformations require senior product executives to operate at lower altitudes
“Here's why you're flying at a lower altitude, because you're trying to change the way people work. You're trying to change people, because you as one of, as you as the senior product and engineering leader are the catalyst for change right now, right? Because …”
Executive intervention in cross-team disputes incentivizes internal political lobbying
“When a leader jumps in and solves problems across their teams, there is a lot of reason for the teams to lobby for their self-interest, right? Which then creates a me versus them.”
Jolly: High demands without psychological safety cause roadmap sandbagging and stagnation
“If you demand without giving people the support to succeed, And letting them know that taking risk is okay and failure is okay. You're going to have a culture where nobody takes risks. Nobody tries to grow. And everybody is, it's a fear-based culture. There's …”
Xero built and shuttered an experimental AI chat-based accounting app
“So for example, we took a bet where we said, and I did this, I'm like, I don't want to wake up in two years and find out that the entire App is agents in a chat. So we're going to go off and build our own app on the side. That is where you're going to do accou…”
Jolly: Senior product leadership requires spending more time thinking than doing
“And I think the more senior you get in any role, especially in the product role, it is about making a few decisions that are big and irreversible. And you need more time to think than to do.”
Jolly: Highly ambiguous strategic decisions cannot be delegated by executives
“I put decisions in three buckets. One is, like, impact on the business, but, like, not make or break for the business. Those are delegated down until somebody escalates. Decisions that have big impact on the business, but what you have to do is fairly certain,…”
Product leaders must spend half their time thinking to remain competitive
“Right now, if you are not spending 50% of your time thinking about what
Your product will look like, what a value it will deliver.
It's just gonna be hard to compete.”
A 100% success rate on major decisions indicates insufficient executive risk-taking
“If every one of the big decisions you take turns out correct, you're doing something very wrong. Like, you're not taking enough risk. Something's very wrong. So the batting average, again, like, I say this, like, on big decisions is, like, if your batting aver…”
Xero's AI financial chatbot hits 95% accuracy on scenario and advice queries
“One is like, we're ahead in the market. Nobody else does it. Second, and we're at like, 95% accuracy or something, like in the answers written on.”
Leaders must maintain multiple probabilistic hypotheses about future market trends
“I think today, if you have one hypothesis on how the world will work, very dangerous. If you have, I'd say, somewhere around two to four hypotheses of where the world will be, and then your probability is on that, and then you line up your bets accordingly.”
Compliance demands ensure traditional accounting SaaS workflows will not disappear
“Will SaaS apps completely disappear? I, will the workflows disappear? SaaS is an overload of terms. Probably not, at least in our category, just because it's compliance. You need, like, you need the workflows to prove that you're doing the right thing to submi…”
Jolly: Carving Out Thinking Time Is Hardest for Middle Managers
“You have the most flexibility to carve out thinking, thinking time, I think at the two ends. Either you're very senior or you're very junior. In the middle, it's very hard to carve out your own thinking time because you're at the behest of too many people and …”
A single underperformer drags down total team performance by up to 30%
“Even one person can push, pull a performance of a team down by 20, 20, 30%, right? Because other people are carrying the weight.”
Jolly secretly called in sick as a director to protect thinking time
“When I was a director, I literally would call in sick to get my thinking time once in a while, because I'm like, I just like, there's like, and I would tell my manager, I'm doing this because you'll respect the fact that I'm, I need to get stuff done, and I ne…”