I work in PE and have been digging into how AI can drive operational improvements in portfolio companies, especially in business services. Early numbers suggest 3-4% EBITDA margin improvement is achievable through automation and efficiency gains in some cases.
Wanted to get a pulse on what others believe:
- Is this actually exciting or just hype?
- What are your expectations for AI in portcos?
- Have you tried implementing anything yet?
- Anyone have wins (or horror stories) to share?
For context, I'm looking at business services companies where labor is 50%+ of revenue - insurance brokers, staffing, etc. Seems like low-hanging fruit but interested in other sectors too.
Happy to share what I'm finding. Anyone down to compare notes?
Focus on the back office first. Finance, legal, compliance all have real verticalized solutions that can actually cut down a lot of time and resources on non-revenue generating activity. That's applicable across your entire portfolio, then you can go into specific issues like calculating broker commissions, streamlining compliance for staffing agencies, or processing small government RFPs.
My firm does this across our portfolio currently and we are in the process of measuring value uplift at our portcos but initial results are positive. The main challenge is the change management in the portco once our deployment ends as they have to use the new technology/tool we build for them to good effect and this can take time to train talent (or hire externally).
See the following thread https://www.wallstreetoasis.com/forum/private-equity/value-creation-in-pe-leveraging-data-science-and-machine-learning
Hey, can I DM you?
I would love to chat briefly and learn more from your experiences.
Very grateful for anything you can share.
Just talk to the AI it's clear you're outsourcing all your thinking to it anyway.
Mods, kill it with fire
I think we have skipped a critical step. In my workplace, building internal tools through code is a lot more reliable and accessible and extremely high-leverage, whereas AI is harder to do well (training an internal model) and higher risk due to hallucinations.
I was thinking more of purchasing available AI solutions for your vertical, as opposed to taking on the risk of doing internal development.
E.g., in your case, it would mean using AI tools even for software development such as Cursor or Lovable.
I was thinking more of purchasing available AI solutions for your vertical, as opposed to taking on the risk of doing internal development.
E.g., in your case, it would mean using AI tools even for software development such as Cursor or Lovable.
Disclosure: I work for Planr (AI-native value creation platform), so take accordingly. But wanted to share some things we're actually seeing work.
@Determined is right that back office is the obvious starting point. But honestly, the bigger wins we're seeing aren't on the cost-cutting side - they're on revenue growth, which most firms aren't even looking at through an AI lens yet.
A few things that have actually moved the needle:
1. Knowing about the missed quarter 6 weeks early, not at the board meeting - by the time you're sitting in that room hearing they missed, it's 3x harder to fix. All the warning signs were in the data - pipeline slippage, conversion drops, deal push rates creeping up - but nobody could see them because the data's siloed across sales, finance, ops, and 14 different spreadsheets. When you break down those silos and look fund-wide, the patterns scream at you.
2. Hidden growth blockers - pushed deals, conversion drop-offs, pipeline issues. Most portco systems literally can't surface this because the data is scattered everywhere. When you aggregate fund-wide, you start seeing things like "why are 3 of our 7 portcos seeing the same conversion cliff at stage 2?" That's a portfolio-level insight you'd never get company by company.
3.A/B/C player identification across KPIs - not just revenue, but GTM metrics, operational efficiency, HR data. See who your real performers are across the fund using the same methodology. Then you can actually act on it - redeploy playbooks, move talent, double down on what's working.
4. Natural language to find the "why" - dashboards are great but the real unlock is asking "why did pipeline velocity drop in Q3 at portco X" and getting an actual answer with contributing factors, not a 3-week project from the ops team.
The change management point above is real though. What's worked is giving portco management teams their own view - single source of truth across finance, HR, RevOps - so it's not just a monitoring tool for the sponsor.
When the CFO and CRO are living in it daily, adoption sticks.
Happy to compare notes if anyone's testing stuff. This space is moving fast.
Soluta est culpa est ut numquam consequatur ipsum. Debitis expedita voluptates dicta cum aut quaerat. Omnis aut id praesentium ad. Omnis sint voluptatem similique sed et libero perferendis. Qui voluptatem voluptas modi aut eum. Qui delectus iusto excepturi.
Eveniet sit vero quia harum aut. In error non voluptas esse sed sed. Nisi voluptas omnis ad et porro quia. Enim omnis aut ea cumque aut et iusto culpa. Aut sequi eos hic sed est. Et est est et provident non quis impedit. Quia deserunt porro labore explicabo voluptatem.
Dolor enim et et omnis doloribus libero. Ullam ut a magnam esse dolorum.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...