AI Agents for Customer Experience Teams to Get Help More Effective and Accurate? Strategic Insights from Ali Merchant, Founder and CEO of Orvera AI

Gain exclusive insights directly from big-name tech leaders, innovators, and visionaries shaping the future of technology. ExtraMile by YourTechDiet brings you candid conversations where leaders share their big wins, remarkable journeys, and inspiring innovations.

Join us for a top-notch Q&A session with transformative leader Ali Merchant, Founder and Chief Executive Officer of Orvera AI. The firm runs AI agents for customer experience teams, resolving customer conversations across voice, chat, email, and digital channels, inbound and outbound in 90+ languages.

The visionary behind the company, Ali, brings 18 years of experience in running contact-center and health-information operations. He speaks and writes about AI accuracy, AI governance, and how human agents and AI agents work together in customer experience.

In this session, we’re pleased to have Ali wherein he shares his early experience in running contact-center and health-information operations, the role of AI agents for customer experience teams, and the shift from traditional conversational AI to agentic AI.

Ali also sheds light on Orvera AI’s voice agents, the importance of AI governance and security, and, most importantly, human agents handling sensitive customer interactions.

Don’t miss out! Dive into the full conversation to see how AI is transforming customer experience to be more efficient, accurate, and faster, with real resolutions.

1. With 18 years of standout experience in contact center operations and business process management, what key challenges motivated you to work towards AI-powered customer engagement?

Ali. I’ve spent 18 years running contact-center and health-information operations, so I know the floor from the inside. The problems never changed much. Volume spikes you can’t staff for. Customers repeating the same detail to three different people. QA teams that could only listen to a small slice of calls, so most mistakes were never seen. And good people burning out on the same 20 questions every day. What finally moved me was accuracy. My team once built a medical records review tool that reached 86% accuracy, and we never launched it, because records work needs 98 to 99%. That set my bar. AI should take a conversation only when it’s accurate enough for that job. Today Orvera’s AI agents average 98% accuracy across our customer engagements, and that’s the point where I think AI earns a place on the front line.

2. Orvera AI is focused on resolving customer interactions rather than simply automating conversations. What does this strategic change mean for the future of enterprise customer experience (CX)?

Ali. For me, the test is simple. Did the customer’s problem get solved, and did they have to come back? A fast answer that leaves the issue open just moves the cost to the next call. So we count first-contact resolution as no repeat contact on the same issue, on any channel, within 7 days. Across our customer engagements, Orvera’s AI agents average 90% on that measure. Resolution also means the AI agent can act in the systems the business already runs: update the order, record the claim details, rebook the flight. That’s why we build on the customer’s existing stack through 500+ integrations. I think this changes the CX scorecard. Containment rates fade in importance. Resolved issues, repeat contacts and customer satisfaction come to the front, and those are the numbers a CFO and a customer both care about.

3. How is agentic AI changing the way enterprise contact centers and BPOs operate? What makes it unique from traditional conversational AI systems?

Ali. Traditional conversational AI followed a script. It recognized an intent and pointed the customer to an answer or a form. An agentic system takes the next step itself. It checks the policy, changes the booking or starts the return, then confirms it back to the customer, all inside the rules the business sets. For contact centers and BPOs, that shifts the work. Human agents spend more of their day on conversations that need judgment and care, and less on lookups and status checks. Supervisors oversee AI agents much the way they coach a team. In our deployments, one trained supervisor oversees 10 AI agents. BPOs feel this most, because their model was built on headcount. In my view, the BPOs that do well over the next few years will sell outcomes, like resolved contacts and quality scores, and AI agents with human oversight are how they get there.

4. Could you share insights about how Orvera AI’s voice agents automate outbound follow-up calls to identify pending requests and take the next steps?

Ali. Outbound follow-up is where a lot of work quietly stalls. Someone has to call back, find out what’s still open and push it forward. Orvera’s AI agents do that across voice, chat, email and text. They start from the open items in the customer’s own systems, reach the right person, ask where the request stands and write the answer back into the system of record. If something is missing, they ask for it, set the next follow-up or pass it to a human agent with the full context. Record retrieval is a good example. Requests for records sit with hospitals and other custodians, and each one needs repeated follow-up. Averaged across our record retrieval engagements, turnaround went from 35 days to 22 days, 93% of requests are on time against the SLA, and 98% of rejections and exceptions are caught.

