How do you build a 100-day AI value creation plan for PE portfolio companies?

Run it in three blocks. In days 1 to 30, audit manual work in each portfolio company and score it on a four-level maturity ladder. In days 31 to 60, put one costed workflow per company live with human approval. In days 61 to 100, promote what runs clean and report the dollar value to the investment committee.

A 100-day AI value creation plan covers

• Readiness: an AI readiness assessment for each portfolio company, scored the same way across the fund.

• Cost: hours of repetitive work converted into a yearly dollar figure per team.

• Pilots: one agentic AI workflow per company, live with a named approver.

• Proof: a value report the investment committee can compare company to company.

Evidence: Grant Thornton is an audit, tax and advisory firm that publishes annual surveys of business leaders, including private equity leaders. Its 2026 AI Impact Survey found "80 % are currently exploring or piloting agentic AI", while "Only 5% have fully integrated AI into operations, compared to 14% across all industries."

Key points

• Most PE portfolios are stuck between pilot and production, so the first 100 days should turn pilots into costed, owned workflows.

• Score every portfolio company with one scale so the operating team can compare a software company and a services business on the same sheet.

• Start each pilot at L2, where AI prepares the work and a person approves it, then promote with evidence.

• Report value in dollars of recovered time per workflow, not in number of AI projects launched.

Why do PE portfolios need a 100-day AI plan now?

Because most PE portfolios have AI pilots without AI results, and the firms closing that gap run AI as a repeatable program. A 100-day window matches how operating teams already plan post-close work.

• Pilots without production: the same Grant Thornton survey says "On AI maturity, 45% are still piloting, 11% above the cross-industry rate."

• No incident plan: it also found "only 7% of PE respondents have a tested AI incident response playbook."

• Low expectations: Bain & Company is a management consultancy that publishes research on private equity for investors and operating teams. StepStone Group is a private markets investment firm that co-published the 2026 GP Outlook with Bain. Their March 2026 release reports "nearly 40% of GPs not expecting material financial impact from AI in 2026".

Firms that run AI as a portfolio program report scale. Vista Equity Partners is a private equity firm that invests in enterprise software companies. On a March 2026 Bain podcast, founder Robert F. Smith said, "We have over 30 that have gone through our agentic factory."

What did a teardown of 13 public PE AI programs show?

They converge on three moves: assess AI maturity company by company, build a shared playbook, and track value in money. In October 2026 I opened 13 public pages: 4 PE firm AI pages or interviews, 3 PE job postings for AI value creation roles, 4 advisory and search firm reports, and 2 surveys.

1. A central team that works inside companies. Hg is a European private equity investor that backs software and services businesses. Its AI transformation page says "Hg Catalyst is our in-house AI product incubator" and reports 1,600+ AI projects live across the portfolio.

2. Maturity assessed in diligence and after close. LLR Partners is a lower middle market private equity firm that invests in growth companies. Its Director of AI Strategy & Value Creation posting asks the hire to "Assess AI maturity, data assets, and competitive positioning during diligence" and to "Build repeatable playbooks, tools, and partner networks to scale AI adoption."

3. AI framed as an operating change. Blackstone is an alternative asset manager whose portfolio spans 250+ companies. Rodney Zemmel, head of its Operating Team, says "It's critical to approach AI adoption as a business transformation exercise rather than only a technology project."

Most funds do not have a central AI team of that size. The plan below runs the same three moves with one audit method, one scoring scale and one pilot per company.

What is the Manual Work Leak Audit?

The Manual Work Leak Audit is the first 30 days of the plan, run once per portfolio company. "A manual-work leak audit is a one-to-two-week review that lists every repetitive task a team does by hand, times each one, and converts the hours into a yearly dollar cost." The full method is on the Manual Work Leak Audit page. Before close, use the AI due diligence checklist for PE operating partners to price what this audit will find.

