CASE STUDY: Solar Edge

A solar manufacturer with installers calling from rooftops is rebuilding technical support on one platform, in a deliberate order: unify first, then let AI carry the troubleshooting.

SolarEdge is a global manufacturer of solar inverters, power optimizers, and energy storage, headquartered in Israel with support operations centered in India and a customer base spanning every major solar market globally. Its technical support organization serves installers and homeowners whose systems, sites, and skill levels are never quite the same twice. In 2025 the customer experience team set out to find a better way to run quality assurance and ended up choosing a single platform for workforce management, voice, chat, quality, and AI. The United States moved to NiCE CXone in one day with no disruption to customers. Global rollout is under way, and NiCE AI Agents (native Cognigy) is the centerpiece of the next phase: a support experience where the caller describes the problem and the system does the work.

AI means everything, but AI also means nothing. We want people to call in and have a simple experience. What do you need? And we solve the problem for them.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

01 Before

Technical Support for an Industry Where Every Site Is Different.

SolarEdge builds the electronics that turn sunlight into usable power. Its customers are installers who put those systems on roofs and the homeowners who live under them. Most of the contact center’s work is business to business, roughly 70 percent, and almost all of it is technical support. Nobody calls to say the system is working.

The work is seasonal and unforgiving. Summer brings the highest production expectations and the most eyes on the equipment, and it brings installers calling from the job site with a customer waiting below.

When installers are calling us, they’re on the roof. It’s 110 on the ground, it’s 130 on the roof, and they need an answer and they need it quick.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

The variety is the hard part. SolarEdge’s products may be identical from one install to the next, but the grid connection, the cabling, the site, and the installer’s experience are not. One caller arrives with every measurement already taken and needs a two-minute answer. The next installed the system alone, has no voltage, and needs to be walked through from the beginning.

Over the years the support operation had assembled the tools it needed one at a time. Chat ran on one platform, voice on another, workforce management on a third, robotic process automation on a fourth. Automated chat could create a case for follow-up but could not resolve anything in the moment. Quality assurance was manual, which meant a handful of reviewed interactions per agent per month, too few to describe the experience customers were actually having. Each tool had done its job. Together they had become an operation the CX team could see only in pieces.

02 Desire to change

It Started With Quality Assurance. It Did Not End There.

The search began narrowly. Michael McCleary, Global Director of Customer Experience and AI Implementation, asked his team to find an AI-based quality assurance tool that could review every interaction rather than a sample. They surveyed roughly ten vendors and sat through the demos. Most solved one problem well and stopped there.

NiCE was on the list for quality management. The conversation kept opening new doors. Each time the team described a fix they needed, the answer led to another capability that was already on the same platform. SolarEdge’s contracts for voice and workforce management were approaching renewal, and the team began asking a bigger question than the one it had started with: what if the point solutions were the problem?

I told them three times, don’t sell me NiCE. And they said, I’m not going to sell it to you. I’m going to show you outcomes.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

The case came together in two parts. A return-on-investment comparison laid out what ten discrete changes would mean for cost and customer experience. And the broader logic was simpler still: ten point solutions will not beat one integrated platform on total cost, on visibility, or on the time it takes to find the right person when something changes.

What set this decision apart was who was at the table. Previous tools had been selected with limited input from the support team itself. This time, the practitioners led the evaluation. The people who take the calls and manage the queues tested the platform against their own processes and brought the recommendation to leadership.

Trust the people who are doing the work and who understand the processes. If we’re saying this is the solution because the one we’ve had hasn’t worked, there needs to be some trust and respect there.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

The hesitation, when it came, was the honest kind. The solution mapped so closely to what the team had needed for years that it sounded too good to be true. Demos are easy. What settled it was working with the tools directly and with a build team that clearly knew what it was doing.

03 NiCE solution

Unify First. Then Let the AI Do the Troubleshooting.

SolarEdge sequenced the rollout on purpose. Voice, chat, and workforce management would move first, in every region, because that foundation had contractual deadlines and because everything else would sit on top of it. Quality management would follow. NiCE AI Agents (native Cognigy), NiCE Copilot, and NiCE RPA would come once the platform was stable and global.

