027 AI for Operators

Matthew Scott of Descript, New AI Operator Jobs, AI in the People Org?, 9 Links

Author

Tom Guthrie
December 10, 2025

Hi there, 

Welcome back to AI for Operators. Here’s what we’ve got for you this week:

  • The Operator: An interview with Matthew Scott, Product Marketing Manager at Descript (listen here)

  • The Essay: AI in the People Org?

  • [New Section] The Jobs: Senior AI roles with $150k-$275k+ salaries 

  • The Links: 9 links, including building AI peers, Cursor’s internal onboarding guide for non-technical hires, and lots from Anthropic.

The Operator

Matthew is a Product Marketing Manager at Descript. Previously, he was Group Strategy Director at Translation and Product Marketing Lead at Airbnb, where he worked from 2017 to 2022.

Matthew Scott, Product Marketing Manager at Descript, shares how his work building and marketing an AI product reinforces his personal AI stack, and vice versa

His Personal AI Stack 

  • Crafting a Context Engine: he uses Claude Code on top of an Obsidian vault, with multiple MCPs plugged in, preserving his data and decisions for compounding use.

  • Beating the Sprawl Problem: sub-agents prune stale notes and surface only relevant history, keeping context windows lean.

  • Increasing Quality: he uses editor agents to do multiple passes for brevity and accuracy. This has moved his output completeness from 50% done to consistently 90% done.

  • Gaining Leverage as a Manager: treat recordings and notes as a mini-CRM for reports to prep better 1:1s and track follow-through.

Building and Marketing an AI Product

  • Building an Agentic Co-Editor: Descript’s product lets non-editors edit by prompt, generate bespoke B-roll, and apply tasteful cuts without learning pro tooling.

  • AI GTM Demands a Faster Cadence: ship and talk daily, not yearly; presence and useful demos are more effective than big, long-horizon campaigns.

  • Consumerization of Enterprise: instead of assuming a technical audience, it’s more effective to aggregate many frontier models in one approachable workspace so even less savvy users can try their capabilities.

Broad takeaway: Different functions have different fundamental building blocks (product marketing: maximum context; design: maximum thoughtful critique). So, instead of designing cookie-cutter automations, break your role down to first principles and identify ways to amplify your unique capabilities.

The Essay: AI in the People Org?

One trend that I’ve been tracking recently is the growing proportion of AI roles and initiatives that are being slotted under the Chief People Officer/HR function.

In some ways, this makes sense: the success of AI initiatives hinges just as much (if not more) on adoption, training, and staffing than it does data pipelines, security, and internal products. If the People team has experience scaling (effective) training, and if the organization is set up to view AI tools both as a way to augment humans and replace certain roles, then this can work.

In many cases, though, there’s a distinct lack of strategic, commercial thinking in the People function. And the last thing that you want for your mission-critical AI deployment is for it to be in the hands of people who prioritize checking boxes off a list versus having a real impact on the P&L and the long-term strategic prospects of the organization.

I remain an advocate for these initiatives to sit with the Chief of Staff, Business Operations team, or a similar function that is nonpartisan (i.e. not in the service of a specific exec’s empire building), commercial, and capable of both being strategic and executing on more tactical initiatives. Someone in one of these roles will be forced - in a good way - to build allies and capabilities in every department, with the blessing and support of the CEO, COO, or CFO who often own the most important parts of the long-term strategic roadmap.

So if you’re a people ops person who owns AI initiatives, or if you’re working in an organization where HR is in charge, it’s not the end of the world - it can turn out well! But, you’ll have to emphasize commercial and strategic impact and make sure it’s not just in service of more bureaucracy.

The Jobs

We’re trying out a new section here where we highlight AI-first jobs. Like it? Want more? Let us know!

  • Own enterprise AI strategy, governance, and rollout across product, CX, and internal operations

  • Comp: $200k-$275k base

  • Lead rollout of AI tools across the credit card lifecycle (application, underwriting, fraud, marketing, servicing)

  • Comp: $150k-$175k

  • Own AI program strategy, the operating cadence, and cross-functional execution for the superintelligence organization

  • Comp: $137k-$222k

  • Scope AI use cases, run prototypes, and establish governance frameworks

  • Comp: $185k-$275k

The Links

Practical

Perspectives

News

Thanks for reading,

Tom Guthrie

If you liked this edition of AI for Operators, share it with another operator you know!