.jpg)
"Nobody wakes up and says my dream is to deploy AI,'" jokes Rob, the podcast host. "They wake up and say, 'Please, for the love of all that is holy, let my team stop drowning in tickets.’”
In this no-nonsense conversation with Saurabh Kumar, CEO and co-founder of Rezolve.ai, the discussion cuts through AI hype to tackle what IT leaders actually care about: measurable outcomes that prove ROI to CFOs.
Drawing from his experience both implementing enterprise technology and building AI solutions, Saurabh delivers a framework for measuring AI success in IT support—complete with realistic timelines, honest numbers, and the admission that "99.9% of the time" organizations are categorizing tickets completely wrong.
Stop accepting vague AI promises. Start demanding upfront ROI visibility. This podcast reveals why traditional metrics like First Call Resolution are "almost pointless," how Rezolve DeskIQ shows exact automation potential before deployment, and why 40-70% ticket reduction beats hyperbolic 100% elimination claims. Plus: the case for predictive metrics that prevent problems instead of reporting them after they've already damaged satisfaction scores.
🎧 Listen to the Full Podcast
Watch the whole episode on the CEO’s insights on AI ROI:
The Three Things Service Managers Actually Care About
Rob pushes Saurabh beyond generic AI talk: "When companies roll out Rezolve.ai, what are the top three things they're secretly hoping will happen in the first 90 days?"
Saurabh, speaking from his experience running IT support organizations, identifies three core outcomes:
Employee Experience: "It's really important to deliver something that people are happy with. Otherwise, they're going to tell all of their co-workers and their boss that they interacted with IT ticketing/ support and it's just not that great."
Cost Control: "At the end of the day, support is really a cost driver inside the organization. The less you spend, the better your bottom line is—generally what every CFO wants to see."
Organizational Productivity: The "next level metrics"—improved agent productivity, increased self-service, and better technology adoption.
The key insight? Success isn't about deploying technology. It's about delivering these three outcomes measurably.
AI Is Different: The Living System Problem
When Rob asks about timelines, Saurabh makes a critical distinction:
"AI is not like that. AI is a living, breathing system. You really have to think about success as something that you measure and improve over time."
Week 2: Early signals appear—survey results from early adopters, usage metrics showing task completion, thumbs up/thumbs down satisfaction indicators, sentiment analysis compared to pre-AI baselines.
Month 1-3: Measurable impact emerges through deflection rates and CSAT scores, though continuous iteration remains essential.
"These are the early indicators—the canary that you can begin to keep an eye on," Saurabh explains.
The Numbers Game: Why 40-70% Beats 100%
Rob doesn't let Saurabh off easy: "Give me some realistic ranges."
Saurabh's response is refreshingly honest: "For most companies, I see these hyperbolic claims around 'we'll eliminate 100% of your tickets.' It really doesn't work like that in real life."
Every organization has unique systems, industry requirements, workforce types, and legacy technical debt. The realistic target? 40-70% reduction in overall ticket volume over time.
Two factors determine velocity:
- Adoption strategy: Communication aggressiveness, policy changes requiring service requests through the AI platform, behavioral nudges like proactive password expiration reminders.
- Deflection precision: A well-defined point of view on what to deflect and how.
This is where Rezolve DeskIQ becomes Saurabh's "Moneyball for IT"—showing you exactly where the wins are hiding before writing a single line of code.
The Rezolve DeskIQ Revelation
Rob asks: "For people hearing that term for the first time, what exactly does Rezolve DeskIQ do?"
Saurabh's explanation is characteristically direct:
"Rezolve DeskIQ is what it says it is—it's IQ about your desk. You run a service desk, and we're giving you insights and intelligence about your service desk."
The platform ingests historical service data from any source—ITSM platforms, emails, Excel sheets, and live chat transcripts. Then a cutting-edge agentic AI platform analyzes each ticket, categorizes it properly, and scores automation potential.
Password resets? Perhaps 100% automation potential. Hardware failures? Perhaps very little based on their nature.
But here's the revelation that makes every IT leader sit up:
"99.9% of the time; those existing ticket categories are almost always mischaracterized."
Agents choose the path of least resistance when categorizing tickets. When AI reads the actual detail and categorizes as a human would, organizations suddenly have "a clean, high-confidence view into what your team is spending time on."
