TL;DR
- Wiz’s GTM Systems team was spending hours just getting context before they could resolve Salesforce cases, slowing down fixes and increasing risk.
- Sweep gave the team a live map of their org, helping them trace dependencies, understand downstream impact, and move far faster with fewer bugs.
- What started as metadata visibility expanded into broader value, including deduplication, self-serve documentation, reduced technical debt, and stronger AI readiness through better system context.
*****
The Story
When Kelly Skelton, Salesforce Administrator on Wiz’s GTM Systems team, describes what it felt like working inside their Salesforce org, he doesn’t start with how much they had been scaling or how overwhelming the complexity had become. He starts with time.
“When a case comes in,” he said, “you could spend three hours on it — two and a half just getting context, and then half an hour actually fixing the problem.”
As it turns out, that ratio is the norm across systems teams throughout the country (evidenced here by our new State of Enterprise Systems report). And that ratio has a cost.
In time, yes, but also in how much work even the strongest of teams can realistically take on, how fast their systems can adapt to an ever-changing reality, and how much risk accumulates with every change.
By the time they started seriously evaluating Sweep, Wiz had already compiled a complex environment due to their breakneck speed of growth. Thousands of users, hundreds of fields. It was enough to hit limits on core objects. Custom objects for everything from compensation requests to post-sale adjustments, all stitched together across teams.
“You’d get a case from another team, and you needed context to resolve something simple,” Kelly said. “But without that context, you’re just digging. And digging.”
Visibility breaks with growth
Not unlike other high-growth companies, Wiz’s system had become complex and difficult to maintain. The team couldn’t see enough of the system at once to move confidently. And when you can’t see, speed becomes dangerous.
“We were constantly producing bugs,” Kelly said. “Breaking stuff, just trying to keep up.”
At the time, the team was small (mostly admins, very few developers) supporting a system that had already outgrown the ways they could reason about it.
Sweep discovered
Sweep entered the picture almost incidentally. Kelly saw it at Texas Dreamin’, then again at Midwest Dreamin’. We struck him: “I hadn’t seen anything that could grab and surface metadata in such a strong way,” he said.
He brought it back to the team as a way to move faster, to understand the system more clearly. Deduplication wasn’t even part of the initial conversation. At that point, they were evaluating Sweep purely for its ability to help them navigate their own org — trace dependencies, understand downstream impact, and finally get a clear view of how things actually worked.
The math changes
It was only later, once they were deeper into the product, that something clicked.
They were already paying for a deduplication tool. And Sweep could replace it.
“It wasn’t really a feature gap we were solving,” Kelly said. “It was more like… we were already looking at Sweep for metadata and documentation, and then realized it could replace something else we were using.”
That changed the conversation internally.
Instead of justifying a new tool, they could justify consolidating one they already had while increasing the output of the team using it.
“We could say, we’re already spending this budget… and now we get that plus new capabilities that are going to speed us up,” said Kelly.
Complexity, dedupe, and speed
“Dedupe is usually difficult because of how complex everything is,” said Lily Zhang, GTM Systems Analyst at Wiz.
“There are so many hands involved, things change all the time. It’s not just volume, it’s how everything connects.”
Lily explained that Wiz needed more than a tool that could merge records: they needed one that could operate inside a system where every change had downstream effects and where those effects weren’t always visible upfront.
When they started using Sweep, the first thing that showed up was exactly what they expected.
Four months in on the tool:
“The time savings we thought we would get, we are indeed getting,” Lily said.
Then they just… trusted the map
Salesforce has ways of showing you where something is used, but only in certain cases, and never completely. “With standard fields, you literally cannot use ‘Where is this used,’” Kelly said. “And we rely on those everywhere.”
Sweep changed that.
“Not having that visibility… that was a major unlock for us,” he said.
And more importantly, they trusted it. “The metadata search is 100% accurate,” he said. “We’ve never found a case where it was incorrect.”
And once they trusted it, everything changed
Once they trusted the output, the way they worked started to shift.
“You’re getting to the answer you need faster,” Kelly said.
“The team wasn’t double-checking every result,” Lily added.
They weren’t baking in the time they previously needed to validate what they were seeing. They were just moving.
And because they understood more before making changes, they broke less.
“We produce far less technical debt and bugs than we did before,” Kelly said.
The questions stopped piling up
At the very same time, something else started happening outside the core admin team.
“I use the documentation all the time,” Lily said. “Sometimes I don’t want to bug someone for something small. I can just build my own understanding.”
That ability to self-serve meant fewer escalations and faster resolution across the board.
“It can be used by literally everybody,” Kelly said.
The agentic layer formed
Over time, the value of Sweep started to look less like a set of features and more like a layer underneath everything they were doing.
“It’s not just that we have this metadata tool,” Kelly said. “The more pieces we layer on, the more valuable it becomes.”
At Wiz, there’s a phrase they use internally when they talk about building with AI: context is queen, and AI is king. (And just like on a chessboard, good luck winning if you don’t have a queen.)
“You can pass fields into an AI model and have it generate something,” Kelly said. “But you don’t have all the context. In Sweep, you have the context. That’s the number one piece.”
What changed most: they could understand what would happen before they made any changes.
Less time digging. More time doing. Less hesitation. More confidence.
In that way, their system’s complexity had become useful.
This wand-logo'd company was magic for this Wiz, you might even say.
