Lars Wiedenhoefer

About Lars Wiedenhoefer

I design AI-powered product systems that make revenue-critical digital journeys measurably more reliable.My focus is mapping behavioral and transactional telemetry to business outcomes: identifying where digital journeys fail, quantifying the revenue impact of friction, and leading remediation that strengthens performance and protects revenue.Over the last several years, I’ve built and operated full-stack observability and intelligence ecosystems connecting replay, product analytics, DevOps telemetry, CRM data, payments signals, and operational metrics. Increasingly, I design AI-assisted investigative systems that compress the path from signal → diagnosis → remediation.This includes:• Building agentic workflows inside Salesforce (Agentforce) that classify, route, and summarize cases using governed LLM patterns• Designing AI systems that reconcile replay, event streams, and operational data to surface recurring friction patterns• Prototyping Customer360-style intelligence portals (Node.js + Snowflake + LLM integration) to demonstrate unified journey insight• Establishing instrumentation governance so defects cannot hide in fragmented telemetryMy approach is economic and systemic:Every friction signal is traced to financial impact.Every AI system is built with oversight, structured outputs, and operational accountability.Every automation is evaluated by how much it reduces investigation latency and improves journey success.I thrive in complex environments where identity, onboarding, payments, and customer care intersect and where AI must operate within real governance, compliance, and performance constraints.

Experience

1. Technical Product Management - Customer Experience

   CapMetro

   Nov 2024 — Present · Austin, TX, US

   Architecting AI-assisted journey intelligence and automation across Salesforce (Service Cloud, Data Cloud, Agentforce), Transit App, and Umo mobility systems — unifying behavioral, operational, and CRM signals into governed decision workflows.• Designed and deployed multi-step agentic workflows inside Salesforce Agentforce, integrating LLM prompts, record collections, structured outputs, and orchestration logic (looping, conditional branching, variable passing).• Delivered a working Customer360 intelligence prototype in 2 days using AI-assisted development (Node.js + Snowflake), providing architectural validation and accelerating strategic decision-making on a $6M Salesforce investment — with zero additional spend.• Built AI-powered case classification, summarization, and routing systems that reduce manual triage and accelerate response time across customer care operations.• Established reusable architectural patterns for AI in CRM environments — defining prompt governance, output structuring standards, and human-in-the-loop oversight.• Led AI-centric development by delivering working prototypes engineers could extend, normalizing AI-assisted workflows across product and CRM initiatives.• Designed AI-powered workflows that surfaced customer recognition insights to bus and rail operators, extending CRM intelligence into frontline operations and enhancing real-world service delivery.• Expanded Salesforce Data Cloud + Marketing Cloud activation to enable behavior-driven segmentation and automated outreach journeys.• Influencing enterprise AI roadmap — defining priority use cases, data requirements, governance controls, and scalable orchestration patterns.Key Impact:• Launched Salesforce Agentforce in production (first public transit agency deployment).• Reduced manual case handling workload through AI-driven routing and structured summarization.• Accelerated automation prototyping through AI-assisted development and rapid iteration workflows.

Education

• RPTU Kaiserslautern-Landau

  Master's, Computer Science