Narrow Scope And Early Buy-In Position Contact Center AI To Succeed

Contact centers are under pressure to put AI to work, and the options multiply with every vendor demo and conference keynote. The teams seeing real gains resist the urge to chase every new capability at once. They start with a narrow, clearly defined problem and bring their people along before the technology arrives. Then they let what they learn from one deployment decide the next. In a business where any inbound call could become a sale, that discipline also keeps automation from costing revenue.

Tewana Ramseur is Vice President of Customer Experience at Bath Saver, where she's responsible for every non-face-to-face customer contact from pre-sales through post-installation service. She was named one of CX Current’s Top 50 Leaders Transforming CX for 2026 for her work rethinking the workflows and operating models behind service. Her team is now building an AI program to support its agents, and they're doing it one deliberate step at a time.

"Figure out what’s most important and then don’t turn the fire hose on," she advises. "What’s the next thing that’s going to take care of the problem you’ve identified, and what’s the one after that?" For Ramseur, that sequencing starts well before any technology goes live.

Start with the purpose and the people

Ramseur’s first test for any new workflow is whether it serves a clear business purpose. "There are all kinds of bells and whistles out there right now, especially as we’re all looking for ways to effectively integrate AI into the work that we do. What you don’t want to do is just jump on that cool thing someone saw on a webinar somewhere," she says.

Once the purpose is settled, she puts transparency ahead of the rollout plan. That includes being candid with her leadership team about the open questions. "I want to make sure they know what I know and what I don’t know, to the extent that I can share that, because I don’t want any of this to come as a surprise to them," she explains. "It’s rare that a new workflow, if it's really going to have impact, doesn’t affect the organization and potentially the people in that organization."

She also recruits ambassadors from the contact center floor early in the process, because adoption decides whether the work succeeds. "You can completely derail a new workflow or a new initiative if you don’t have the people on board, understanding it and trusting in you and the decisions you’re making." The same restraint governs scope. After a small first release, she's asked her leaders to think about expansion, but within firm limits. "When you start talking AI, your imagination just goes crazy and there’s all this great stuff we could do. We could, but we’re not going to do that. We’re going to keep it right here," she says.

Customer tolerance sets the guardrails

Where those expansion limits sit depends on where a customer is in their relationship with the company. On the pre-sales side, the goal is to book in-home sales appointments. When the office is open, Ramseur wants people handling those conversations on every channel, including chat and SMS. The AI agent has a narrower job: capturing the opportunities that come in after hours. "We're not here 24/7, so where AI comes into it for us is in scheduling those appointments so that we don’t lose an opportunity."

Once someone becomes a customer, the calculation shifts. "They’re a lot more tolerant of interacting with the AI agent on channels other than voice, because at that point they just need an answer. They need an update on something," Ramseur notes.

The company installs tubs and showers, so weekend water problems are common. AI can help triage whether a technician needs to visit the home or the issue can wait until the office reopens. Throughout, Ramseur keeps a person within reach for a customer base that can skew older. Anyone who wants a human during business hours can get one.

Begin where revenue is at stake

Ramseur’s team tends to start from the revenue side of the business. Her first target was a habit customers bring from everywhere else. "Many of us have figured out that if a company is going to make money on the reason we call, they’re going to answer the phone," she points out.

Because Bath Saver’s sales line answers quickly, it became the front door for nearly every call to the company. That included service requests, calls for the marketing department and job seekers. "One of the first things we realized we could do was use an AI agent to filter those calls to the place that they need to be, and that stops the appointment center from being a receptionist," Ramseur says. It sends the sales team the calls that matter and everyone else where they belong.

Building the agent with Cresta as the AI partner meant mapping every reason customers call and where each call should land. The team identified the keywords that signal intent and the follow-up questions that separate one appointment type from another. They also found that one large "other" bucket could route to a single destination. "We worked on defining what those call flows were and then worked with our partners on building an agent that would follow those trees to get the calls to the right place," Ramseur shares.

Rethink the team, not just the workflow

Ramseur treats the first deployment as a way to learn the technology and to work out what roles her organization will need to support it. That rethink is already underway. "I’m telling my team that a 'hybrid environment' isn’t working from home and in the office anymore. It’s AI agent, human agent," she says.

Quality assurance is shifting from a scorecard toward a source of insight. Her Voice of the Customer program manager, who's worked from social media and third-party platforms, will soon have every interaction on every channel to draw from. That data will feed operational suggestions that head off problems before they happen.

Next on the list are service efficiencies, followed by the chat channel once the agent has learned enough to handle it well. "You don’t want too much coming at you from too many directions, or I don’t think you’ll realize enough of the benefit of what this thing can give you and the information it can provide," she says.

Her advice on choosing a technology partner comes back to the people. "A technology partner shouldn't be so driven by getting you deployed that they don’t recognize the space, the need and the focus that has to be on the people to get you there," Ramseur concludes.