Presenters
Source
🚀 The Future of SaaS: From Clicking Buttons to Conversational Insights
In the rapidly evolving world of software, we are witnessing a fundamental shift in how humans interact with technology. According to Diya Jolly, the Chief Product and Technology Officer at Xero, the era of clicking buttons all over a dashboard is ending. Instead, SaaS is transforming into a space where users primarily review insights and engage in natural conversations with their data.
In a recent episode of the Product Podcast, hosted by Carlos Gonzalez de Villaumbrosia, Diya shared her blueprint for scaling complex platforms and integrating AI into a global business serving over 4.5 million customers across 180 countries. With a market cap of $9 billion and $1.2 billion in annual revenue, Xero provides a masterclass in balancing innovation with the precision required for financial services.
💡 The Core Fundamentals: Empathy at Scale
Whether you are building for advertisers at YouTube, enterprise security at Okta, or small businesses at Xero, the core of product management remains the same: customer empathy.
Diya emphasizes that understanding the jobs to be done is non-negotiable. While consumer products allow for rapid iteration based on personal use, B2B requires a deeper, more intentional effort to touch the user’s reality.
🛠️ Strategies for Deep Understanding:
- Shadowing Workflows: Diya’s team spends entire days with accountants and bookkeepers to observe their daily struggles.
- Advisory Councils: Xero maintains both a Small Business Advisory Council and an Accountant/Bookkeeper Council to stay close to market needs.
- The “Driver” Mentality: Much like Uber PMs must spend time as drivers, product people must use their own products (demo orgs) and set up competitor software to understand the landscape.
🚪 Decision Making: One-Way vs. Two-Way Doors
One of Diya’s most striking philosophies involves making decisions with imperfect data. She adopts the Amazon-coined principle of One-Way and Two-Way Doors.
- One-Way Doors: These are high-stakes, hard-to-reverse decisions. Pricing is the ultimate example. Diya notes that it doesn’t take a genius to reduce prices, but raising them is incredibly difficult. Therefore, the strategy is to start high and maintain the option to offer offsets or lower prices later.
- Two-Way Doors: Roughly 80% to 90% of daily decisions fall here. These should be treated as experiments. If a feature doesn’t work, you simply walk back through the door and try again.
The Challenge: In large organizations, committees can stall progress. Diya combats this by establishing clear ownership. For AI initiatives, she empowers a “pod” of leaders (AI, Product, and Engineering) to make the final call, holding them accountable via adoption and value metrics rather than seeking total consensus.
🏗️ Org Design: The CPTO Model
Diya holds a unique dual role as Chief Product and Technology Officer. To maintain high standards, she structures her team by functional depth (Engineering, Product, Design, AI) but operates through virtual pods.
- Functional Hardlines: Stellar engineers want to learn from stellar engineering managers. This keeps the craft sharp.
- Virtual Pods: PMs, Designers, and Engineers form a dedicated squad with a common mission for a specific product area.
- Strategic Flexibility: Diya adjusts the structure based on industry shifts. Currently, she maintains a singular Engineering leader to oversee massive platform changes required for AI data flow, ensuring information moves seamlessly across the organization.
🌐 Scaling Globally with Hyperlocal Precision
How do you build a product that works in 180 countries when tax laws change at every border? The answer lies in Platform Thinking.
The Tradeoff: Building a different app for every country is impossible to scale, but building a single “one-size-fits-all” app ignores local regulations.
The Solution:
- Identify Common Surfaces: An invoice looks roughly the same everywhere (items, costs, customer names). Build these as core services.
- API-First Flexibility: Use APIs to connect the common surface to regional-specific engines, such as local sales tax calculators or country-specific payment rails.
- Regional Teams: Empower local teams to manage the specific regulatory nuances while the core platform remains stable.
🤖 The Rise of the “Super Agent”
The most provocative part of the conversation centered on Jax (Just Us Xero), the platform’s new Super Agent. Diya argues that slamming an AI chatbot on top of an old UI is a mistake.
🎯 The “Jax” Philosophy:
- Orchestration: Instead of a dozen disconnected AI features, a Super Agent manages specialized sub-agents (payroll, invoicing, bank reconciliation).
- Conversational UI: Users can simply text Jax on their phone: Hey, what’s my cash flow? or Should I lease or buy this delivery van?
- Contextual Intelligence: Jax doesn’t just give a number; it looks at historical data. It might warn a bakery owner that summer is a slow season and suggest waiting until August to expand the fleet.
🛡️ Guardrails and the “Human in the Loop”
In accounting, 100% precision is the only acceptable standard. LLMs are notorious for “hallucinations,” which is a non-starter for financial records.
Q: How do you ensure accuracy in a regulated environment? A: Diya implements a dual-layered safety system:
- Algorithmic Guardrails: Xero uses traditional, non-AI algorithms to check LLM outputs. If an LLM returns a value outside a pre-calculated “safe bound,” the system blocks the answer.
- Human-in-the-Loop: AI is used to speed up work, not replace knowledge. For Bank Reconciliation, the AI suggests matches, but the accountant is given an easy workflow to “eyeball” and approve them. This saves users over 22 hours per month while maintaining total control.
🎯 Why Focus on SMBs?
While many tech giants chase the Fortune 500, Xero remains intentionally focused on the Small and Medium Business (SMB) segment.
The Strategy: Diya believes that trying to serve everyone results in a hodgepodge product. SMBs are a unique “hybrid” persona—they require the simplicity of a consumer UI but the robustness of an enterprise workflow. By staying focused, Xero can solve the specific “daily pain” of the 1-to-50 employee business, providing a “junior employee” in the form of AI that helps them run their business more intelligently.
✨ Key Takeaway
The future of SaaS isn’t about more features; it’s about less friction. By combining deep customer empathy with a conversational “Super Agent” layer, platforms like Xero are turning static systems of record into dynamic partners in business growth. 🚀📈👨💻