The customer rebuilt a revenue-critical inbound sales call center, replacing a custom-built tool that had become a compliance liability. Within months of going live with Zingtree, the new platform was carrying 100% of volume on the largest and most profitable segment of sales traffic.

The challenge: A custom-built scripting tool that broke when anyone touched it

After years of staffing fluctuations and shifting priorities, the customer’s custom-built scripting solution started to become a compliance liability due to the lack of ownership. The code base had slowly drifted out of compliance with the regulations governing credit card collection in the company's primary market.

Three pain points dominated:

• The development cycle had become a bottleneck. Even simple changes took days to weeks, because nobody on the current team had deep knowledge of the legacy code. 
• The agent experience was a static, one-size-fits-all flow that couldn't adapt to caller intent or product mix. 
• Every attempt to plug new tech into the old platform meant building backwards compatibility that no one wanted to own.

"It was more the threat, the looming threat of downtime and missed revenue that was the scariest thing for us. Beyond that, there were so many revenue and growth drivers we wanted to put in, but we just couldn't." – Staff Product Manager, performance marketing & consumer sales

Implementation: A five-month build, a phased pilot, and a J-curve that closed faster than planned

Build and QA took five months. A pilot launched in January 2026 with a small handful of onshore agents, expanded to a 50/50 split at an offshore center in February, and reached 100% of volume on the largest and most profitable slice of traffic by spring. Niche flows for small business, commercial, and foreign language calls are on the roadmap for Q2 and Q3.

"The J curve was, for the initial group, fairly small. There were so many new capabilities we added to the flow that kind of offset the friction of shifting to a new system." – Staff Product Manager

Solution: Nested workflows, connected objects, and a platform anyone on the team can edit

Nested, iframed workflows let the sales operations team build multiple modules that appear on the agent's screen simultaneously, with a parent "framework" workflow progressing through the macro steps of the sales call. 

Connected objects bridge Zingtree to the company's full internal stack: data science models, recommendation engines, fraud and PII checks, and other decisioning APIs the engineering team already operates. The team didn't have to rebuild any back-end logic to migrate. They wrapped existing services in APIs and plugged them into the workflow. 

"One of the best things about Zingtree is the connected objects. The ability to reach out to any REST API, take in that response, and integrate it with the overall workflow. The back end we owned and supported, we didn't really have to change much of it. We were just able to plug it into Zingtree." – Staff Product Manager

‍

The same pattern is how AI is running in production today. The company's in-house AI teams build the underlying capabilities; Zingtree surfaces them inside agent workflows through connected objects, putting AI-driven recommendations and decisioning in front of agents at the moments where they move the sale.

"Some teams already in-house focus on building out AI capabilities, and we've plugged a lot of that into our flow already. That's through building different processes and then using a connected object to bring in the work." – Staff Product Manager

‍

By February 2026, the team had published more than 250 workflows, made over 16,000 edits, and run more than 3,500 agent sessions through the new platform.

Results: Revenue per net call, a compressed migration dip, and a cycle-time reduction the business can feel

Revenue per net call is the team's biggest sales metric, capturing both how many callers convert and how much revenue each generates. 

Every workflow change shipped on Zingtree now influences that number by lifting through-rate at a step in the funnel, tailoring the product pitched to caller intent, or trimming talk time. 

For the first time, the team can build different agent experiences for different caller types inside a single platform: expanded scripting for hesitant buyers, streamlined flows for high-intent callers ready to close. 

"Revenue per net call is probably our biggest metric. Each step we've built out in Zingtree, getting more through-rate on each step is ultimately going to increase that revenue per net call number." – Staff Product Manager

‍

Changes that used to take days to weeks now take days to hours, and many production fixes ship in under an hour. In a regulated vertical where compliance issues have to be patched quickly, that speed is itself a revenue protection mechanism: the legacy platform's slow fix cycles had been a downtime risk, and downtime in this business is missed revenue.

"People who have never touched the flow before in the old system now have the power to do it. More hands make the work a lot faster." – Staff Product Manager

‍

The team had modeled the migration as a couple of weeks to a month of performance dip, a planned cost of switching. In practice, the dip closed in roughly two weeks, compressing what they had planned out as the cost of switching. 

Looking ahead: AI inside a controlled, regulated workflow

AI Actions are next. The team is interested in how Zingtree will let them democratize the building of AI-powered experiences the same way it democratized rule-based flow building, with guardrails layered on top of generative output and the ability to bail out of a generation that doesn't hit the mark.

"When you're in a more regulated industry, you have to be extremely scrutinous around the phrasing of certain things. You need rules living on top of the AI prompts, or other AI systems acting as the judge of the output of the other model." – Staff Product Manager

About the company

A leading performance marketing company that partners with consumer brands to accelerate digital and phone-sales experiences, operating national sales call centers across regulated consumer verticals.

About Zingtree

Zingtree is the agentic context and decision layer that orchestrates complex, high-stakes support and sales workflows safely, accurately, and at scale.