Announcing “Looking for Leverage”: A Newsletter for PE-Backed Leaders
Guest post: Where can we get gains from AI internally?
Get revenue recognition right, and no one notices.
Get it wrong, and it’s your name on the 10K.
As a CFO, I know how high the stakes are. Rev rec isn’t just an accounting issue—it’s a risk management issue. We created this free report to highlight the top revenue recognition challenges CFOs are facing in 2025—and what the right solution should look like to stay compliant, efficient, and out of the headlines.
Ok, I couldn’t wait any longer.
Mostly Metrics is branching out.
(Now that I’m full time, what’d ya think I’d do all day? ✍️)
Here’s a sneak peak of our NEW newsletter: Looking For Leverage 👀
It’s written specifically for execs working at PE backed companies.
I’m talking the CEOs, CFOs, CROs, CMOs, CTOs, and CPOs at portfolio companies.
Why write this?
There are many (too many?) newsletters about sexy, VC-backed, hyper growth companies…
But there are very few resources for the C-Suite operating more mature companies in environments that call for constraints around margins, cash flow, and time horizons.
This is the playbook I wish I had.
We’re opening the list to the first 1,000 subscribers today to give them a taste 👇
Note: It’s a free newsletter, and Mostly metrics readers will NOT be added automatically.
You have to opt in to this 🔥 content.
AI in PE Portfolio Ops: Headcount Cuts vs. Efficiency Gains
Many folks are obsessing over how to wedge AI into their product roadmap to drive revenue.
Not enough are talking about how to embed AI internally to save time and money.
Given the maturity of their operations, private equity-backed companies are in a prime position to lead here. But first, you need a realistic understanding of what AI can do today, and then translate that into specific, tactical use cases with real vendors, not just vague platitudes.
Let’s break it down.
AI in PE Portfolio Ops: Headcount Cuts vs. Efficiency Gains
PE-backed companies are using AI in two very different ways across internal ops:
Headcount Reducer – Replacing repetitive, structured work (think: high-volume, low-variance tasks).
Efficiency Booster – Augmenting existing teams so juniors can perform like mids, and mids like seniors (think: steroids)
Knowing which bucket a tool falls into helps you model costs, reassign headcount, and avoid overpromising.
1. AI is Cutting Headcount in Highly Repetitive Roles
These are the clearest direct cost-saving use cases where AI is outright replacing jobs:
Sales & Marketing: Fewer SDRs, More AI-Driven Prospecting
Vendors: 11x, Clay, Ema, Unify
Tasks: Lead sourcing, personalized outreach, follow-ups
Impact: Shrinks SDR teams by 20–50%, depending on outbound strategy
Limits: Can’t talk on the phone (yet)
Think of AI SDRs as scouts. They don’t close the deal, but they’ll tell you which alley to send your AEs down before they waste valuable time.
You can also afford to stretch the field, and test going after a wider array of companies that might not have been in your core ICP, since the outbound motion is now nearly infinitely scalable.
Customer Support: AI-Handled Low-Level Requests
Vendors: Decagon, Intercom, Forethought
Tasks: FAQs, password resets, invoice requests
Impact: Reduces level one support by 20–30%, and crucially might eliminate the need for overnight shifts
Limits: Needs strong help docs and structured inputs. You need to feed AI the playbooks you train your employees on, and teach it what “done” looks like.
AI isn’t killing the chatbot—it’s redeeming it (except Bank of America, that experience still sucks).
Finance: Automating Billing & Revenue Recognition
Vendors: Metronome, Tabs, RightRev
Tasks: Contract review, billing, receivables, revenue recognition, and reporting
Impact: 10–20% headcount reduction in AR for high-volume orgs
Limits: Bad inputs = bad outputs. Still need humans for edge cases and messy contract data.
Fun fact: Most self-inflicted revenue leakage happens on renewal, not on net new sales.
A 98% accurate AR tool is great... until you mess up a bunch of auto renews and have $3 million in ARR leakage (happened to a CFO I know).
That’s why you need to laydown real tracks for an agent to work on.
2. AI is Supercharging Efficiency, Not Cutting Jobs
Now for the upside stories—the ones where AI acts as a performance steroid.
These are cases where AI helps existing teams do their jobs 20-100% faster, making junior employees perform at mid-to-senior levels.
Engineering: AI is a Force Multiplier, Not a Job Killer
Vendors: GitHub Copilot, Devin, Replit, Cursor, Bolt
Tasks: Write code based on prompts
Impact: Increased output and code coverage
Junior engineers write 30-50% more code per sprint.
Code coverage (test automation) increases from 50% → 80%, reducing QA cycles.
Dev teams ship features 25-40% faster, accelerating time to revenue.
Limits: AI won’t replace architects. It can build, but not think through system design.
Remember: Engineers have reused code for decades. Most code is filled with large chunks of open source (why recreate Google Maps or a ‘Contact Us’ form from scratch?).
AI is great at finding building blocks to leverage, and then also stitching together + testing those building blocks.
Product Development: AI Speeds Up Prototyping
Vendors: Replit, Bolt, v0, Lovable
Tasks: Turn designs into working code
Efficiency Gains:
Cuts handoff time 2–3x
Frees PMs to spend 50% more time with customers
Limits: Just because you can get to a prototype faster, doesn’t mean it was actually, like, a good idea.
No, Figma isn’t going anywhere. But the back-and-forth between design and build just got way shorter.
Marketing Attribution & SEO: Smarter, Not Cheaper
Vendors: HockeyStack, Growth X
Tasks: Multi-touch attribution, programmatic content
Efficiency Gains:
30–50% boost in attribution clarity
2–5x faster content scale
Limits: AI still needs human judgment on brand voice and strategy
Want to test 50 subject lines before lunch? Now you can. But someone still has to know what “good” looks like. That’s what Don is for.
The Big Picture: AI Changes the Shape of Work, Not Just the Cost
Where AI replaces roles: Support, sales, and finance (high-structure, high-volume tasks)
Where AI boosts speed: Engineering, product, GTM (creative or technical work with leverage)
PE-backed companies should be thinking less about “how much can we cut?” and more about “how much faster can we scale?”
And by the way - humans are terrible at gauging exponential change.
We overestimate what we can get done in a day and wildly underestimate what can happen over ten years. AI is no different.
A lot of us are expecting too much, too fast, from tools that are still learning to walk. But that doesn’t mean we should wait. The smart move is to run controlled experiments now - to figure out what 5%, 10%, even 20% better looks like.
Those gains may feel incremental at first, but they compound quickly. So best to start today. Pick one use case from the above and let it rip.