Should we standardize the ATS across every portfolio company first?

No. A full ATS migration takes 9 to 18 months and stalls the value you bought the portfolio to create. Start with a shared data layer that pulls from each brand's existing system. You can standardize core processes and run AI on top of disconnected systems now, then consolidate platforms later when the timing makes sense. Lead with the layer that moves margin in 90 days, not the migration that moves it in two years.

Will a shared-services model kill the brand identity that made each firm valuable?

Only if you centralize the wrong things. Keep the client-facing identity, the recruiter relationships, and the niche expertise separate. Centralize the back-office work clients never see: payroll, compliance, candidate sourcing automation, and reporting. Done right, the brands keep their face and lose their overhead. The market still sees three specialized firms. You see one cost base.

How do we know if the shared-services model is actually working?

Track spread per producer and back-office cost as a percent of gross profit, by brand, every month. If spread per producer climbs while back-office cost as a percent of gross profit falls, the model is working. If headcount grows faster than gross profit, it is not. Set a baseline in the first 30 days so you can prove the change instead of guessing at it.

Scale a PE Staffing Portfolio Without Adding Headcount

Q1 2026 was the most active staffing M&A quarter in three years. Most acquired firms run disconnected tech stacks. A shared-services AI operations model is the fastest way to expand margin across a multi-brand portfolio without growing headcount.

Q1 2026 was the most active quarter for staffing M&A in three years. PE firms are buying and bolting on at a pace we have not seen since 2021. But here is what shows up after the deal closes: three, five, sometimes eight acquired firms, each running its own ATS, its own payroll system, its own reporting, and its own definition of a placement.

That is not a portfolio. That is a pile of companies that happen to share an owner.

The instinct is to grow your way to the EBITDA target. Add recruiters. Add a few ops people per brand. But headcount that grows in lockstep with revenue does not expand margin. It holds margin flat while the top line gets bigger. PE math does not reward that.

The fastest lever you have is operational architecture. Build a shared-services AI operations model once, run it across every brand, and you grow gross profit without growing the cost base under it. That is where the margin lives.

Why Disconnected Stacks Quietly Eat Your EBITDA

Most acquired staffing firms were run lean and profitable as standalone businesses. The owner knew everything. The systems were good enough for one company. Then you put five of them in a portfolio and the cracks show up fast.

You cannot see across the brands. One firm closes a contract role in 6 days. Another takes 19 for the same kind of req. You have no idea why, because the data sits in five systems that do not talk to each other. You are paying for the same software five times. You are running five separate compliance processes. You have five people doing the same back-office task that one person with the right tools could do for all five.

None of this is dramatic. It is a slow leak. A few points of margin here, a few there. Across a portfolio doing $200 million in revenue, a slow leak is a very expensive problem.

The mistake leaders make is treating this as a software problem. It is not. It is an architecture problem. You do not need one system. You need one operating layer that sits across the systems you already have.

The Three Layers That Drive Margin Across Brands

A shared-services AI operations model has three parts. You build them in order. Each one earns its keep before you fund the next.

Layer One: The Shared Data Spine

You cannot manage what you cannot see across brands. Before you touch AI, you need one place where every brand's data lands in the same shape. Same definition of a placement. Same definition of a submittal. Same way of counting spread.

This does not mean ripping out anyone's ATS. It means pulling the data out of each system nightly and standardizing it in one warehouse. Now you can compare brand to brand on the same terms. Now you can see that Brand C's time-to-fill is double Brand A's and ask the real question: what is Brand A doing that the others are not?

Get this right and you have already paid for the project, just from killing duplicate software contracts and seeing your real numbers for the first time.

Layer Two: Centralized AI-Driven Back Office

This is where headcount stops scaling with revenue. The work that every brand does the same way moves to a shared team running AI tools.

Candidate sourcing. One AI sourcing engine feeds qualified candidates to recruiters across every brand, instead of five separate teams searching the same talent pools.
Screening and matching. AI ranks inbound applicants against open reqs so recruiters spend their hours on the top 10, not the top 200.
Compliance and onboarding. One automated process handles I-9s, background checks, and document collection for all brands.
Payroll and billing. One team, one toolset, every brand.

The recruiters and salespeople who carry the brand stay where they are. They keep the client relationships and the niche expertise that made the firm worth buying. What they lose is the administrative weight that slowed them down. A recruiter who used to spend 40 percent of the week on sourcing and paperwork now spends it selling and closing.

Layer Three: One Reporting and Decision Rhythm

The data spine gives you visibility. This layer turns it into action. Every brand reports on the same dashboard, every week, using the same metrics. The ops leader for the portfolio runs one review rhythm, not five.

When spread per producer drops at one brand, you see it the week it happens, not the quarter it happens. When a sourcing change works at Brand A, you roll it to the other four in days. The portfolio learns as one organization instead of five that never compare notes.

The Metrics That Tell You It Is Working

Margin expansion is the goal, but it is a lagging number. Watch these to know you are on track before the EBITDA shows it:

Spread per producer, by brand. This should climb as the back office takes weight off your producers.
Back-office cost as a percent of gross profit. This should fall as the shared team absorbs work the brands used to do separately.
Time-to-fill variance across brands. This should shrink as best methods spread.
Software spend per brand. This should drop as duplicate contracts die.

If gross profit is growing faster than headcount, the architecture is working. If you are adding a body for every increment of revenue, you built five companies again, not a portfolio.

Start Here This Week

Pick your two largest brands. Pull last quarter's data from both into a single spreadsheet and force them into the same definitions: placement, submittal, spread, time-to-fill. You will spend a day fighting over what each term means, and that fight is the point. It shows you exactly where your portfolio is flying blind.

That one exercise tells you where the margin is hiding and what to build first. Do it before you approve another headcount request.

Scaling a PE-Backed Staffing Portfolio Without Adding Headcount