The six steps, run in each portfolio company:

1. Interview each person. Thirty minutes, recorded, with everyone in the focus area.

2. Map one row per workflow. Owner, trigger, steps, tools, minutes per run, runs per week and failure mode.

3. Time three live runs. Time three live runs of each candidate workflow with its owner, who signs off the baseline minutes.

4. Sort every step. Each step goes into one of four buckets: delete, plain code, agent or human decision.

5. Cost each workflow. Runs per week times minutes, divided by 60, times the loaded hourly rate, times 52.

6. Rank by annual dollars. Mark the top 3 and the first build. The output is a costed workflow map with signed baselines.

Illustrative worked example. A 12-person team spends 8 hours each a week on manual work at a $60 loaded hourly rate. That is $5,760 a week and $299,520 a year. If one workflow accounts for half of it, getting it back is worth $149,760 a year.

The dollar figure turns "we should use AI in finance" into a ranked list the board can approve.

How do you score each portfolio company's AI readiness?

Score each workflow on the AI Automation Maturity Ladder, then score the company on the readiness checklist below. "It is a four-level scale that describes how much of a workflow AI does and how much a person still does." The level names come from Relevance AI's four levels of AI autonomy. Relevance AI is an AI agent platform that lets business teams build and run groups of agents.

The AI Automation Maturity Ladder page has the full promotion rules: move one workflow up one level at a time, and only with evidence.

Portfolio scoring checklist (score 1 point per yes, per company)

• [ ] Leak audit done: the top 3 workflows have a yearly dollar cost.

• [ ] Runbook written: each of the 3 has a trigger, steps, owner, tool, done condition and exception on paper.

• [ ] Data reachable: the systems each workflow touches have API or export access.

• [ ] Owner named: one executive owns each workflow's outcome.

• [ ] Approver named: one person approves AI output at L2, with a backup.

• [ ] Baseline signed off: the owner has timed three live runs and signed off the baseline minutes.

• [ ] Eval cases approved: 5 to 10 cases from real past runs, graded by the workflow owner.

• [ ] Handoff contract drafted: the agent's limits and escalation rules are on one page.

• [ ] Incident plan: someone knows how to pause the agent and roll back its output.

• [ ] Value tracked: recovered hours and dollars report monthly to the operating team.

With 8 to 10 points, a company is ready to pilot in days 31 to 60. With 4 to 7, it needs SOP and data work first. Below 4, run the leak audit and park the pilot.

What happens in days 31 to 60?

Each ready company puts its top workflow live at L2, where AI prepares the work and a person approves it. Good first picks repeat weekly and follow written rules: lead routing, CRM data hygiene, collections follow-up, ticket triage, month-end reporting.

1. Write the runbook. Trigger, steps, owner, tool, done condition and exception, plus a log of the judgment calls a person makes.

2. Write the eval cases. Take 5 to 10 real past runs and have the owner grade them. No build starts without approved eval cases.

3. Sign an Agent Handoff Contract. The Agent Handoff Contract is a one-page document that sets out what the agent may decide alone, what sends a case to a human, who resolves it and how to reverse bad output.

4. Go live at L2 Copilot. Ship only when the eval cases pass. The agent does the work, the owner approves each output and every run is logged.

Most portfolio companies stall here: nobody inside has time to write runbooks and build the agent. That is the work Forward Deployed Engineering does: it starts with an audit and AI roadmap, documents SOPs as the knowledge layer, then builds agents one at a time, each with a handoff pack (SOP, run log, rollback path and owner). Each agent goes live at L2 Copilot, and the company owns the workspace, agents, docs and run log. Book a free coffee chat to get yours scoped.

What happens in days 61 to 100?

Promote what runs clean, demote what does not, and write the portfolio playbook. A workflow moves from L2 to L3 only after three clean reviewed runs.

• Promote with evidence. Workflows that pass go to L3, where a person reviews only flagged exceptions.

