Why Your Bots Fail and How Agents Fix Your Customer Support
Audio Brief
Show transcript
This episode discusses Intercom's AI agent Fin, focusing on its strategy, impact, and disruption of the customer service industry.
Here are three key takeaways from the discussion. First, Intercom prioritizes building a single, standardized AI product that universally improves for all customers, contrasting with custom platforms or bespoke services. Second, their AI agent, Fin, performs complex, multi-step tasks, with "resolution rate" serving as the North Star metric for success. Third, Intercom employs outcome-based pricing, charging per successful resolution, which demonstrates confidence and actively disrupts traditional customer support models like BPOs.
Intercom's "product over service" philosophy ensures that improvements to Fin's core model benefit all users simultaneously. This approach centralizes advanced AI optimization, delivering superior quality without customers needing to engage in complex individual prompting or custom development. It fosters a flywheel effect where continuous enhancement scales efficiently across their client base.
Fin's agentic capabilities allow it to execute sophisticated business procedures, such as processing refunds or returns, by integrating with external APIs. This goes far beyond simple Q&A, enabling genuine automation of complex workflows. Its success is measured by the "resolution rate," representing the percentage of conversations Fin autonomously resolves without human intervention.
Intercom's unique outcome-based pricing, billing per successful resolution, directly aligns their financial incentives with customer value. This model reflects strong confidence in Fin's performance, which is often proven through A/B testing against competitors. The agent's efficiency frequently leads enterprises to end contracts with Business Process Outsourcers for tier one support, highlighting its disruptive market impact.
This discussion highlights how a product-first AI strategy, focused on outcome-driven metrics and pricing, can fundamentally transform industries like customer service.
Episode Overview
- Fergal Reid, Chief AI Officer at Intercom, discusses the strategy and impact of their AI agent, Fin, a tool designed to autonomously resolve customer service inquiries.
- The conversation explores Intercom's philosophy of building a single, horizontal AI "product" that improves for all customers, contrasting it with customizable platforms or service-based solutions.
- Key themes include the disruption of the traditional customer service industry, the importance of "resolution rate" as the North Star metric, and the agent's ability to perform complex, multi-step tasks.
- The discussion covers Intercom's unique outcome-based pricing model, where they charge per successful resolution, aligning their success directly with their customers' value.
Key Concepts
- Product over Service Philosophy: Intercom focuses on building a standardized, highly opinionated AI product (Fin) rather than a generic platform or a custom engineering service. This approach creates a flywheel effect where improvements to the core model benefit all customers simultaneously.
- Agentic Capabilities: Fin is designed as an "agent" that can perform complex, multi-step business procedures, such as processing refunds or returns by integrating with external APIs, going beyond simple Q&A.
- Resolution Rate as the Core Metric: The primary measure of Fin's success is its "resolution rate"—the percentage of conversations it resolves without human intervention. This metric serves as the North Star for product development and performance benchmarking.
- Disruption of BPOs: The deployment of Fin often leads customers to end contracts with Business Process Outsourcers (BPOs) for tier-1 support, showcasing its disruptive impact on the customer service industry.
- Self-Serve with Enterprise Support: While Fin is designed for easy, self-serve onboarding, Intercom has developed professional services to help large enterprises navigate the complex business transformation and integration required for full adoption.
- Outcome-Based Pricing: Intercom utilizes a unique pricing model where they charge per successful resolution. This aligns their financial incentives directly with the customer's success and demonstrates strong confidence in the product's performance.
- Performance-Driven Competition: Intercom encourages prospective customers to A/B test Fin against competitors, confident that its superior resolution rate will be empirically proven.
Quotes
- At 0:11 - "We don't really want our customers, you know, writing individual prompts and trying to get down into the machine learning..." - Fergal Reid explains that Intercom handles the complex AI optimization centrally to deliver higher quality.
- At 9:15 - "the typical pattern is, a customer of ours will deploy it, and they will often end a contract they had with a business process outsourcer..." - Reid describes how Fin is disrupting the customer support industry by replacing outsourced human agents.
- At 10:28 - "Our core metric that we care the most about, we call resolution rate, and it's essentially the percentage of times when Fin is involved in a conversation that it successfully resolves the conversation..." - Reid defines Intercom's North Star metric for its AI agent.
- At 17:52 - "Please just do an AB test, do a trial and and see how you get on." - Reid explains their straightforward advice to potential customers, encouraging them to test Fin's performance directly against competitors.
- At 28:05 - "We bill a dollar per resolution... If we resolve the question, we get a dollar. If the end user is not happy and they talk to the human, we get no dollar." - Reid explains their outcome-based pricing model, which ties their revenue directly to the AI agent's success.
Takeaways
- A product-first strategy for AI, focusing on a single, standardized core that improves for everyone, can create a more powerful and scalable solution than bespoke services.
- The true value of AI in customer service lies in its ability to be an "agent" that executes complex business processes, not just a bot that answers questions.
- Tying your business model directly to customer outcomes, such as charging per successful resolution, is a powerful way to build trust and demonstrate product confidence.
- Defining a single, clear North Star metric like "resolution rate" is critical for focusing product development and proving value in a competitive market.
