Tech Disruptors Salesforce on Measuring Agentic Work Outcomes
9 snips
Mar 25, 2026 Madhav Thathai, GM and EVP of Salesforce AI with prior leadership at Google Maps Enterprise and Dell, discusses agentic enterprise frameworks and how agents shift work measurement. He explains Agentic Work Units, hybrid reasoning that mixes LLM flexibility with deterministic workflows, session-trace tooling for optimization, and how monetization and budgeting evolve around agent consumption.
AI Snips
Chapters
Transcript
Episode notes
LLMs Enhance Rather Than Replace SaaS
- LLMs supercharge existing SaaS stacks rather than replace them overnight.
- Madhav Thathai says LLMs speed deployment, enable natural-language interfaces, and create new agentic outcomes across data and workflows.
Slackbot As A Personal Work Assistant
- Slackbot is an example of an employee agent that knows conversations, context, and connected systems to act as a personal assistant.
- Madhav says Slackbot can execute tasks across Salesforce, ServiceNow, Workday, and other apps in the flow of work.
Four Systems That Make An Agentic Enterprise
- Salesforce defines four systems for an agentic enterprise: context, work, agency, and engagement.
- The framework ties together data (context), encoded workflows (work), agent platform/tools (agency), and interfaces like Slack or WhatsApp (engagement).