PE add-on integration: the 100-day playbook.

CT Acquisitions states the failure mode more directly than most advisory content manages: the biggest 100-day-plan mistake is making the plan in the first 100 days. By close you should already hold a draft covering what diligence taught you about the business, the team, the customers, the operations and the financials. The first hundred days are the only period in which an acquired organisation accepts large changes without renegotiating them, and spending three weeks of it writing a document burns the window. Diligence answers whether the numbers are real; the 100-day draft asks who does the work and by when.

The economics improve with every deal. Add-on acquisitions accounted for 72.9 percent of all buyouts in 2025, per PitchBook data published in Cherry Bekaert's private equity report as of the end of that year. This is the operating rhythm, not an occasional event, and BowMerge is right that every add-on needs its own 100-day plan even where it overlaps the platform's standing workstreams.

Generic day-1 checklists cover the employee announcement, customer and vendor notification, retention agreements, control of the bank accounts, and a communicated decision about the brand. Do all of it; none of it is specific to your industry.

What is specific is that in a trade services business the revenue arrives by telephone, and on the night of close that telephone is answered by an arrangement nobody at the platform has inspected. Sometimes it is two people in an office who also run dispatch. Sometimes it is a voicemail box after five that nobody has emptied since March. In every case it is now yours, and it is leaking at a rate you cannot currently state.

So the addition to day 1 is instrumentation, not intervention. Get the call detail records, and the recordings where consent and state law allow. Establish in writing the answer rate, the after-hours behaviour and the booking rate per answered call before anyone from the platform touches a process. It takes a day, and it makes every synergy claim for the next two years defensible instead of arguable.

For what an unmeasured inbound leak is worth at platform scale, the call-center leakage calculator runs on your own inputs and returns the number on the page.

The pull in month one is toward moving the acquired brand onto platform systems, because it is visible progress and it is what the investment committee asked about. Resist it for thirty days. The cost of waiting is one month of duplicated reporting; the cost of not waiting is a permanently contaminated baseline you will need every time someone asks whether the integration worked.

What to collect per brand before anything moves: inbound call volume and answer rate by hour and day of week; booking rate per answered call; jobs completed per technician per day; average ticket by trade; membership counts and renewal behaviour; revenue mix; accounts receivable ageing. Most of it is retrievable through the field service system the brand already runs.

Month one also buys an honest technical inventory. V7 Labs describes the steady state: a platform that has completed four add-ons is, on average, running four different accounting systems, three CRM instances, and two or three payroll platforms simultaneously. Nobody planned that; it is what happens when each deal defers rationalization to the next one.

If the acquired brand's records turn out worse than diligence suggested, data readiness for mid-market AI covers what has to be true before any layer above the data is worth building.

The rest of the integration hangs off this decision, and it is usually made implicitly by whoever moves first. Make it explicitly, in the pre-close draft, with the reasoning written down.

The industry default is brand retention. CT Acquisitions lists four options, retain the target brand long term, transition to the buyer brand within six to twelve months, run a hybrid co-brand, or create a new combined brand, and names retention as the most common for PE platforms, family offices and search funds, with no day-1 visible change for tuck-in acquisitions where the target brand is retained. Uprise RI reported the same pattern from the customer's side in 2023: acquirers typically keep operating the acquired company under its original name. Profitability Partners' 2026 acquirer guide names two archetypes, platforms that keep your brand and team intact and centralize the back office, and platforms that fully absorb you.

The workable rule: preserve everything the customer can see, centralize everything they cannot. The asset is the name people already dial, plus decades of local search presence and review history; rebranding converts it into a marketing project.

Preserve the name on the truck, the local number, the greeting a caller hears, the local review profile, and pricing where the market genuinely differs. Centralize dispatch capacity, after-hours answering, payables, payroll, procurement, insurance and the reporting layer. Operators who get this wrong centralize the greeting, the one piece of the back office a customer actually hears.

One honest caveat on the payback numbers. Vendor content quotes utilization gains from centralized dispatch somewhere between ten and thirty percent. We went looking for a primary source we could stand behind and did not find one; every figure we traced terminated in another vendor's marketing page rather than in a study with a method. We will not repeat a number we cannot source, which is exactly why the day-1 instrumentation matters. Your own before-and-after is the only defensible version of that statistic, and it costs one day of work to be able to produce it.

Once the brands stay separate to the customer, centralizing dispatch stops being a database exercise and becomes a routing and capacity-visibility exercise. That reframing makes it achievable inside a hundred days.

Four things have to be true. Every brand keeps its own inbound number and greeting, so a caller to the plumbing brand never hears the HVAC brand's name. Whoever answers can see real capacity for that brand's board, so it never offers an arrival window the dispatcher cannot honour. The booking is written into the system that brand actually runs, against the correct business unit, rather than into an email a coordinator retypes. And one capacity view sits above all of it, with no brand migrated to produce it.

The one-queue temptation is worth naming. Pooling every brand's calls into a single queue with generic scripting is cheaper to stand up, and it is how platforms erase the local feel they paid for. Brand-aware routing costs slightly more to configure and preserves the asset. If you take one implementation decision from this page, take that one.

