PE Portfolio Company: 30-Person Backoffice Becomes 6 Humans + 14 Virtual Employees, EBITDA Up 18 to 22% in 8 Months

The operating partner ran a thesis across the portfolio: every portco was paying vendor margins on operations the portco should have owned. Across the portfolio, the combined backoffice cost line ran into the eight-figure range, and the savings story alone was not the unlock. The unlock was that every dollar of backoffice spend was funding a vendor’s institutional knowledge instead of the portco’s. The operating partner knew where this ended. At exit, the buyer would underwrite the portco’s operations as a function of vendor contracts, not as a function of the portco’s own AI capability. That is the difference between an EBITDA multiple and a revenue multiple.

Reliable Group ran the Blueprint across the portfolio in parallel. The Blueprint produced one AI-native org chart per portco, a Virtual Employee roster with role-by-role unit-of-work pricing, a hiring plan for the offshore team inside each portco’s COPO entity, and the joint unit economics across both levers.

Each portco’s 30-person backoffice was redesigned around six humans (judgment, exceptions, escalations, relationships) and fourteen Virtual Employees handling defined workflows: claims processing, document extraction, compliance auditing, vendor reconciliation, customer support triage. Persistent memory architecture sat inside each portco’s data boundary. Governance controls were defined per role with audit trail.

Reliable Group set up a Company-Owned, Partner-Operated entity inside each portco. The entity hired the six humans. RG handled facilities, payroll, compliance, ongoing operations. The portco owned the entity, the team, the data, and the AI workflows.

Per-portco EBITDA expansion in the 18 to 22 percent directional range, after netting Reliable’s operating fee. Per-portco backoffice operating cost reduced an estimated 35 to 45 percent versus the prior vendor stack. Time-to-deploy from Blueprint sign-off to first production Virtual Employee per portco: 7 weeks. Combined token-cost run-rate across the portfolio came in 12 percent under the Blueprint forecast, validated against the unit-economics model RG built first inside its own operations.

Every claims pattern, every exception edge case, every workflow improvement, every fine-tuning decision now accrues inside the portco’s COPO entity. At exit, the buyer underwrites the portco’s operations as a tech-enabled services business with proprietary AI capability inside the entity, not as a vendor-managed cost center. That changes the multiple range the buyer is willing to defend.

The decision the operating partners kept coming back to: did the institutional knowledge sit inside the portco or inside a vendor. The Blueprint scoped both levers together because the operating partners did not want to optimize one and leak the other.