#373 What Do Your Colleagues Do All Day? (The Value of Institutional Knowledge & AI for Process Reengineering) | Jennifer Smith, CEO at Scribe
Summary
Jennifer Smith, CEO of Scribe, argues that enterprise AI fails because models lack context on how specific companies operate. She advocates using AI to automatically map workflows, turning ephemeral tribal knowledge into machine-readable "specialized intelligence." This approach helps identify inefficiencies and best practices, enabling targeted automation rather than broad, ineffective agent deployment.
Topics discussed
Intro: AI skills evolution and the myth of job replacement
Guest intro: Jennifer from Scribe on process automation
The gap between general AI intelligence and specific business context
Treating general AI as rented utility vs. owned specialized IP
The challenge of documenting ephemeral, head-based processes
Automating process discovery to avoid manual documentation
Using LLMs to parse raw data and map workflows automatically
Case study: Reducing non-selling time for sales reps
Leveraging workflow variation to find and spread best practices
Quick wins: Improving morale by fixing tedious process steps
Defining clear business goals before starting AI transformation
Tracking digital work in physical industries like agriculture
Why individual 'n-of-1' roles still have repeatable processes
Avoiding RPA failures by understanding real work vs. happy paths
Managing and sharing AI skills and agents across teams
Calculating ROI: Projections, accuracy, and tool adoption
Automating time tracking to remove cognitive overhead from employees
The mindset of accountability and ownership in an AI-native world
Human agency: Defining goals, values, and constraints for AI
Conclusion: Compounding specialized intelligence over time
Listen ad-free on Castria