Forget ROI for a moment—just having accurate categorization is transformative.
Real Results: 90%+ Satisfaction and 9/10 Success Rates
Saurabh shares client outcomes that validate the framework:
Client 1: "We had a client get 90%+ satisfaction. This is unheard of in the IT support industry. Most people just hate that interaction."
What drove it? Faster service (hours to seconds), easier omnichannel access, Teams integration, 24/7 automated resolution.
Client 2: "Nine out of 10 times the system was giving the users what they needed and they said 'this solved my need.' That's a very high bar to get to."
Mini Case Study on AI ROI
Black Angus Case Study: The restaurant group reduced after-hours support calls from 90% to 10%—an 80% reduction—through Desktop Automation in Microsoft Teams, without adding headcount.
The Most Overrated Metric in IT
Rob's playful question—"What is the most overrated metric in IT?"—triggers Saurabh's most passionate response:
"The most overrated metrics in the entirety of the service industry are what I would call the backward-looking set of metrics, which is pretty much all the metrics that we track in the service industry today."
First Call Resolution?
"You can define what a first call is in a way that the metric always looks healthy. When you control that definition, you've already won."
Mean Time to Respond?
It measures what has already happened. "That time has passed. I can't go back and fix it. People who had that bad interaction are going to be unhappy about it."
The future?
Predictive metrics powered by AI:
- Identifying which tickets need escalation before they breach SLA
- Detecting complexity patterns requiring specific expertise
- Recognizing users with persistent issues
- Proactively routing to the right person in real time
"It's not about reporting what happened a month ago—it's about responding in the moment to what's happening now."
The Simplest First Step Tomorrow
Rob's final question: "If someone listening wants to start making their AI measurable, what's the simplest first step they can take tomorrow?"
Saurabh's advice for service desk leaders:
"Start with something like Rezolve DeskIQ. Get a measurable insight into how you're going to change your service even before you've started to work on changing it."
For other AI implementations: "Be very clear on what two or three things you're trying to change and what's the easiest way to measure that change. If you define that measurement ahead of time, then the entire program aligns to that outcome."
What doesn't work? "Saying 'I've now delivered a project and I'm going to figure out how to measure the success once it's all done.' That doesn't really work well."
Key Takeaways
- Define success metrics before implementation, not retroactively after deployment
- Expect realistic 40-70% ticket reduction, not hyperbolic 100% elimination promises
- Use Rezolve DeskIQ for upfront ROI visibility through historical data analysis
- Recognize that 99.9% of tickets are mischaracterized in existing systems
- Shift from backward-looking metrics to predictive metrics enabling real-time intervention
- Treat AI as a living system requiring continuous measurement and improvement
Conclusion
As Rob jokes at the end, "May your tickets be few, your MTTR short, and your dashboards full of those lovely little green arrows that are pointing up." But the real message? AI success isn't about trusting the technology—it's about proving it with data before, during, and after deployment.
Ready to see your exact automation potential?
Request a Rezolve DeskIQ analysis and discover your ROI roadmap before committing to implementation.
FAQs
Q1: What is Rezolve DeskIQ?
A: Rezolve DeskIQ analyzes historical service desk data to identify automation opportunities before implementation, providing exact ROI visibility upfront.
Q2: How quickly can we see ROI?
A: Early signals within 2 weeks, measurable impact in months 1-3, with 40-70% ticket reduction over 6-12 months.
Q3: Why not promise 100% ticket elimination?
A: Organizations have unique contexts. Realistic 40-70% ranges account for varying systems, industries, and technical debt rather than making hyperbolic claims.
Q4: Can Rezolve.ai integrate with our existing ITSM?
A: Yes. Rezolve.ai seamlessly integrates with ServiceNow, Jira, Zendesk, and other ITSM platforms while working across Teams, Slack, email, phone, web, and mobile. Rezolve.ai delivers instant IT, HR, and shared services support through Agentic Sidekick 3.0, achieving around 70% ticket auto-resolution. Notably, the platform integrates with existing ITSM tools or functions as standalone AITSM, working across Microsoft Teams, Slack, email, phone, web, and mobile.
.webp)
.jpg)
.png)
.png)