The first 100 days after close set the operating rhythm for the whole hold. If AI isn't in that plan with owners and numbers, it tends to drift into scattered pilots that never reach the P&L. This guide lays out a phased AI plan for a newly acquired portfolio company: what to do in each 30-day block, how to pick initiatives, how to govern them, and which KPIs to track.
The short answer
- Days 0–30: diagnose and baseline. Days 31–60: pilot two or three initiatives. Days 61–100: scale what worked and stop what didn't.
- Pick initiatives with a simple weighted score for value, feasibility, and time to impact, and limit yourself to a handful.
- Run AI inside the broader 100-day plan, with the same governance and KPIs, not as a separate innovation track.
Why AI belongs in the main 100-day plan
Most portfolio companies are past the "should we try AI" stage, and few are past the pilot stage. Bain, which advises PE firms, reported that in a September 2024 survey of investors, most said their portfolio companies were testing generative AI, but nearly 20% had operationalized use cases with concrete results. Bain also describes Vista Equity Partners requiring quantified goals for generative AI in each company's annual operating plan.
That's the point of putting AI in the 100-day plan. The same plan already covers management changes, reporting, pricing, procurement, and add-on readiness. AI should compete for resources against those priorities, using the same yardstick: impact on EBITDA, cash, or growth within the hold.
- Days 0–30Diagnose and baseline
Inventory AI in use, score initiatives, and baseline the KPIs.
- Days 31–60Pilot
Run two or three pilots against the baseline.
- Days 61–100Scale or stop
Scale what moved a KPI. Stop what didn't.
Days 0–30: diagnose and baseline
Don't start by buying tools. Start by learning where the work and the money are.
- Pick up the diligence findings. Turn AI risk items, such as unvetted tools or unclear data rights, into week-one actions.
- Map the core workflows. In a field service business: booking, dispatch, job execution, invoicing, renewals. In an energy infrastructure manufacturer: RFQ, quoting, engineering, procurement, production, testing, dispatch.
- Measure the baselines. Pull 12 months of history for the KPIs you might move: technician utilization, first-time completion, time to invoice, RFQ-to-quote time, on-time delivery, days sales outstanding.
- Build a long list. Interview the people doing the work. Expect 20 to 40 ideas.
- Score and cut. Score each idea and pick two or three to pilot.
How to score initiatives
Rate each idea from 1 to 5 on three factors, weight them, and rank by the total.
| Factor | What a 5 looks like | Suggested weight |
|---|---|---|
| Value | Material, measurable impact on a P&L line or cash | 50% |
| Feasibility | Data exists, workflow is clear, an owner has time | 30% |
| Time to impact | Measurable results within 60 days of starting | 20% |
Weights are a starting point. Adjust them to your thesis, then hold them fixed so scores stay comparable.
Illustrative example: an AI-assisted dispatch tool scores value 4, feasibility 3, time to impact 3. Its weighted score is (4 × 0.5) + (3 × 0.3) + (3 × 0.2) = 2.0 + 0.9 + 0.6 = 3.5. An automated invoice-from-job-completion workflow scores 3, 5, 5: 1.5 + 1.5 + 1.0 = 4.0. The less glamorous idea ranks higher, which is often the right answer in month one.
Days 31–60: pilot
A pilot should test whether the initiative moves a KPI, not whether the technology works.
- Define success before you start. Write down the KPI, baseline, target, and measurement method. For example: cut average time from job completion to invoice from five days to one, measured across two branches.
- Keep scope tight. One branch, one product line, or one team. Enough volume to measure, small enough to fix fast.
- Put an operator in charge. The owner should be the branch manager or operations lead whose numbers will change, supported by technical people.
- Review weekly. Track adoption as well as outcome. A tool nobody uses has no value, however good it is.
- Set guardrails. Decide what data the tool can touch and where a person must review AI output before it reaches a customer.
BCG, which sells AI consulting, found in a 2024 survey of 1,000 executives that about 70% of AI challenges relate to people and process, 20% to technology, and 10% to algorithms. Plan your pilot effort in roughly that proportion: most of it on training, workflow changes, and follow-up.
Days 61–100: scale or stop
- Make a call on each pilot. Scale it, extend the pilot with a specific fix, or stop it. Stopping is a good outcome if it frees people for something better.
- Plan the rollout. Sequence branches or sites, budget training time, and name the owner at each location.
- Move KPIs into the monthly operating review. From here on, AI initiatives report like any other value-creation lever.
- Refill the pipeline. Re-score the long list with what you've learned and pick the next wave.
- Write the value case for the board. State the measured impact, the annualized run rate, and what it took.
Governance that doesn't slow you down
Keep it light and tied to existing structures.
- Sponsor: the CEO or COO, with the operating partner as the investor-side counterpart.
- Owner per initiative: one named operator accountable for the KPI.
- Cadence: weekly initiative check-ins, a monthly review inside the main 100-day steering meeting.
- Policy: a one-page acceptable-use policy covering approved tools, data that must never leave company systems, and when human review is required.
- Decision rights: who can approve a new AI tool, and the spending limit before it goes to the steering group.
KPIs to track
Track three layers for each initiative: adoption (are people using it?), operational KPI (is the workflow better?), and financial impact (what's it worth?). In field services, that might be share of jobs dispatched with AI suggestions, then jobs per technician per day, then gross margin. In energy infrastructure, share of RFQs processed with the tool, then RFQ-to-quote days, then win rate and revenue.
We help portfolio companies build and run these plans, so weigh our view accordingly. The template below is free and works without us. Use it with your portfolio company: score the long list, lay out the three phases with owners and due dates, and track each KPI from baseline to target.
Checklist for your 100-day AI plan
- Have the diligence AI findings been turned into week-one actions?
- Are baselines measured for every KPI you plan to move?
- Is every idea scored on the same weighted criteria?
- Are you piloting no more than two or three initiatives?
- Does each pilot have a named operator, a target, and a measurement method?
- Is there a one-page AI acceptable-use policy?
- Do AI initiatives report in the main 100-day steering meeting?
- Is there a scale, extend, or stop decision for each pilot by day 100?
Where to go from here
If you've just closed, or are about to, and want help shaping the plan, tell us about the company. You can also read how we support each stage from diligence to exit on our Private Equity page.