The architecture is not hypothetical to us. At Jim Glaser Law, a nameable client who takes reference calls, the commissioned build runs five channel-specific voice agents, one per acquisition channel, so attribution lands on answered calls rather than form fills. Swap channel for brand and that is the multi-brand answering layer: an agent and a number per brand, one capacity view behind them, the brand written onto the booking at creation. Our own inbound line is answered by one, at (617) 675-9067.

Underneath sits the ordinary state of a roll-up mid-execution: the platform on ServiceTitan, one acquisition on FieldEdge, another on Housecall Pro, a third on Workiz, finance on Sage Intacct or QuickBooks. Each is a different integration surface; the per-system detail is in the individual playbooks: ServiceTitan AI integration, FieldEdge AI integration, Housecall Pro AI, Workiz AI integration and Sage Intacct AI integration.

The deliverable is unglamorous, and it is the thing operating partners consistently say they lack: one weekly view covering every brand, sourced from each brand's own system, with definitions that survive scrutiny.

The technical work is a scheduled read from each field service system into a store the platform controls, plus a normalisation layer. The genuinely hard work is the definitional argument that precedes it. A completed job in one brand's configuration is a job the technician marked done; in another it is a job that has been invoiced; in a third it is one that has been paid. Until someone with authority rules on each definition, a consolidated dashboard is three brands lying to each other in a shared font.

Two rules keep the phase from expanding. Read only at first: no write-back into any brand's system until the reads are trusted. And publish narrow: five or six numbers the operating partner will act on, not forty nobody reads. The cross-brand reporting shape is covered on workflow automation and revenue operations AI.

By day sixty you hold a baseline, a consolidated view and a real inventory of systems, so the standardization question can be answered on evidence. The answer is usually narrower than the deal model assumed.

Split the estate into three buckets. Standardization is nearly free and obviously right on payroll, benefits administration, insurance and procurement, because no customer experience depends on them. It is expensive and has to be argued on the field service platform itself, because migrating an operating company's dispatch and history is the highest-risk project on the list. And it is actively wrong on anything carrying local identity: the phone number, the review profile, the domain, the brand.

The middle bucket is where platforms overreach. Forcing every brand onto one field service system is defensible when the trades and operating models are similar and the migration cost is smaller than the fragmentation cost you measured. It is much weaker on a mixed platform where a residential HVAC brand and a commercial plumbing brand run different work in different ways. There, an orchestration layer above several systems delivers the consolidated visibility the sponsor wanted at a fraction of the cost. The sponsor-facing version of that argument is on ServiceTitan Pro versus a custom AI build for PE platforms, and the configuration-versus-construction line is on Pro Services versus commissioned custom AI.

Calibrate expectations. CT Acquisitions puts full integration for a lower-middle-market acquisition at roughly eighteen to thirty months, and V7 Labs puts synergy realization, where it happens at all, twelve to twenty-four months after close, with systems work routinely slipping out of year two. Day 100 is where the plan is reviewed and the next two quarters get scheduled, not where integration finishes.

Most vendors writing about PE home services integration imply a portfolio of roll-up case studies. We will be specific about ours instead.

Nameable and verifiable: Jim Glaser Law, a client who takes reference calls, running five channel-specific voice agents across 3,787 AI-handled calls and 5,514 minutes, with per-channel attribution on answered calls. That is a legal services operator, not a home services platform. In this vertical we also run an AI intake layer for a multi-location home services operator, 1,486 calls and 2,203 minutes handled. We work with a legal platform as well, which has not cleared us to publish a name or numbers. Across all clients, more than 6,000 live calls have been handled by commissioned AI layers.

What that means here: the multi-brand answering architecture we would build for a platform is a structural transfer of a system running in production. What we do not have is a published PE roll-up case study with before-and-after utilization numbers. If a competing vendor shows you one, ask which platform, ask to call the operating partner, and ask what the baseline measurement method was.

The wider diligence set worth asking any vendor, us included, is on how to choose an AI consultant for home services, and the sector data we publish is in the home services platform AI benchmark.

A 45-minute diagnosis call names one constraint in a sentence. A prototype runs against your real data inside seven to ten days. If it demonstrates the workflow, the production build follows on a four to seven week cycle at a fixed fee in the $45,000 to $180,000 band, with integration depth setting the number: read-only cross-brand reporting at the bottom, a brand-aware intake layer writing bookings back into two or three field service systems at the top.

What the platform holds at handoff: the integration code, the prompts, the model selection, the sync design and the credentials. No retained licence, no recurring platform fee. The reasoning is on why code handoff matters, the structure is on the pricing page, and the first three weeks are described on how we work.

What we would decline. A migration project, because moving an operating company off its field service system is implementation work a specialist integrator does better. A replacement for anyone's CRM, dispatch or invoicing. A general-purpose assistant over the portfolio, which demos well and changes no number. And any engagement where the brand question is still open, because a brand-aware intake layer built before that decision gets rebuilt after it.

If build-or-buy is still open, start with build, buy or commission rather than with us. If the platform is not ready, why mid-market AI rollouts stall in month four is a cheaper read than finding out at month four.