003 AI for Operators

Head of Biz Ops at AOA Dx, review of Lex, 7 links

Author

Tom Guthrie
June 18, 2025

Hi there, 

Thanks for joining AI for Operators!

Here’s what we’ve got for you this week:

  • The Operator: Emma Martin, Chief of Staff at AOA Dx

  • The Review: Lex - an AI-first, browser-based word processor

  • The Links: 7 articles, including an AI-only marketing department, a McKinsey report, and where Finance leaders are implementing AI

The Operator

Emma Martin, Chief of Staff at AOA Dx

Emma is a 2x Chief of Staff, currently a leader at AoA Dx, a company pursuing a revolutionary breakthrough in early-stage cancer detection for women.

In our third episode of the AI for Operators podcast, Emma breaks down her learnings, operating in a highly-regulated space, including:

  • Become a design partner, not just a buyer. AOA-DX piloted the tool mentioned above as a design partner; in return they got a built-for-them solution before the rest of the market. Getting early access and shaping the tool can create a strategic advantage.

  • Prioritize ROI demonstration. Four months of usage has shown substantial time savings and fewer compliance errors. These are the types of results that help justify more AI budget.

  • Price and privacy are the gating factors. Startups need enterprise-level data isolation without enterprise-level pricing. Solving for both is difficult (but not impossible).

  • Make AI a reflex across every function. The 12-month vision includes tools for R&D experiment design, clinical study protocols, and off-site logistics so every employee asks the model first.

  • Governance before enthusiasm. Roll out guardrails, mandatory training, and usage audits early; proactive expectation setting is key.

  • Build the daily habit. Emma now consults ChatGPT before Google for everything from ops workflows to dinner recipes, showing that constant reps are needed to embed AI thinking into company culture.

The Review

In this section, we’ll try to bring you quick, useful review of tools that you may not have encountered (or highlight use cases that may be new to you). If we ever do sponsored reviews, we’ll be very clear that they're sponsored. In the case of Lex, Nathan (the founder) and I have spoken a couple of times before, but he and his team didn’t influence this review in any way (and I have no financial interests in Lex).

What It Does

Lex is a browser-based word processor that integrates AI into the writing process, enabling you to draft, research, rewrite, version, and polish long-form pieces in one place. Real-time collaboration, version history, and saved prompts keep teams aligned without the clutter of traditional tools. Persona-driven prompts, style guides, and knowledge bases add brand voice and automation to every document.

Why Ops Leaders Should Care

The operator’s key tools are the spreadsheet, the deck, and the doc. Lex’s vision - to create more powerful writers by seamlessly integrating AI - means that it could be faster and easier for you to create better writing, from internal docs like board updates and all-company emails to external docs like marketing collateral and long-form essays. If your company has a writing-heavy culture, then it’s worth seeing if you could save time (and create more) with a tool like Lex.

Key Features (Pros & Cons)

Pros

  • Single workspace for drafting, AI assistance, and editing (easier than flipping between tools)

  • “Rewind” timeline and manual save points make version control simpler

  • Style Guides and Knowledge Bases let you lock in brand voice and surface internal references

  • Prompt Library and personas speed up brainstorming, outlining, and rewrites

  • Inline chat with multiple models means you can switch from GPT-4o to Claude 3 mid-draft

  • Generous free tier plus affordable per-seat pricing for small teams.

Cons

  • One big shared space: lacks true team-level workspaces or opt-in visibility, which can get noisy

  • No native SEO, internal-link suggestions, or media search—external tools still needed for web publishing polish

  • Limited integrations with CMS, Notion, Airtable, or image libraries, so final hand-offs stay manual

  • Mobile app is still in beta, and there’s no desktop offline mode.

Live Experience

I tried several different use cases with Lex to get a feel for the product and had other members of my team try it out. Their perspectives have been incorporated throughout the review.

Using Lex’s chat feature to critique my essay on communities

First, I used it to edit an essay that I wrote a few weeks ago but haven’t published yet. It was able to analyze the document and identify places where my argument was weak and provide options to strengthen it. This was somewhat helpful, but some of the suggestions didn’t make sense (most likely a model problem, not a Lex problem).

Second, I uploaded a novel I’m writing and asked it to find logical inconsistencies and plot holes. It didn’t find any. I don’t think I trust this assessment. I asked it to find typos. It found some and hallucinated others.

Finally, I used it to help write some sections of this newsletter. I’ll let you be the judge of how effective it was in doing that!

Other Options

Lex isn’t the only tool out that uses AI to help writers. I put together an overview below of some of the more popular ones, along with who they’re probably best for, based on their features and price point.

It will be particularly interesting to see whether the smaller, nimbler startups wholly focused on these use cases will end up winning against the much larger incumbents like Google Docs and Notion where most of your writing probably already lives.

Bottom Line

Lex has some interesting features and can be quite powerful, if you take the time to configure it with the right prompts and context. If you write for a living, or people on your team do and you’re not a huge enterprise, it’s worth trying out.

The Links

  • Meta snaps up 49% of Scale for $14.3B: The strategic rationale here isn’t crystal clear, but here’s what is: Meta needs to catch up in the race for AI supremacy, this is their bet (for now), the biggest companies in the world are willing to spend almost unlimited amounts of money to compete in this race, and they’re also willing to turn to increasingly convoluted deal structures to avoid antitrust scrutiny.

  • McKinsey’s new “agentic AI” playbook: The consulting giant makes the case for centralizing AI efforts and the indispensable role that CEOs play in making sure that AI initiatives are prioritized and effective. Bullish for Chiefs of Staff and operators who work closely with CEOs. But (see Ethan Mollick’s video below) are they thinking too small?

  • Where 200 finance leaders are deploying AI today: While data analytics and forecasting are the most common use cases today, many organizations are using AI to catch anomalies and even close the books. Focus on the stats and ignore the vendor sales pitch - this can be helpful intel for how to bring AI to your finance department.

  • How ChatGPT is changing education: Students are usually among the first to figure out novel ways to cut corners and, unsurprisingly, it seems like all of them are using AI. Worth reading for a couple of reasons: 1) these students are your future employees, 2) an increasingly corporatized educational system has more in common with corporate America than most people assume.

  • AI-powered roll-ups are the hot new VC thesis: People will make money here, but I’m guessing that one of the things they’ll discover is that the operators deploying AI are just as important (if not more so) as the AI systems that these roll-up founders are leading with. M&A integration is not a fun or easy thing.

  • [Video] Wharton’s Ethan Mollick argues the real AI win isn’t cost savings but 10× creativity: Too many companies are only thinking about AI in terms of headcount reductions or efficiency gains. What would happen if you tried to do more instead of less? Even if you can’t spend all your time thinking about that question, it’s probably worth spending more of it.

  • “Our marketing team is just me and ~40 AI agents”: A former Google PM Lead and current Founder/CEO shares how he built a marketing team that’s all AI.

What if you could wake up tomorrow and actually own your company’s AI strategy?

That’s what the Chief of AI Fellowship is built for. Over six weeks, you’ll:

  • Demystify AI - pick the right tools for the right jobs

  • Build real automations that save hours every week

  • Craft an AI roadmap that your CEO can’t ignore

  • Brainstorm and problem-solve with a cohort of top Chiefs of Staff, BizOps, and Strategy leaders

Cohort #3 kicks off July 24. Seats are going fast and the waitlist is in the 100s.
Ready to stop playing catch-up and start setting the pace?

Thanks for reading,

Tom Guthrie

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