Town's Strategy: Neutral AI Agents for Work
Town founder Jean-Denis Grès explains the pivot from tax software to a horizontal AI assistant. The strategy focuses on model neutrality, data ownership, and product-led growth to outpace vertical incumbents and model labs.
Strategic Pivot: From Vertical Tax to Horizontal AI
Town’s trajectory illustrates a critical lesson in AI entrepreneurship: vertical specificity does not guarantee venture-scale growth if the workflow lacks high-frequency interaction. The company initially built an AI tax preparation tool, a vertical with clear technical feasibility but poor go-to-market dynamics. Tax preparation is an annual, low-frequency event with limited word-of-mouth potential, making it difficult to achieve the rapid network effects required to outpace incumbents with established distribution. Recognizing this structural limitation, the founders pivoted to a horizontal AI assistant focused on high-frequency knowledge worker workflows, specifically email and calendar management. This shift allowed Town to tap into a massive, recurring TAM where AI can provide daily value, creating a stronger foundation for product-led growth.
The Neutrality Advantage
A core differentiator for Town is its stance on model neutrality. Unlike model labs that may lock customers into their own infrastructure, Town positions itself as a neutral layer that routes requests to the best-performing model for a specific task. This strategy addresses a key pain point for enterprises: the uncertainty of which model provider will dominate over the next three to five years. By remaining agnostic, Town reduces vendor risk and allows businesses to optimize for cost and performance dynamically. Furthermore, Town emphasizes data ownership, ensuring that the context and memory generated by the AI remain portable assets for the user. This contrasts with the 'Forward Deployed Engineer' (FDE) model used by some labs, which often creates lock-in by building custom, non-portable software.
Product-Led Growth and the Tinkerer Gap
Town’s go-to-market strategy prioritizes product-led growth (PLG) over sales-led enterprise deals. The thesis is that consumer-grade experiences drive adoption; if the tool is not genuinely useful to the end-user, enterprise compliance features cannot save it. The company targets SMBs and mid-market firms first, where adoption friction is lower, to build a flywheel of feedback and feature development. A significant challenge identified is the 'tinkerer gap': most businesses lack the technical expertise to discover AI’s full potential. Town addresses this by proactively suggesting automations based on aggregate user behavior, effectively removing the need for users to be 'tinkerers.' This approach democratizes AI adoption, allowing non-technical users to capture high-ROI efficiencies, such as automated price monitoring or meeting prep, without manual configuration.
Conclusion
Town’s strategy highlights the shift from app-centric software to agent-centric workflows. By focusing on horizontal utility, model neutrality, and user-centric design, Town aims to become the default interface for AI-driven business operations. The company’s success will depend on its ability to scale this neutral, portable model while maintaining a product experience that feels indispensable to daily work. As AI capabilities improve, the value will shift from the model itself to the orchestration layer that connects intelligence to specific business contexts.
Key insights
-
Vertical AI applications with low-frequency interactions, such as annual tax preparation, struggle to achieve venture-scale growth due to limited network effects and word-of-mouth potential. High-frequency horizontal workflows offer a more robust path to market penetration.
Impact: Founders should evaluate workflow frequency and recurrence when selecting AI verticals to ensure sustainable growth and competitive advantage over incumbents.
-
Model neutrality allows AI application companies to avoid supplier lock-in and optimize for the best-performing model for each task. This reduces vendor risk and enhances cost efficiency for enterprise customers.
Impact: Neutrality becomes a key differentiator against model labs, attracting enterprises that fear being locked into a single provider’s technology stack.
-
Consumer-grade product experiences are a prerequisite for enterprise adoption. Building compliance features before ensuring user value leads to poor adoption and product failure.
Impact: Prioritizing user experience drives organic demand and makes enterprise sales easier, as IT departments are more likely to approve tools that employees actively use.
-
The 'tinkerer gap' limits AI diffusion because most users lack the technical skills to discover and implement AI automations. Proactive suggestion engines based on aggregate behavior can bridge this gap.
Impact: Automating the discovery of AI use cases expands the total addressable market to non-technical users, significantly increasing product penetration and value realization.
-
Data ownership and portability are critical trust factors for enterprise AI adoption. Users must feel secure that their business context and memory are not being used for training or locked into a specific platform.
Impact: Emphasizing data portability differentiates neutral AI platforms from labs, reducing switching costs and building long-term customer loyalty.
Action items
-
Audit current AI workflows for frequency and recurrence. Pivot away from low-frequency verticals if they lack strong network effects or word-of-mouth potential.
Impact: Ensures the business model supports venture-scale growth by focusing on high-impact, recurring user interactions.
-
Implement a model-agnostic architecture that allows dynamic routing to the best-performing AI model for specific tasks. Market this neutrality as a key risk-mitigation feature.
Impact: Attracts enterprise customers concerned about vendor lock-in and optimizes operational costs by leveraging the most efficient models.
-
Prioritize consumer-grade user experience in product development. Delay heavy enterprise compliance features until the core product demonstrates strong organic user adoption.
Impact: Drives bottom-up adoption and reduces friction in enterprise sales cycles by ensuring the tool is genuinely valued by end-users.
-
Develop proactive suggestion engines that analyze aggregate user behavior to recommend high-ROI automations. Remove the need for users to manually discover AI capabilities.
Impact: Bridges the 'tinkerer gap,' expanding the user base to non-technical employees and increasing the overall value delivered by the platform.
-
Establish clear data ownership policies that guarantee user portability of context and memory. Ensure no training on user data without explicit consent.
Impact: Builds trust with enterprise customers and differentiates the platform from model labs, reducing switching costs and enhancing long-term retention.
Quotes
“We're neutral. We will use whatever supplier we think is the best for you in the moment.”
“It is risky always to be competing with your suppliers.”
“We're trying to build a product that works with people to find more and more opportunities for them to deploy AI to be better at their job.”