In the world of AI-driven customer service and experience, autonomy is both the promise and the problem.
The advancement of agentic AI has helped bring contact centers into a new era of automation. But when machines start improvising, who’s keeping them in check?
At Scorebuddy, that question has a clear answer.
The company’s GenAI Auto Scoring solution is designed not only to enhance quality assurance (QA) for human agents but also to keep a watchful eye on the growing number of autonomous AI tools operating on the front lines of CX.
These checks are essential in making sure that your AI is working for you rather than against you, as Emmanuel Doubinsky, Scorebuddy’s CEO, explains:
“Agentic AI needs to be monitored. Because it will start making decisions that you don’t necessarily want to happen.”
What is Agentic AI? And Why Does It Matter?
Before deciding on a strategy to effectively oversee your agentic AI offerings, it is essential to understand precisely what the technology is.
With its rise in popularity over the past 12 months, the actual definition of agentic can sometimes get lost amongst the excitement.
Doubinsky explains how “many people mistake agentic AI with AI agents.
“AI agents automate front-end support; agentic AI is about letting AI decide the workflow itself. It doesn’t just follow rules, it figures out what the rules should be.”
Agentic AI turns deterministic workflows into non-deterministic ones.
In practice, that means customer journeys are no longer tightly scripted; instead, the AI adapts in real time, bringing in documents, surfacing context, and choosing next steps.
It’s a leap forward in automation. But it’s also a compliance and accountability headache.
Quality Assurance Needs a Redesign
The benefits of agentic AI have been preached far and wide, with efficiency being one of the most apparent.
However, Doubinsky argues that “when you deploy agentic AI in the contact center, it’s not just about saving time; it’s about making sure your AI behaves like a good agent.”
While that may seem relatively straightforward, it is often a lot harder than it sounds.
Agentic AI can be “very creative in addressing unexpected problems,” but that same creativity can introduce hallucinations, bias, and behavior that falls short of regulatory or brand standards.
In other words, just because the AI can respond doesn’t mean it should.
To manage that risk, Scorebuddy’s QA tools are evolving … fast.
How GenAI Auto Scoring Bridges the Gap
Scorebuddy’s GenAI Auto Scoring fights fire with fire by leveraging elements of agentic AI itself, but with tight controls.
“We are starting to use it, but with a lot of precautionary measures,” says Doubinsky.
“We don’t promote it as a front-end tool. It’s a backend service, and we’re cautious with how it behaves.”
That caution is critical, especially as Scorebuddy looks to automate QA processes without sacrificing accuracy.
By blending agentic AI with human oversight, the platform enables organizations to:
- Automatically score a large percentage of QA criteria across every interaction
- Use customizable scorecards to match both human and AI agent performance to compliance benchmarks
- Deliver analytics that surface emerging issues before they escalate
This, in turn, frees up human evaluators to focus where their expertise is most valuable: edge cases, coaching, and complex problem-solving.
For Doubinsky, “it’s not about replacing the human.
“It’s about asking them to do what they’re best at. And that’s understanding context, nuance, and making judgment calls.”
Building Trust Through Feedback Loops
Another major benefit of GenAI Auto Scoring is the impact it’s having on agent morale.
Surprisingly, automation has actually made QA feel fairer to many agents.
Doubinsky explains that previously, only two percent of calls were samples, with agents often feeling that they did not accurately reflect the overall level of customer service being provided.
“Now, with auto-scoring analyzing 20%, 30%, or even 100% of conversations, agents see it as less biased.”
More comprehensive data also means more targeted coaching.
As AI handles routine queries, agents are dealing with more complex tasks, which necessitates better support.
GenAI Auto Scoring doesn’t just flag issues; it helps managers pinpoint where individual agents need training or recognition.
“It’s a self-serving circle,” Doubinsky adds.
“AI enables you to better coach humans and make them better at their jobs.”
Walking the Walk
Hearing about a vendor’s capabilities in theory is one thing, but the proof is always in the pudding.
One of the clearest examples of Scorebuddy’s solutions in action is the company’s work with Intercom – an AI-first company that, in Doubinsky’s words, doesn’t just talk the talk: “they promote it, they do it, they live it.”
To become Intercom’s QA partner, Scorebuddy had to prove its AI-first credentials.
Together, the two companies are pushing the boundaries of what responsible AI in QA can look like.
But the partnership isn’t just technical, it’s also cultural.
Intercom’s rollout of AI automation was deliberately collaborative.
“They didn’t impose it,” says Doubinsky.
“They deployed it progressively, with full buy-in from the agent population. We learned a lot from that.”
Responsible Autonomy Starts with QA
Agentic AI is already reshaping how contact centers function. However, the shift from rule-following bots to decision-making systems means QA can’t be an afterthought; it must be the foundation.
Scorebuddy’s GenAI Auto Scoring is a step toward that future.
By balancing autonomy with oversight, and automation with transparency, it helps ensure that both humans and machines are delivering the kind of customer experience that builds genuine trust.
As Doubinsky puts it:
“You never finish proving you’re AI-first. That’s the beauty of it. It’s continuously changing.”
