← back
Anthropic in the Enterprise — Alexander Bricken & Joe Bayley
Original: Anthropic in the Enterprise — Alexander Bricken & Joe Bayley
Takeaway
Get back to the core user problem, invest in evals and architecture, and use interpretability-aware models for trust-critical enterprise workflows.
Summary
- Anthropic's interpretability team works in four stages (understand, detect, steer, explain) — papers 'Towards Monosemanticity' and 'Scaling Monosemanticity' isolate features like 'famous NBA players' #304 and gave us Golden Gate Claude.
- Customer guidance: move beyond chatbots/summarization to product-defining bets — e.g. hyper-personalize course content rather than ship Q&A.
- Sonnet 3.5 (Oct 2024) leads SWE-bench; Claude is available via API, Claude for Work, AWS Bedrock and GCP Vertex.
- Common enterprise pitfalls include skipping evals, weak architecture design and treating Claude as a drop-in for legacy summarizers.
- Anthropic's applied AI team built MCP and 'Building Effective Agents' research, and embeds with customers to design architectures and prompts.
anthropicenterpriseinterpretability
Original description
Enterprise AI implementation often fails due to overengineering, poor data infrastructure, or lack of testing - despite pitfalls many industry leaders across finance, customer support, and productivity sectors have achieved transformative results with Claude. This session will dissect Anthropic's Applied AI approach across multiple deployment models (API, cloud partnerships, and enterprise solutions), revealing how to build effective AI roadmaps while avoiding these common pitfalls. Drawing from real-world case studies of frontier model deployment, we'll demonstrate practical strategies for cross-functional collaboration, developing robust evaluation frameworks, and making informed build-vs-buy decisions that align with your organization's AI maturity and safety requirements. Recorded live at the Leadership Track Session Day from the AI Engineer Summit 2025 in New York. Learn more at https://ai.engineer and purchase tickets to our next event, the AI Engineer World's Fair, in SF June 3 - 5 here: https://ti.to/software-3/ai-engineer-worlds-fair-2025 About Alexander Alexander Bricken is a member of the Applied AI Product Engineering Team at Anthropic, leading technical implementation for Financial Services. He specializes in building production-grade Claude solutions, working directly with customers while collaborating with the Product Research team. Before Anthropic, he spent 2 years at Palantir, where he balanced his responsibilities as an AI Engineer and Account Strategist, specifically working with North American Insurance customers. He has a background in machine learning, data science, and mathematics as one of the founding students of Minerva University.