Jay and Jeff dig into why Harvey went from SaaS to AI implementation, why agent infrastructure has to be a company-wide investment, and why the personal-agent wave is a platform war, not an app war. Then they get honest about AI writing: 78% of readers stop reading when they suspect AI wrote it.
KEY TAKEAWAYS
- Paid AI is still niche: A16Z data shows only ~3% of US consumers pay for AI themselves, but the youngest buyers adopt at 4x the rate of the oldest.
- SaaS to AI implementation: Harvey went from app to harness to its own models, and now deploys that know-how inside clients as forward deployed engineering.
- Centralize agent infrastructure: When every department builds its own agents, context fragments. CS, support, product and engineering should share one customer brain.
- No plumbing, no agents: Messy CRMs and unused data warehouses mean you often have to build the infrastructure before agentic work pays off.
- Platform war, not app war: Most personal agents look the same. You'll buy into ecosystems (Meta, Amazon, Google, Apple) based on where your data lives.
- Don't use the words it suggests: LLMs write in magazine headlines and encourage your first-draft impulses. Readers sense you've become "artificially flavored."
- Readers punish AI writing: 78% stop reading suspected AI content and 71% avoid the author going forward.
- Use AI as researcher and editor: Let it gather sources, opposing views and data, then write yourself. Save generation for templated assets built from verified sources.
CHAPTERS
- 00:00 - Welcome and AI in high school
- 01:38 - Only 3% of consumers pay for AI
- 04:19 - Meta Muse negotiates on Facebook Marketplace
- 06:30 - Harvey's shift from SaaS to AI implementation
- 10:59 - Why agent infrastructure should be centralized
- 13:19 - You're never that far behind
- 14:11 - Selling agentic work without the plumbing
- 14:53 - Tool FOMO: Instinct, Town and Muse
- 17:27 - A platform war, not an app war
- 20:01 - Jay goes back to writing on his own
- 23:28 - Two rules for writing with LLMs
- 25:38 - Readers abandon AI-written content
- 28:03 - AI for research, not writing
- 29:48 - AI slop in client presentations
- 31:19 - Claude in Slack and trusting AI messages
- 32:18 - Branded two-pagers in seven minutes
About the Show: Chief Customer Officer Podcast is a show about real strategies for customer-led growth in the AI era—from leaders actually executing, not just talking about it.
Your Hosts:
- Jay Nathan – CEO of Balboa Solutions and Co-Founder of ChiefCustomerOfficer.io
- Jeff Breunsbach – Head of Customer Success at Junction and Co-Founder of ChiefCustomerOfficer.io