Drawing from his experience at Square, Gokul advises that in a multi-product portfolio, some products should be optimized for profit while others are for retention. Understanding this distinction is crucial for setting the right team goals and building a sticky ecosystem.
Gokul presents a framework of eight moats—data, workflow, regulatory, distribution, ecosystem, network, physical, and scale—to evaluate a software company's durability. He argues that companies with a score of four or more are highly defensible against threats like AI.
Gokul argues that single-function tools in a vertical have a limited ceiling. To achieve a decacorn valuation, vertical SaaS companies must aim to own the entire software stack for their niche, like ServiceTitan with its 30+ products for field services.
Gokul argues that brand is no longer a strong moat for B2B companies. As AI makes data portability and product replication easier, he predicts switching costs will approach zero, making business customers more rational and less loyal to brands.
For a system of record like Salesforce to survive the threat of AI agents built on top of them, they must actively commoditize their complement. This means identifying their core profit pool (data vs. workflows) and aggressively building and offering the other for free to neutralize new entrants.
A "bolt-on" AI strategy will fail. Successful integration isn't about adding an AI feature; it's about fundamentally re-evaluating and rebuilding the entire product experience and its economics around new AI capabilities, creating entirely new user interactions.
Market sizing fails to predict the biggest hits because they often create "non-consumption markets." Companies like Shopify succeed not by capturing existing spend, but by creating a product so remarkable that it convinces users to pay for a new category of tool they never previously budgeted for.
Gokul is a huge fan of the trend toward very young founders, noting he's invested in more dropouts recently than in the past 15 years. He believes they are "AI maxing"—natively adopting AI tools to live and breathe differently, giving them an operational edge.
The biggest AI opportunities lie in replacing human labor costs, not just competing for existing software budgets. Gokul observes this shift happening in stages: companies first cut outsourced BPO spend, then freeze hiring for roles that leave, and only later resort to layoffs.
Judging an early-stage company on its current gross margins is a mistake. The key indicator of future profitability is its potential pricing power. A defensible, sticky product that can consistently raise prices over time is a much stronger signal than one that relies solely on falling costs.
At the seed stage, if you're right about a truly exceptional company, the entry valuation hardly matters. Gokul cites a 200x return on an expensive seed deal. However, by Series B, a high price can crush your multiple, even if the company continues to perform well.
Gokul has reversed his stance on remote work for startups. He now argues that being fully remote kills early-stage companies because it drastically slows down iteration speed and hinders crucial founder alignment. He advises being in-person at least three days a week.
AI is splitting software into two categories: "access products" and "work products." While access tools can stick with seat-based pricing, work products (e.g., AI that processes legal contracts) must adopt outcome-based pricing, as value is tied to output, not the number of users.