Turning a year of disconnected AI pilots into one funded, phased rollout with the numbers to defend it.
178%
Run-rate ROI once both phases reach ceiling
$1.95M
Three-year net savings
~5.5 mo.
Payback period
The situation
A private equity-backed global payments technology company had spent more than a year on disconnected AI pilots: several vendors, several business units, no established sourcing process to drive toward a decision. Its sponsor had surfaced vendor options, and the CEO had set a clear mandate: deflect call volume, improve handling efficiency, and scale the contact center without adding headcount.
Why this was hard without help
Five vendors entered the evaluation, and each priced AI differently: per minute, per resolution, and hybrid structures that couldn't be read side by side. Without a common framework, the decision risked coming down to whichever demo looked most impressive, not which platform the numbers actually supported.
What Resourcive brought
The company's VP of Operations and Global Head of Procurement brought Resourcive in to give the effort structure. A weighted scorecard normalized every proposal into a common one-year and three-year cost view, and client stakeholders scored the finalists independently before the results were aggregated, which put two vendors in the final round.
The team did not select the cheapest option, which came in roughly $1M below the winner, and did not simply follow the strongest demo. It chose the vendor whose per-minute pricing could be forecast and defended to the board, where a competing per-resolution model left too much room for interpretation.
The rollout
- Phase one: a virtual agent across a 25-agent portfolio handling roughly 35,000 calls a month, ramping deflection from a 40% start toward a 70% ceiling.
- Phase two, about six months later: a second virtual agent across 45 agents spanning two further business units, ramping toward a 60% deflection ceiling.
- From day one: Agent Assist across every human-handled call and case, cutting average handle time by roughly 15% while deflection was still ramping behind it.
Contract minutes were sized against the resulting financial model rather than the vendor's initial estimate, keeping the commitment tied to what the rollout would actually use.
Results
Documented outcomes
- 178% run-rate ROI once both phases reach their modeled deflection ceilings, on roughly $1.10M in net annual savings.
- $1.95M in three-year net savings across a deliberately conservative, phased ramp.
- $612K in negotiated cost avoidance against the vendor's initial proposal.
- Roughly 44 FTE equivalents freed once the rollout reaches ceiling.
- Payback in approximately 5.5 months from first phase go-live.
Figures reflect the two funded phases at modeled call-deflection ceilings of 70% and 60% respectively, plus Agent Assist at 15% on all human-handled calls and cases. Additional portfolios beyond these 70 agents represent future upside not counted here.
