A pattern observation from the last twelve months of portfolio diagnostic work: sponsors are increasingly asking what their portfolio companies should be doing about AI. The question is usually framed at the wrong altitude.
The question is not whether portfolio companies should invest in AI. The question is where in the hold period the investment makes sense, what it should produce, and how the operating layer should be positioned to absorb it.
Portfolio Value Architecture is Fulcrum’s structured operating method for sponsor-backed companies. It sequences three phases across the hold period. Frame for diagnostic and stabilization in the first 180 days. Floor for operating cadence installation through the mid-hold. Focus for pre-exit posture in the final 12 to 18 months. AI belongs in each phase. The work it should do in each phase is different.
What follows is the operating partner’s view on where AI fits, sequenced to the hold.
The Frame phase runs in the first 180 days of the hold period, typically months three through six. The work is diagnostic. The deliverable is a written briefing that names the operating signals observed, quantifies their compounding cost, and prioritizes the intervention sequence.
The questions that belong in the Frame phase are operating questions about the company’s current AI posture:
- Where is AI capital currently being deployed, and against which operating outcomes?
- Which initiatives have an accountable owner inside the company, and which sit with external vendors or consultants?
- What operating cadence exists for reviewing AI initiatives, and what decision rights are resolved?
- Which functions have the workflow documentation required to support AI deployment, and which do not?
- What data infrastructure exists, and what does its condition imply about realistic AI timelines?
The output of this diagnostic work is not a list of AI tools to consider. It is an honest assessment of whether the operating layer can absorb AI investment at all, and if so, where the readiness is highest.
“A sponsor who funds an AI initiative before this diagnostic has happened is funding a guess. A sponsor who funds an AI initiative after this diagnostic has happened is funding a sequenced operating move.”◆ Frame Phase · Portfolio Value Architecture
The Frame phase will, in many cases, surface that the operating preconditions for productive AI deployment are not yet in place. That finding is more valuable than any AI strategy the portfolio could produce in the same window.
The Floor phase runs through the mid-hold, typically months six through thirty-six depending on engagement scope. The work is operating cadence installation. The deliverable is an installed rhythm of decision-making, accountability, and execution discipline that holds the value creation plan in place.
AI gets installed in the Floor phase, but not the way most companies install it.
The mistake we observe most frequently in mid-hold companies is treating AI as a parallel workstream. Companies will run their normal operating cadence on one track and an AI initiative on a separate track. The two tracks rarely converge. The AI initiative produces outputs that the operating cadence does not absorb. The cadence runs forums and meetings that do not surface the AI initiative’s results. The capital deployed on AI does not produce operating leverage because the operating layer does not interact with it.
In practice, this means several things.
The regular operating reviews must include the AI initiatives in scope. Not as a separate agenda item, but as part of the function they are intended to affect. An AI initiative in finance operations belongs in the finance operating review. An AI initiative in sales belongs in the commercial review. If it does not belong in either, it does not belong in the portfolio.
The decision-rights structure must explicitly address AI deployment. Who can approve a pilot. Who can move a pilot to production. Who can shut one down. Who owns the outcome. Without these decisions resolved, AI initiatives stall.
The measurement infrastructure must tie AI outputs to operating metrics, not to AI metrics. Model accuracy, response time, and adoption rates are not operating outcomes. Cycle time reduction, margin improvement, and accuracy improvement are. The Floor phase work installs the discipline of measuring the right thing.
The Focus phase runs in the final 12 to 18 months before exit. The work is pre-exit posture. The deliverable is a company prepared to defend EBITDA quality, withstand buyer diligence, and avoid the operating signals that compress the multiple at the moment of sale.
AI posture affects exit in three specific ways. Sponsors approaching the Focus phase should treat each of them deliberately.
First, AI-driven cost takeout shows up in the QofE process. Buyers will scrutinize the durability of AI-driven margin expansion. Whether the cost reduction is structural and defensible, or whether it depends on continued vendor relationships, key-person knowledge, or unstable workflow assumptions. The Focus phase work is to harden these dependencies before the buyer asks.
Second, AI maturity is starting to influence multiples in certain sectors. In categories where buyers are paying premiums for AI-mature targets, the absence of AI capability is a discount factor. In categories where AI maturity is not yet priced, the presence of capability creates an asymmetric upside. The Focus phase work includes positioning the company’s AI posture in the management presentation accurately, with neither inflation nor understatement.
Third, key-person and vendor concentration risks tied to AI initiatives should be resolved before exit. Companies that built AI capability around a single internal champion or a single external vendor carry transition risk that diligence will surface. The Focus phase resolves these risks while there is still time.
The answer reveals whether the spend is on track or whether the company is running the advisory-without-authority pattern. It reveals whether the operating layer is positioned to absorb AI as a capability or whether AI is being run as a parallel initiative that will produce overhead instead of leverage. It reveals where the company sits in the readiness sequence for hold-period AI investment.
The companies that can answer this question crisply are positioned. The ones that cannot are exposed.
Portfolio Value Architecture is built to install the operating answer.