March 2026 Summaries

Fin has introduced a new AI model called Apex, touted as the most advanced, efficient, and cost-effective technology in the customer service agent category, surpassing leading models like GPT-5.4 and Opus 4.5. Fin's latest development is a result of extensive research by their AI group, leveraging proprietary data to create a custom model that significantly enhances customer issue resolution rates, speed, and cost-effectiveness, with one notable client improving resolution rates from 68% to 75%. This advancement signals a shift in the AI industry towards more specialized and domain-specific models, moving away from general-purpose models to post-training as a competitive edge. The customer service sector, alongside coding and the legal industry, is experiencing substantial impacts from generative AI, highlighting the importance of companies developing their own models to maintain differentiation. The broader AI industry is witnessing a transformation with open-weight models closing the gap with frontier labs, suggesting a need for labs to innovate and create specialized models to stay competitive, potentially leading to new data partnerships or acquisitions. Fin's achievement with Apex exemplifies the potential of custom models in meeting specific industry needs and sets the stage for future advancements in deploying AI across customer service operations.

Fin Labs Paris has introduced Monitors, a new product designed to enhance the observability of AI-driven customer support operations, complementing existing tools like Insights and Recommendations. Monitors allow businesses to specify which conversations are reviewed and evaluated using Custom Scorecards, thus ensuring support quality aligns with business priorities. This system replaces traditional, less effective QA methods with a scalable solution that provides structured, repeatable evaluations. It includes a Review Queue for flagged conversations, enabling teams to implement improvements directly within Intercom. The comprehensive observability suite enables support leaders to confidently evaluate AI and human interactions to maintain high service standards. By integrating QA data with performance metrics, businesses can continuously monitor and optimize support quality, identifying areas needing improvement and ensuring conversations meet predefined criteria. Monitors aim to provide transparency and control in AI operations, facilitating perfect customer experiences through informed evaluations and real-time alerts for potential issues.

Kaizen, the Japanese philosophy of continuous improvement, emphasizes the power of small, incremental changes to achieve significant long-term results, a concept that has been effectively adopted by companies like Toyota. Intercom applies this philosophy to its customer service operations through the "Fin Flywheel," a process that enhances AI-powered customer support by continuously training, testing, deploying, and analyzing the AI system. This approach ensures that Fin's performance improves with each interaction, while human support teams at Intercom are encouraged to identify and implement improvements based on customer interactions, turning them into proactive problem-solvers. By embedding continuous improvement into both AI and human processes, Intercom creates a dynamic system where learning and refinement are integral, ultimately leading to a competitive advantage.

Leading the support function for a company with an AI-driven customer service platform presents unique, exciting, and daunting challenges, as the team uses the same advanced technology offered to customers, allowing them to be both the user and advocate for customer experience improvements. Since Intercom’s strategic pivot in 2022 to focus on an AI-first framework, their AI tool, Fin, has significantly transformed customer support by resolving over 81% of support volume, enabling the company to handle a 300% increase in demand without proportional headcount growth, saving an estimated $7.5M–$9M annually. This transformation involved restructuring the support team, adopting early AI features, optimizing knowledge management, and introducing roles such as Conversation Designer to ensure a seamless customer journey, which collectively shifted the team from reactive support to a consultative role that contributes to customer retention and expansion. Intercom's experience with Fin highlights the potential of AI in redefining customer service, emphasizing continuous improvement through the Fin Flywheel framework, which focuses on training, testing, deploying, and analyzing AI performance, while also preparing support staff for more complex, consultative roles. The case study underscores that successful AI integration requires not just technological adoption but also a comprehensive reevaluation of support processes, knowledge management, and team structures to fully harness AI’s transformative potential.

In 2023, the pricing model for AI agent Fin evolved from being based solely on resolutions, where the AI fully resolves a customer's issue, to a more comprehensive outcome-based approach. This shift acknowledges that the value of AI lies not only in complete automation but also in effectively handling complex tasks that may require collaboration with human agents. As Fin's capabilities have expanded to tackle more intricate queries, the definition of success has broadened to include procedures where Fin gathers information and performs actions before handing off to human agents. This new pricing strategy allows for more nuanced use of Fin, empowering teams to design workflows that meet compliance needs and incorporate human oversight when necessary. The transition to outcome-based pricing reflects the company's commitment to aligning pricing with value, maintaining fairness, and ensuring competitiveness, as Fin continues to develop into a comprehensive "Customer Agent" capable of managing the entire customer experience.

Recent updates to Fin have addressed key challenges in AI-powered customer service by removing constraints related to complexity, voice quality, Shopify integration, and helpdesk operations. Enhancements include the ability to handle complex queries using natural language, advanced simulations for testing, and improved voice features such as pronunciation rules, new voice options, and noise reduction capabilities. The streamlined Shopify setup allows for rapid deployment, enabling Fin to serve as a comprehensive customer agent that assists in sales and support. Additionally, Fin's native integration with Intercom helpdesk has been bolstered by new call metrics and automated updates for holiday office hours. These advancements aim to enhance Fin's functionality and user experience, providing teams with greater control and transparency in their customer interactions.

AI is transforming the customer experience by extending beyond traditional support roles, as businesses recognize the success of AI in support and plan to scale its use across other departments. The 2026 Customer Service Transformation Report reveals that over half of the businesses surveyed intend to expand AI applications, with customer support teams playing a strategic role in this transition due to their existing expertise in managing complex customer interactions. A notable example is WHOOP, a fitness wearables company that used AI to enhance sales interactions, resulting in a significant increase in sales. However, the expansion of AI across different departments poses a risk of fragmenting customer experiences if not managed cohesively, highlighting the need for a unified approach that keeps the customer at the center. This shift presents a unique opportunity for support teams to lead AI integration efforts, setting the standard for shared practices and ensuring seamless customer journeys. As customer service leaders take on expanded roles in AI implementation, the vision of a unified Customer Agent capable of managing the entire customer journey is becoming a reality, promising consistent and context-aware interactions.