One customer's words changed how an entire support department thinks about its mission. After an interaction with SysAid's team, they said it was the first time a support department had given them a "wow experience."
Asaf Goldstein, Senior Director of Global Customer Care at SysAid, turned that quote into a rallying cry. "That's exactly what I'm using with my entire department," he said. "We need to replicate this. We need to duplicate that experience."
The results speak for themselves. Unhappy customers dropped from 40 per quarter to just 5. Twelve percent of tickets now resolve completely through AI. And critical issues that once lingered get resolved in 15 minutes.
The Shift from Reactive to Proactive
SysAid has provided IT service management software for 25 years, offering ticketing, asset management, license management, and AI tools for IT departments. Goldstein manages a team of around 30 individuals across tier one, tier two, tier three, and technical account management.
The traditional support model is inherently reactive. Customers call when something breaks. They are frustrated. Support tries to save the day.
AI changes that equation.
"The most important thing that support leaders need to understand is that support needs to change from being reactive to being proactive and to drive even revenue and not just be a cost center," Goldstein explained.
With AI analyzing patterns across thousands of customers, SysAid can now identify issues before customers even notice them. In multiple cases, they have contacted customers proactively: "Look, we identified something. We're already on it. Don't worry." Within minutes, the issue was resolved.
The industry is moving rapidly in this direction. According to Salesforce, by 2027, 50 percent of service cases are expected to be resolved by AI, up from about 30 percent in 2025.
The Discord War Room
One example illustrates the power of AI-driven proactive support. SysAid started a community using Discord. When two or three customers mentioned a critical issue, AI automatically identified it, checked internal tools, confirmed the problem existed, and within five minutes had engineering and DevOps assembled in a war room. Fifteen minutes later, the critical issue was resolved.
"The customers were actually happy," Goldstein recalled. "They were down for 15 minutes, but they said like, wow, okay, I'm happy. It was solved so quickly."
This aligns with research from Freshworks showing that first response time for tickets has dropped from over 6 hours to less than 4 minutes with AI-powered support. In some cases, resolution times have been slashed from nearly 32 hours to just 32 minutes.
The 12 Percent Solution
SysAid achieves 12 percent monthly ticket deflection through AI, meaning those issues resolve completely without human intervention. This matches industry benchmarks: Freshworks reports that AI agents now deflect over 45 percent of incoming customer queries in some industries, with retail and travel companies seeing deflection rates above 50 percent.
But the impact extends beyond the deflected tickets.
"My agents can work on actual complicated tickets, on the things that require the human touch, the things that require a remote session or something that is more complicated," Goldstein said. "And we leave the more basic stuff to AI."
When tickets do reach agents, AI provides context: it already tried certain troubleshooting steps with the customer. Agents can skip the basics and start from advanced troubleshooting. They handle fewer tickets, but each ticket receives deeper attention.
The AI Manager Role
When SysAid began implementing AI, they made a critical decision: assign a dedicated project manager whose job was to review every AI interaction, validate answers, and continuously improve the data pool.
"If you don't pave the way in the beginning, AI is going to make a lot of mistakes because in the beginning, it just tries based on the data that you feed him," Goldstein explained. "But if you don't have someone that actually performs the relevant guardrails and validates the data pool answers and even formats it in the way that your organization sounds, because someone is more formal, someone is more going easy with the answers."
This human oversight proved essential. First impressions matter with AI. If a customer gets a wrong answer initially, they will demand a human immediately next time. If they get the right answer, they will try AI again.
The AI manager tracks two primary metrics: quality score (how confident the system is in each answer based on data pool matching) and tickets deflected. Secondary metrics like first response time and customer satisfaction correlate directly to AI success.
The Danger of Over-Automation
Goldstein shared a personal frustration that illustrates the risk of over-automation. He tried to get a simple invoice from an airline company. The AI kept promising it could help but failed repeatedly. It refused to connect him to a live agent for 30 minutes.
"That's exactly when you're over trying with AI instead of understanding, okay, this is where I need to stop and move it to a real person," he said.
Research confirms that 85 percent of consumers say their issues eventually require human intervention. The goal is not to eliminate humans but to route customers appropriately.
"You need to find the right balance that when we see that we can't help a customer using AI, it will direct them to an agent and tell the customer, 'Hey, look, I already gave the agent all of the relevant information,'" Goldstein explained.
What AI Cannot Replace
After 25 years in ITSM, SysAid understands what technology can and cannot do. Some things require human presence.
"There are countless examples that we had where customers asked for a remote session or asked someone to actually jump on a call with them, even travel abroad and be with the customer to help them solve a complex issue," Goldstein said. "Those things can never be replaced with AI."
When customers are unhappy, human empathy becomes essential. Before Goldstein joined SysAid, quarterly surveys showed around 40 unhappy customers out of 300 who left feedback. Now that number is 5.
The approach: every unsatisfied customer gets a personal call and email from a team lead, not an agent. They want to understand what could have been done better. They validate multiple times that everything is solved before closing the ticket.
"AI doesn't understand those small sentiments," Goldstein noted. "When it comes to okay, I'll just close the case and I won't ask if there is anything else. I want to validate there aren't any other issues."
The Personalized Wow Experience
AI enables personalization at scale. When agents engage with customers, they already have comprehensive context.
"We tell them, 'Hey Russ, I know that you're the chief recognition officer in the Business Intelligence Group. I know that you already opened five tickets with us about XYZ. I know that you're using Azure and Slack,'" Goldstein explained. "You already feel confident that I know what I'm talking about. You don't need to explain yourself to me."
This context comes from AI analyzing customer history, platforms, previous interactions, and even job titles. The result is that personalized touch that creates wow experiences without requiring agents to spend time researching before every conversation.
The Future: AI Managers, Not Customer Care Managers
Goldstein sees a fundamental shift coming in how support teams operate.
"At some point, I would say a few years from now, we would only have people that will manage AI instead of managing customers," he predicted. "People that will enhance the AI with more connections and more information and enhance the data pool and the coverage. We will see more jobs like AI manager rather than customer care manager."
For tier one agents doing basic troubleshooting and how-tos, the warning is clear: those roles will likely be automated. Goldstein advises his team to develop deeper technical skills, platform expertise, and capabilities that AI will take longer to replicate.
"If you don't have the ability to understand how AI can help you, then you're in danger of being replaced," he said. "I don't remember life before utilizing AI. And I can say that it definitely helped me get to those results and not replace me."
Principles for AI Implementation
For customer support leaders considering AI implementation, Goldstein offers several principles:
Establish guardrails first. Define what AI can and cannot do. Ensure it cannot share information about different customers. Set clear limits.
Assign a dedicated validator. Someone must review AI interactions, improve answers, and ensure the system sounds like your organization.
Monitor continuously. Do not implement and walk away. Constantly validate that AI performs as expected.
Keep expanding. Identify new tasks AI can handle. Goldstein shared an example of using an AI agent to coordinate a technician visit for a water company issue, never speaking to a human but having an amazing experience.
Remember the goal. It is not about reducing headcount. It is about creating wow experiences that drive loyalty, reduce churn, and increase adoption.
The combination of AI efficiency and human empathy creates something neither can achieve alone. As Goldstein put it: "The good experience comes from the combination of AI with people. And that's how you can create loyalty, how you can create trust."
Enjoying insights from industry leaders? Subscribe to The Winners' Circle podcast on your favorite podcast player and never miss an episode. Listen and subscribe at bintelligence.com/podcast.