5. As AI voice agents and omnichannel automation transform customer experience, how can businesses improve their operations and efficiency without losing the human connection that builds customer loyalty?

Ali. Start by deciding where people add the most, then give them more time there. Most customers want a fast, correct answer to a routine question, at any hour, in their own language. AI agents do that well. Orvera’s AI agents answer instantly on every call and chat, in 90+ languages. The human connection matters most when someone is upset, confused or facing a big decision, like a claim after a storm or a missed connection late at night. That’s where a human agent should step in, with the whole conversation in front of them so the customer never repeats themselves. Agent Assist helps there too, by guiding the human agent live. Done this way, efficiency and loyalty move together. Across our customer engagements, CSAT averages 90% and average handle time is 30% lower. The risk is using AI as a way to hide from customers. People notice that fast.

6. Contact centers in enterprise organizations typically deal with personal, financial, and healthcare-sensitive data. How can businesses ensure proper governance, security, and oversight as they scale AI agents?

Ali. Governance has to be part of the design at the start. Three things matter most to me. First, every AI decision is logged, so you can always see what the AI agent said and did. Second, every conversation is scored, human and AI. We score 100% of conversations against the customer’s own scorecard, including the actions taken in their systems, and the customer’s QA team calibrates and reviews those scores. Third, people stay in charge. One trained supervisor oversees 10 AI agents and can step in at any time. On security, Orvera is SOC 2 Type II certified, HIPAA compliant and GDPR compliant. On models, we stay model-agnostic. A governed orchestration layer sits above third-party models, so a business can move to a better model without re-engineering or lock-in.

7. As AI agents have increasingly started working with customers. How will companies decide when to use AI for independent tasks and when to involve humans?

Ali. I use one test. Is the AI accurate enough for this specific job? That’s the lesson from the 86% tool we never launched, and it applies to customer conversations too. Order status, appointment changes and first notice of loss are well defined, and AI agents can handle them end to end. Situations with real emotion, exceptions to policy, or a decision that changes someone’s life should go to a human agent, with the AI handing over everything it already knows. The line moves over time, and the only safe way to move it is to score every conversation, human and AI, and look at the results every week. When the AI is consistently accurate on a task, give it more. When it slips, pull it back.


Check Out Our Other Informative Interviews:

Strengthening Healthcare Interoperability with Healthcare Data Orchestration: Key Takeaways from Sunita Pradhan, Chief Growth Officer at ELLKAY

Bridging Animal and Human Biology with AI-Driven Drug Development: In Conversation with Dr. Jo Varshney Founder and CEO at VeriSIM Life

Orvera AI Reviews & Recognitions

Orvera AI G2 Reviews Rating
Orvera AI Gartner Peer Insights Rating
Orvera AI Everest Group PEAK Matrix 2026
Orvera AI CMP Core Performing Voicebot 2026
Orvera AI CMP Up and Coming Chatbot 2025
Orvera AI CMP Core Performing Automated QA QM 2026

Ali Merchant

Ali Merchant is Founder and Chief Executive Officer, Orvera AI, which runs AI agents for customer experience teams. He has spent 18 years running contact-center and health-information operations. His view on AI is practical: an AI agent should take a conversation only once it is accurate enough for the job, and every conversation, human and AI, should be scored. He speaks and writes about AI accuracy, AI governance and how human agents and AI agents work together in customer experience.

LinkedIn

Orvera AI

Orvera AI runs AI agents for customer experience teams. Orvera’s AI agents resolve customer conversations across voice, chat, email and digital channels, inbound and outbound, in 90+ languages. On the same platform, Agent Assist guides human agents live, AI Quality Management scores 100% of conversations, human and AI, and Voice of Customer shows why customers get in touch.

Orvera builds, deploys and runs the platform on the customer’s existing stack through 500+ integrations, and full deployment, end to end, takes 3 to 6 weeks. The platform is model-agnostic. A governed orchestration layer sits above third-party models, so customers can adopt better models without re-engineering or lock-in. Every AI decision is logged, and one trained supervisor oversees 10 AI agents.

LinkedInFacebookTwitter