Two live deployments on one runtime answer diligence questions that a hundred slides of modelled projections cannot. The transfer test is whether the system runs without the founder in the room.
What a Home Services Private Equity Buyer Should Ask Before Trusting the Pitch
As of 2026, more than three billion dollars has been deployed into AI roll-ups targeting American service businesses. General Catalyst allocated 1.5 billion dollars to its Creation Strategy. Thrive Capital launched a dedicated vehicle and brought OpenAI in as an equity partner. Long Lake reached 100 million dollars in EBITDA in under two years and agreed to take American Express Global Business Travel private for 6.3 billion dollars. The capital is moving fast. The diligence, in too many cases, is not keeping pace. Every one of those players is capital-first: they buy the business, then scramble to build the AI. The question a home services PE buyer should be asking before signing an LOI is not whether a vendor has an AI story. It is whether the AI is already running the work, or whether it is still a model on a slide.
The Conviction-Execution Gap Is the Real Risk in Every Home Services Deal
The gap between what PE funds believe about AI and what is actually running in their portfolio companies is the single largest unpriced risk in home services M&A right now. According to FTI Consulting's 2026 Private Equity AI Radar, which surveyed 200 senior decision-makers across three continents, 95 percent of funds report their AI initiatives meeting or exceeding expectations. In the same survey, only 36 percent of portfolio companies have AI deployed in production, and just 7 percent describe it as fully integrated across the enterprise. Forty-three percent have not implemented AI in any meaningful way.
That gap is not a technology problem. It is a diligence problem. Funds are underwriting the pitch, not the production system. When a vendor walks into a pre-LOI meeting with a home services target and presents modelled projections of dispatch efficiency, churn reduction, and collection improvement, the right question is not whether the numbers look attractive. The right question is: where is this running today, and can it run without the founder in the room?
Why modelled projections are not the same as verified mechanisms
A modelled projection starts with an assumption about what the system will do once deployed. A verified mechanism starts with a system that is already doing it. The difference matters at exit. When a buyer underwrites a twelve-times EBITDA multiple on a home services platform, every dollar of projected AI-driven margin improvement that turns out to be a model rather than a mechanism is a dollar that will not survive the hold period. The EBITDA does not compound. The multiple does not hold.
The specific failure mode looks like this: a vendor shows a 20 percent reduction in windshield time, a 15 percent improvement in first-call resolution, and a projected drop in days-sales-outstanding from 47 to 19. Each number is plausible. None of them is attached to a running system. The diligence team accepts the model because the math is internally consistent. Post-close, the integration takes 14 months, the founder who held the tribal knowledge leaves at month six, and the system never reaches the state the model assumed.
The Transfer Test: Whether the System Runs Without the Founder
The transfer test is the most important question in any AI vendor diligence process, and it is almost never asked directly. It is not a metaphor. It is a specific operational check: if the founder of the target business, or the lead engineer of the AI vendor, were removed from the picture on a Tuesday morning when a technician misses a job, a customer disputes an invoice, and a compliance deadline triggers, does the system handle all three without a human escalation?
Most systems fail this test not because the technology is weak but because the system was built around a person rather than a process. The dispatch logic lives in the dispatcher's head. The dunning sequence depends on someone manually reviewing the aging report. The compliance calendar is a spreadsheet that one operations manager updates. When the founder exits at close, or the key employee takes a competing offer, the system reverts to a manual operation wearing an AI costume.
What passing the transfer test actually looks like in a home services fleet
A system that passes the transfer test has three observable properties. First, the agents share state: no two agents contact the same customer simultaneously, no job is dispatched twice, and no invoice is sent before the geofenced checkout confirms the technician left the site. Second, the hard 20 percent of decisions, the ones that require judgment, are escalated to a human with full context already assembled, not handed off as a raw problem. Third, every action is logged and auditable, so a new operator stepping in on day one can reconstruct exactly what the system did and why.
The WeLaunch orchestration brain is built around these three properties. The Facility19 control tower runs eight agents plus one brain across a twenty-truck fleet, handling dispatch, compliance, and overtime. Technician-hours recovered, compliance deadlines met without manual intervention, and dispatch collisions eliminated are all measurable in production, not in a model. See how the orchestration brain handles dispatch and compliance in a live fleet deployment.
The Five Questions Every Home Services PE Buyer Should Ask Before Trusting the Pitch
These questions are not designed to be adversarial. They are designed to separate a running system from a well-constructed narrative. A vendor with a live deployment will answer all five without hesitation. A vendor with a pitch deck will deflect at least two of them.
What ServiceTitan and Jobber Tell You, and What They Do Not
ServiceTitan and Jobber are the two most widely deployed field service management platforms in home services. Both are genuinely useful tools. ServiceTitan provides intelligent dispatch, a full CRM, recurring invoicing, and detailed job-costing flyouts. Jobber gives smaller operators a clean scheduling interface, fast invoice generation, and on-site payment capture. Neither platform runs the work. Both platforms record it.
The distinction matters in diligence because a target running ServiceTitan or Jobber will have clean data, organized job histories, and legible reporting. That data is valuable. But the presence of that data does not mean the back office is automated. It means the back office is documented. The dispatcher is still making the routing call. The billing coordinator is still reviewing the aging report. The operations manager is still deciding whether to escalate the missed job or absorb the overtime. The software gave them better information. It did not replace the decision.
