002 AI for Operators

In this episode, Emily walks us through her step-by-step process for using ChatGPT to make herself more effective at project management, including the following:

• 30-minute charters. Emily trained ChatGPT on her template; one prompt + clarifying questions = a polished charter in half an hour.

• Detailed project plans. Feed the signed charter back and ChatGPT spits an Excel timeline (tasks, dependencies, risks); even a complex data-governance rollout took <60 min.

• Clarifying-question hack. ChatGPT asks what’s missing, surfacing scope gaps and hidden blockers.

• Hallucination guardrails. She tells ChatGPT to label unknowns “TBD,” so creative fills are flagged and facts stay clean.

• Memory pruning = speed. Deleting one-off trivia, she keeps reusable patterns; each new charter gets faster.

What It Does

Generate well-designed presentations with AI.

Why Ops Leaders Should Care

If you work at a deck-first company, this could be a lifesaver. While it’s probably not quite there yet for the most high-stakes external use cases (board meetings, investor presentations, etc.), it’s a great pick for All Hands presentations, department meetings, project status updates, or other internal use cases where speed and legibility are more important than tweaking the final 1% of the design. 

Likewise, if your team handles a high volume of less customized customer presentations (initial sales meetings, customer success QBRs, etc.), this could be a useful way to shift their focus from perfecting presentations to driving business outcomes.

Key Features (Pros & Cons)

Pros:

• Fast

• Rich, interactive content, including videos, embeds, forms, etc. - and can be consumed as a live presentation, a scrollable webpage, or exported into pdf, ppt, or google docs

• Collaborative: multiple people can work on a deck together in real time

Cons:

• Themes and formatting can seem a bit “samey” after a while - “Gamma starts off looking impressive, but quickly becomes underwhelming after a few hours of use. The AI-generated slides are extremely repetitive” (G2 review)

• Content quality can be lackluster for complex subjects - Reddit commenter said that it was “not consulting ready at all” (albeit this was 7 months ago)

• Users cannot freely reposition every element on the slide; there are some limits to the customization

• No offline mode - if the airplane wifi isn’t working, you won’t be able to work on your deck

Live Experience

In addition to speaking to users of Gamma, reviewing online commentary about Gamma, and reviewing product features vs. competitors, I tried my hand at a few different use cases.

Investor update to presentation: 

• I pasted in one of my recent investor updates - Gamma had decent judgement on vs. how to separate the content into individual slides

• The AI-generated images were not great - I would not use them

• The AI-generated graphics were okay, but really only served to display the content in a different format rather than adding a layer of meaning as great graphics can

• Each of the slides had a different format (under the same theme); the formats were somewhat visually engaging

Blog post to presentation: 

• I took a blog post that I wrote and asked Gamma to create a presentation

• The design was clean and readable

• The AI-generated graphics were quite relevant in places but it missed obvious opportunities to amplify the text in the graphics (e.g. one of the key concepts was that of a ‘barbell’ and it did not carry that over to the graphics creation)

• The text was reflected accurately and the slides had good titles

Sales proposal email to presentation: 

• I took an email to a potential customer outlining different options and asked Gamma to create a presentation based on it

• This was the best of the three presentations despite having the least text to work with (potentially because the concepts were quite simple, comparatively?)

• The formatting was effective and made the different packages easy to understand

• The AI-generated headlines were fairly strong and the copy was straightforward

Other Options

Beautiful.ai, Decktopus, Plus.ai, Canva, Prezi

Bottom Line

Out of the box, Gamma will create slick-looking presentations quickly, which can be useful for a number of use cases. However, for complex, conceptual, or client-ready presentations, you’ll still need to put a fair amount of work into tweaking the text, graphics, and layout to get to a good end-result (for now, at least).

• 16 Changes to AI in the Enterprise: Andreessen Horowitz’s latest CIO survey says LLM budgets will jump 75% next year, mapping 16 shifts every ops chief should bake into 2025 plans.

• Voice AI is Working - Here’s Where it Wins First:  As operators, many of us still think text-first, but underestimating the utility of voice is a mistake.

• OpenAI’s Malicious-Use Report: A board-ready rundown of covert influence ops, model-poisoning, and scam rings you’ll need to block before any customer-facing AI rollout.

• Generative AI Legal Checklist (Cleary Gottlieb): Copyright, indemnity, and disclosure landmines distilled into a punch-list for execs moving pilots from sandbox to production.

• Don’t build this: Get brutal feedback from AI on your startup idea

• o3 Pro Can Replace 10-15% of Human Hours Worked Today: (According to o3 Pro) 

• The Gentle Singularity: New post from Sam Altman lays out the next few years of AI impact.

• Human-AI Relationships: OpenAI’s Joanne Jang explains why users bond with chatbots (and what OpenAI is doing about it).

• Claude Code Best Practices: A leaked cheat-sheet shows how Anthropic teams script Claude like a colleague.

Thanks for reading,

Tom Guthrie

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