How Building with AI Can Double the Throughput of Your Engineering Team
AI Engineer Europe 2026 · 21:49
AI customer agents and customer service software
Intercom, renamed Fin in 2026, builds AI customer agents for service, sales and ecommerce. Its Intercom helpdesk remains a separate product, giving support teams shared inbox tools, centralized ticketing, multilingual support and proactive messaging for onboarding and customer guidance. Fin’s service agent handles customer queries across chat, email, WhatsApp, SMS, phone and Slack; its sales and ecommerce offerings extend the agent into inbound sales and shopping.
Founded in 2011 by Eoghan McCabe, Des Traynor, Ciaran Lee and David Barrett, the company remains led by CEO and chairman Eoghan McCabe. Its technical approach includes Apex, a proprietary model purpose-built for customer support, and retrieval-augmented generation to ground answers in knowledge-base content. Its engineering team developed Fin-cx-reranker using ModernBERT-large to reorder retrieved passages by relevance before answer generation, tailoring retrieval to English customer support.
As of August 2026, the company reported more than 30,000 companies using its products and over $400 million in annual recurring revenue. Salesforce signed an agreement in June 2026 to acquire Fin for approximately $3.6 billion, subject to purchase-price adjustments. The acquisition remained pending as of August 26, 2026, when Salesforce revised its expected closing to the coming weeks, during its third fiscal quarter of 2027.
AI Engineer Europe 2026 · 21:49
Affiliations reflect their AIE appearances, not necessarily current employment.
Start here to understand Scanlan’s use of code changes per R&D employee as a throughput measure and the role of coding-assistant experimentation.
Brian ScanlanAI Engineer Europe 2026
Start here for Bar’s approach to choosing high-value support use cases and scoping an MVP around knowledge-base answers, authentication, and routing.
Peter BarAI Engineer World's Fair 2025
Scanlan describes reusable agent skills and hooks alongside Claude Code-related pull-request automation, grounding coding-agent adoption in concrete development practices.
Bar compares chained speech-to-text, LLM, and text-to-speech architectures with direct voice-to-voice models. His talk also covers call observability and integrations with existing support-team workflows.
Affiliations reflect each recorded session, not necessarily current employment.