One day, three systems, no surprises

Working with implementation partner TTEC, SolarEdge moved its United States operation off its previous voice, chat, and workforce management platforms and onto NiCE CXone on a single day. The team expected problems. There were none, apart from one language-queue setting on SolarEdge’s side.

I looked at our WFM manager and said, are you sure? There’s got to be something routing wrong. We checked everything. Then I called the global VP and said, I need you to understand this went as smoothly as I told you it was going to.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

Customers noticed nothing, which was the goal. Agents log in, go ready, and take the call. Workforce management, which had been a point of friction on the previous platform, became easier to run: changing queues and assigning roles now takes a fraction of the effort. Real-time analytics that used to require five windows and a business intelligence layer stitched from ten sources now live in one place.

The design principle: one letter

Everything SolarEdge is building on top of the platform is shaped by a constraint most contact centers never face. In solar, a troubleshooting step given incorrectly can kill someone.

If they pull the AC lines out rather than the DC lines, that carries a serious safety risk. It’s one letter. So, we need a lot of safety guardrails.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

That constraint sets the architecture. The AI Agent handles intake: it gathers the serial number, the error code, the measurements, and the photos, and it works through the troubleshooting steps that are safe to automate. The moment a step carries safety risk, the interaction hands to a human agent with everything already attached. The agent does not re-ask for the serial number or the measurements. They read what the customer already provided and pick up at the decision that requires judgment.

The efficiency follows from the safety design rather than competing with it. Today an agent spends ten to twelve minutes manually creating a case after a troubleshooting call. With the AI Agent collecting the inputs and NiCE RPA creating the case and processing the return, SolarEdge expects to remove ten to fifteen minutes from every call that involves troubleshooting, and to resolve the routine ones without an agent at all.

04 Results

A Stable Foundation and a Clear Line of Sight

SolarEdge is the first to say it is early. The capabilities that will move the operational metrics are not yet in production, and the company is more interested in getting them right than in getting them fast. What can be said now is narrower and, for the team running the operation, more important.

The migration held. The largest region moved in a day with no customer impact and no agent disruption. Workforce management went from a source of friction to a tool supervisors like using. And for the first time the CX team can see the whole operation, across channels and queues, in a single view rather than reconstructing it from disconnected reports.

The accountability changed too. When something goes wrong on a unified platform, the team knows what is down, where, and why, and there is one partner responsible for it. On the previous stack, a dropped chat could produce a week of finger-pointing between vendors before anyone owned the fix.

It’s nice to say, look, if something’s down, something’s down. We know why and we know where. Before, it was pulling teeth to find who was responsible for what.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

The most durable result may be cultural. The practitioners who chose the platform own it. That is a different footing for the AI phase than the one SolarEdge’s earlier tools started from.

05 Future

A Three-Minute Call

The target is specific. An installer calls, describes the problem, and answers a few questions. The AI Agent runs the troubleshooting it is allowed to run. If the issue resolves, it resolves. If a replacement is needed, the case is created, the return is processed, and the part ships, without a human touching it. If a safety-critical step arrives, a person is already there with the full picture. Three minutes, and the installer is back to work.

Getting there means finishing the global rollout of voice, chat, and workforce management, standing up automated quality management, and then bringing NiCE AI Agents, NiCE Copilot, and NiCE RPA live in combination. SolarEdge has told its customers publicly that good things are coming. Those things are the AI phase.

I don’t believe AI should just replace people. AI should be a tool to help people. If you build it out right and you guardrail it, it is a full solution. Not for everything. But for a lot.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge

The ambition is larger than SolarEdge’s own queue. Technical support for a product where every site is unique has been treated as something automation cannot reach. If the design holds, SolarEdge will have shown otherwise, and shown it in an industry where getting it wrong is not an option.

If the answer is ‘that’s just how it’s always been,’ you’re doing it wrong. Start over. Understand the core reason you’re looking for a new tool or a new partner. Take the advice of the people who deal with the problem every day. And really target what you’re looking for in AI, what it means to your customers, and what it means to your company.

Michael McCleary

Global Director of Customer Experience and AI Implementation, SolarEdge