The Decisions That Make Or Break Startups

Go-to-Market for AI in Legacy Industries

When bringing AI into a legacy industry like accounting or law, founders have three primary go-to-market strategies:

1. Build AI Software for Incumbents: This is the most common approach. The goal is to identify a high-value, specific problem within the industry that can be solved with AI in the first few months. The key challenge is finding the right initial customers—the true early adopters who are empowered and incentivized to integrate new technology. Founders should heavily pre-qualify leads to find these champions, as they are crucial for initial feedback and traction. A successful tactic is to embed with a potential customer, as one YC company did by working out of a law firm’s office for months to build their MVP.

2. Start a “Full-Stack” AI-Powered Firm: This involves creating a new company that performs the industry service (e.g., an accounting firm) with the intention of automating operations from the ground up.

   • Core Metric: The most important metric is the percentage of work that is automated. This metric’s trajectory is more important to investors than early revenue.
   • Common Failure Mode: A significant risk is scaling the manual side of the business too quickly. Founders might achieve initial success and hire dozens of practitioners (e.g., accountants), effectively becoming a traditional service firm with some software. This dilutes the technology focus and makes it harder to prioritize automation. A forcing function, like capping non-technical hires, can enforce discipline.

3. Acquire an Existing Firm: This strategy provides an immediate customer base but presents the immense challenge of changing an established culture, which is often intractable. This is the least common approach.

Growth vs. Defensibility: Enterprise vs. Mid-Market

When deciding between targeting large enterprises or the mid-market, the most important factor for an early-stage startup is the pace of learning.

• Enterprise Drawbacks: Enterprise sales cycles are notoriously long, often lasting many months. This slow feedback loop starves a startup of the rapid learning it needs to iterate and find product-market fit.
• Mid-Market Advantage: Mid-market companies and smaller businesses generally have faster decision-making processes, allowing startups to learn, ship, and iterate more quickly.
• The Exception: If the problem you’re solving only exists at the enterprise level, you have no choice but to start there. In this case, the advice is to target the smallest possible company that experiences the pain or dramatically reduce the product’s scope to land a small team within the enterprise.

Ultimately, qualifying the individual you are selling to is often more critical than the company segment. An empowered and motivated buyer in a mid-size company can move much faster than a bureaucratic enterprise.

Pivoting, Conviction, and Technical Difficulty

When to Pivot, Even with Traction

Pivoting is one of the most difficult decisions for a founder, especially when the current product has some traction (e.g., hundreds of thousands in ARR). The key question to ask is not “Is the revenue number going up?” but “Are our users truly valuing the product?”

Often, founders with lukewarm traction discover that no two customers use or describe the product the same way, and none would be devastated if it disappeared. This indicates a lack of a core, high-value proposition.

The ultimate leading indicator for a pivot is the founders’ own loss of conviction. If you stop believing your current path will lead to a massive company, it’s time to explore new directions. A pivot is a reset that requires immense energy, so it’s crucial to explore a range of new ideas to find one that restores that deep-seated conviction.

The Value of Technically Difficult Problems

Founders should not pivot away from an idea because it is technically difficult. On the contrary, technical difficulty is a powerful moat. If a problem is hard to solve, few others will attempt it, giving a skilled team a significant advantage.

To tackle a project with a long build time (e.g., 6+ months), the key is to reduce the scope:

• Build an internal version first: Build the simplest possible tool for yourselves to use. This allows you to become your own first user and engage with customers in a consulting capacity while you build the full product.
• Use existing APIs: Before building complex backend infrastructure (like a real-time bidding platform), build a best-in-class frontend on top of another company’s API. This gets you to market faster, helps you acquire users, and provides the learnings needed to build your own system later.

A critical warning: Do not use technical complexity as an excuse to avoid talking to customers. Even if the product isn’t ready, you can learn immense amounts by living their problems alongside them.

Hiring and Open Source Strategy

When to Hire

For an early-stage startup, the answer to “Should we hire?” is almost always no. Hiring too early slows you down. The right time to hire is when you are at a breaking point because the business is working so well that you don’t even have time to conduct interviews.

• Leading Indicator: A specific part of the company—sales, engineering, support—is consistently failing or about to fail because of overwhelming demand.
• Hiring is not a success metric: The focus should be on capital efficiency. Many top founders now aim to reach significant revenue milestones with the smallest possible team.
• The Exception: Opportunistic hires. If a truly superlative person (e.g., your smartest friend, the best engineer you know) becomes available, it can be right to bring them on.

When to Open Source an Enterprise SaaS Product

While common for developer tools, open-sourcing can be a powerful strategy for any enterprise SaaS product, particularly in markets with high-security, privacy, or compliance concerns.

The primary benefit is trust. By making the code available, you dramatically shorten sales cycles. Customers in sensitive industries like healthcare can verify the code and have the option to self-host, eliminating concerns about sharing private data with a third-party startup. Even if they never read the code or self-host, the knowledge that they can is often enough to establish the trust needed to close a deal. However, self-hosting comes at a cost and must be priced as a premium feature.

• Tags:
• Open Source
• Go-to-Market
• Feature
• Enterprise AI
• Hiring
• Startup Strategy
• Pivoting