• Demote on failure. If a promoted workflow fails its checks, drop it one level and fix the SOP.

• Copy the winners. A collections or lead-routing workflow that worked in one company becomes a template for the next.

• Report to the investment committee. One line per workflow: baseline, hours recovered, yearly dollar value, current level, owner.

The report uses the leak audit's math, so day 100 numbers compare directly with day 30. Tie each line to the thesis lever: revenue growth, margin or working capital.

How does agentic AI fit portfolio companies with different maturity?

Set each company's day-100 target from its score. A services business with paper-based workflows may spend 100 days getting SOPs written and one workflow to L2, while a software company with clean CRM data may reach L3 on two workflows.

• Shared tooling, local owners. The operating team picks the approved tools and contract template; each company names its own approvers.

• Revenue workflows go through GTM Engineering. For pipeline, routing and CRM work, the GTM Engineering service starts with an audit and a costed workflow map, then a 30-day Sprint with up to 3 workflows live in the CRM. The forward deployed engineer for GTM guide explains how the two roles split the work.

What are the common mistakes?

Each one has a fix that fits inside the 100 days.

1. Counting projects, not dollars. The board sees a list of pilots and no margin change. Fix: report recovered hours and yearly dollars per workflow.

2. Skipping the SOP. The agent guesses at rules nobody wrote down. Fix: document the process before building.

3. Going straight to autopilot. Errors reach customers before anyone sees them. Fix: launch at L2 and promote after three clean reviewed runs.

4. One playbook for every company. A services business and a software company get the same target. Fix: score each on the ladder and set its own day-100 level.

5. No rollback plan. The agent writes bad data and nobody can undo it. Fix: put the rollback condition in the handoff contract before launch.

FAQ

What is an AI value creation plan in private equity? It is a time-bound program that finds where AI can cut cost or grow revenue in each portfolio company, puts the best workflows live with owners and measures the dollar value. A 100-day plan audits, pilots and promotes.

What is an AI readiness assessment for a PE portfolio? It scores each portfolio company on the same checklist: costed workflows, written SOPs, reachable data, named owners and approvers, a baseline, a handoff contract, an incident plan and value tracking. Each workflow is also placed on the four-level AI Automation Maturity Ladder.

How long does a manual-work leak audit take per company? One to two weeks for one team. The audit lists every repetitive task done by hand, times each one and converts the hours into a yearly dollar cost, so operating partners get a ranked shortlist of the top 3 workflows.

Should portfolio companies start with AI agents or AI assistants? Start one level lower than the ambition. Most workflows begin at L1, where people use AI as a tool, and the first pilot goes live at L2, where AI prepares the output and a person approves it. A workflow earns L3 after three clean reviewed runs.

Who owns AI value creation, the PE firm or the portfolio company? Both. The operating team owns the scale, the contract template, the approved tools and the reporting. Each portfolio company names the executive who owns each workflow's outcome and the person who approves AI output.

How do you report AI value to the investment committee? Use one line per workflow: the baseline, hours recovered, yearly dollar value, current ladder level and owner. Because the leak audit and the day-100 report use the same math, the numbers compare directly and tie to the thesis lever.

Plan AI value creation across your portfolio

Run the Manual Work Leak Audit in one portfolio company, score it on the AI Automation Maturity Ladder and use the checklist above to pick the first pilot. Kevin Payne is an AI engineer and GTM operator who installs AI workflows for B2B SaaS revenue teams. His three services (GTM Engineering, Forward Deployed Engineering and Content Engineering) are on the services page.

Want to test this plan on one portfolio company before rolling it out across the fund? Book a free coffee chat to talk through which company to start with as the pilot.

About the author: Kevin Payne is an AI engineer and GTM operator who installs AI workflows for B2B teams. He has 10+ years in B2B SaaS demand generation and RevOps across Techstars, a16z and YC portfolio companies, and his writing has appeared in HubSpot, GetResponse, Outbrain and more.