Summary: “NordLayer AI Scout is a browser extension and desktop sensor architecture that collects data from AI-specific tools in one place, and it’s a really useful tool.”
Oxylabs is a global web intelligence platform and a unicorn valued at $3.6 billion With nearly 800 employees, the company helps businesses collect large amounts of public web data Its products include proxy networks, web scraping tools, and ready-made datasets used for work such as price tracking and market research.
As more teams began using AI tools, Oxylabs needed a way to see how employees were actually using them.
We spoke with Martynas Jočys, Data Engineering Tech Lead, and Matas Šeškauskas, Senior Risk Manager both of whom work at Oxylabs. We discussed how NordLayer AI Scout helped the company gain visibility into AI usage, spend, and prompts without building a custom solution in-house.
Like most companies, Oxylabs first relied on the analytics pages and APIs of each AI tool. But that approach quickly hit its limits:
The team tried to fix it themselves. They built a unified dashboard for Cursor and Claude Code that pulled data from the tools’ APIs and combined it with HiBob employee data. But keeping its accuracy was an uphill battle. Constant API updates, pricing changes, discounts, and caching caused the data to drift.
Martynas Jočys recalls:
“Each tool counted usage differently and identified users in different ways. We needed to bring that data together to better understand which AI tools people used and whether their subscriptions suited their work.”
Meanwhile, the board and upper management wanted to track AI spend The data team, on the other hand, wanted to know if AI tools were actually making work more productive or just faster.
You can’t secure what you can’t see.
Get visibility and control into AI activity within your network.
From their perspective, the stakes were even higher. According to Matas Šeškauskas, Senior Risk Manager, the biggest shadow AI risks were data leakage, loss of intellectual property, regulatory noncompliance, and third-party vulnerabilities especially with machine-to-machine access and high-frequency querying in the mix.
A couple of scenarios worried the team most. The first was machine-to-machine access and high-frequency querying where coding agents and automated workflows talk to AI tools with no human watching what goes in or comes out. The second concern was security-related tokens—credentials or API keys—that could end up in prompts. In an engineering-heavy company, both scenarios were real possibilities.
“We had no AI visibility solution in place, and we were determined to find one. The NordLayer AI Scout team offered us a test run, and we took them up on it.”
With no AI visibility solution and an in-house build already on the drawing board, Oxylabs decided to test NordLayer AI Scout. The company-wide rollout, including preparation, took less than 2 weeks
Today, NordLayer AI Scout consolidate Oxylabs’ AI usage insights from apps and browsers into one place. Here’s what the team gained:
Oxylabs was already planning its own tracking tool: a narrowly scoped, CLI-based solution that would have taken significant engineering time to build, not to mention ongoing maintenance. Instead, the company rolled out NordLayer AI Scout across 469 devices in under 2 weeks According to Martynas Jočys, the timing was perfect.
“NordLayer AI Scout was already doing what we were about to build, but on a much broader scale, covering both app-specific and browser data. It was faster for us to just try it out than to build a similar solution ourselves.”
For Oxylabs, the standout feature of NordLayer AI Scout was its ability to provide insights that help the entire organization use AI more effectively Martynas Jočys says that it opens up ways to boost employee productivity:
“Our own solution would have focused purely on usage—tokens, models, spend. The AI adoption findings let us optimize how AI is used across the company. We can already spot cases where we could help employees use AI tools more effectively.”
With session-level analytics, the team can see how many messages a session takes, which models are used, and whether expensive models are being used for simple prompts. All of this means smarter AI usage company-wide.
NordLayer AI Scout has built-in data loss prevention (DLP) capabilities, which help prevent leaks of sensitive information in an AI tool This is where Oxylabs gets the most value from the solution. The risk team doesn’t comb through raw AI activity manually. Instead, NordLayer AI Scout’s detection engine flags the findings for them. Custom filters surface what needs attention.
“What made NordLayer AI Scout stand out for us is the combination of DLP insights and faster triage.”
Next, Oxylabs plans to connect alerts to a Slack channel, so the team can move from conducting periodic reviews to responding to findings as they come in.
Oxylabs wanted to manage AI subscriptions in a smarter way, and NordLayer AI Scout made it possible.
Every month, the data engineering team adjusts Claude Team seats across 3 tiers—Standard, Premium, and Individual pro licenses. The goal is to match each employee’s license to how they actually use the tool. Heavy users get the higher limits they need. Light users don’t occupy expensive seats that they never use.
“From a pricing perspective, there are different tiers. We initially assigned licenses based on API usage, but now we adjust them based on actual usage. That’s exactly the kind of decision this visibility supports.”
For Martynas Jočys, AI usage and spend visibility is the first non-negotiable step before making any AI governance decisions.
“Once you have data, that’s when real decision-making starts. Otherwise, you’re just blindly guessing or copying decisions from other companies. If you want to build your own AI governance strategy, you need visibility first.”
Today, Oxylabs is running NordLayer AI Scout in monitoring mode It uses security review features and is working toward integrating Slack alerts. The company collects the data it needs before setting policies on what’s allowed, what’s blocked, and how subscriptions should be assigned.
NordLayer AI Scout met Oxylabs’ needs because it delivered broader visibility than anything the team planned to build internally:
Martynas Jočys summarized why NordLayer AI Scout is a useful tool:
“For me, NordLayer AI Scout is a browser extension and desktop sensor architecture that collects data from AI-specific tools in one place.”
Martynas Jočys and Matas Šeškauskas offered advice to engineering-heavy companies on how to manage AI usage and risks:
Oxylabs chose NordLayer AI Scout to address the questions every company adopting AI eventually faces: who is using what, how much does it cost, and is it actually paying off?
The company replaced fragmented dashboards from different AI tools with a data-driven foundation for smarter AI decisions
Want the same visibility into how your teams use AI? Talk to our sales team to see what NordLayer AI Scout can do for you.
Joanna Krysińska
Senior Cybersecurity Copywriter
Joanna writes about zero trust, network security, access control, and threat prevention, but her cup of tea is compliance and how regulations shape security controls. