To deploy LLMs in high-stakes environments like finance, combine them with deterministic checks. For example, use a traditional algorithm to calculate cash flow and only surface the LLM's answer if it falls within an acceptable range. This prevents hallucinations and ensures reliability.
Resist the pressure to serve disparate customer segments like SMBs and enterprise with one product. Their needs are fundamentally different. Focusing intensely on one segment allows for deeper innovation and superior product-market fit, avoiding a compromised, 'hodgepodge' solution that pleases no one.
To truly understand B2B customer pain points, data and interviews are insufficient. Product teams must immerse themselves in the customer's environment, such as by working for an advertiser for a week or shadowing an accountant for a day, to gain firsthand workflow experience and develop deep empathy.
To implement a cohesive AI strategy in a large organization, avoid siloed decision-making. Instead, empower a dedicated leadership pod (Product, Engineering, AI) to own the end-to-end vision. This prevents features from being diluted into a 'lowest common denominator' by committee.
AI is fundamentally changing SaaS interaction. Instead of users clicking buttons to take action, AI will perform the tasks. The UI will then transform into a surface where users primarily review AI-driven outcomes, get insights, and make corrections, often interacting via conversational language.
When scaling a product internationally, identify universal components (like an invoice template) and build them as a common service. Then, use a well-designed API layer to connect to region-specific logic, such as different payment rails or tax rules, avoiding fragmented, country-specific products.
As companies deploy numerous task-specific AI agents (e.g., payroll, payments), the user experience risks fragmentation. Xero's solution is a 'super agent' that manages all sub-agents, orchestrating actions, transferring information, and applying user preferences globally to create a cohesive system.
When launching a new product, err on the side of a higher price. This strategy provides the flexibility to reduce prices later if needed—a much easier maneuver than attempting significant price increases on an established user base. As one advisor noted, 'it doesn't take a genius to reduce prices.'