When a vendor tells you they have integrated with ServiceTitan or Jobber, the follow-up question is: which decisions does your system make autonomously, and which decisions does it surface to a human? If the answer is that the system surfaces everything and the human decides everything, the integration is a data pipe, not an orchestration layer. Read how the WeLaunch orchestration layer differs from a field service management integration.
One Brain Across the Portfolio: The PE Case for Horizontal Orchestration
The capital-first AI roll-up model has a structural inefficiency that becomes visible at scale. Each acquired company gets its own AI implementation project. The accounting firm gets one set of agents. The pest control company gets another. The HVAC fleet gets a third. Each implementation is bespoke, each one requires a deployment team, and each one starts the transfer-test clock from zero. The brain does not compound across the portfolio. It replicates, at cost, one company at a time.
The alternative is a horizontal orchestration brain that is portable across verticals. The same agent framework, the same MCP connectors, the same shared-state architecture that prevents double contact in a facility management fleet can be redeployed into a home services platform, a pest control lifecycle, or a legal billing operation without rebuilding the brain. The vertical agents, the ones named for specific workflows like dispatch, checkout, and dunning, are the proof points. The brain is what the PE partner is actually buying.
According to FTI Consulting's 2026 Private Equity AI Radar, among portfolio companies that have moved AI into production, 95 percent report meeting or exceeding their business case expectations, but only 7 percent have achieved enterprise-scale deployment across the portfolio.
That 7 percent number is the gap the horizontal brain closes. A single runtime redeployed across every portfolio company means the second deployment is faster than the first, the third is faster than the second, and the diligence question at each new acquisition shifts from "can we build this?" to "how long does the redeploy take?" That is a fundamentally different investment thesis, and it is the one WeLaunch is built to support. See one brain running across multiple verticals on the WeLaunch platform.
The Density Argument: Why the Loop Compounds After Close
The financial case for an AI-native back office in home services is not just about margin improvement during the hold period. It is about what happens to customer acquisition cost as the system accumulates route data, review data, and service history. Every completed job produces a data point about the customer, the address, the technician who performed the work, and the review that followed. A system that reuses that data to find the next customer on the same street, route the same technician back to the same neighborhood, and trigger the renewal before the customer thinks to shop around is not just more efficient. It is compounding.
The 64,000-customer lifecycle that WeLaunch has sized and automated in home services is not a static number. It is a loop: lead, book, dispatch, service, review, invoice, collect, and back to lead. The roughly ten-times model ROI on that lifecycle comes from the loop running autonomously, not from a one-time efficiency gain. At a twelve-times EBITDA exit multiple, a recovered churn dollar that re-enters the loop is worth twelve dollars at exit. A system that prevents the churn in the first place, by catching the eleven-month anniversary cliff before the customer cancels, is worth more than a system that recovers the customer after the fact. Read how the anniversary cliff and collection leakage interact in a home services churn model.
The office is empty. The work is done.
Frequently Asked Questions
What is the transfer test in private equity AI diligence?
The transfer test is the question of whether an AI system continues to run a business's core operations without the founder or key operator present. A system that passes the transfer test handles dispatch, billing, compliance, and escalation autonomously, with full audit logs, and does not revert to manual operation when institutional knowledge walks out the door at close.
How do modelled projections differ from verified mechanisms in a home services AI pitch?
A modelled projection assumes the system will produce a specific outcome once deployed. A verified mechanism is a system already producing that outcome in a live environment. The difference matters at exit because modelled projections that do not survive deployment compress the EBITDA multiple the buyer underwrote at entry.
Why is a single AI metric not enough to evaluate a home services vendor?
A single metric, such as a reduction in windshield time, does not show whether the system traded one outcome for another. Two or three metrics moving together, for example windshield time, technician utilization rate, and average collection time, demonstrate that the system improved the operation without degrading a parallel variable.
What is the difference between a field service management platform and an orchestration brain?
A field service management platform like ServiceTitan or Jobber records work, organizes data, and surfaces information to human decision-makers. An orchestration brain makes decisions autonomously, dispatches work, triggers billing, suppresses double contact, and escalates only the decisions that require human judgment, with full context already assembled.
How does a horizontal orchestration brain benefit a PE portfolio versus a bespoke AI implementation?
A bespoke AI implementation starts the deployment clock from zero at each new acquisition. A horizontal brain is portable: the same agent framework, shared-state architecture, and MCP connectors redeploy across every portfolio company, making each successive deployment faster and cheaper than the last, and compounding the operational advantage across the portfolio rather than replicating it at cost.
What should a PE buyer ask about AI governance before closing a home services deal?
Ask whether every agent action is logged and auditable to the individual decision level, whether the system has a defined escalation protocol for the decisions it does not make autonomously, and whether the vendor can demonstrate that agents share state to prevent double contact. Governance is what makes autonomy safe to underwrite, and a vendor that cannot answer these questions specifically has not built governance into the system.
Take the Next Step
If you are evaluating a home services acquisition and want to see what a live, production-grade orchestration brain looks like before the next diligence call, two options are available.
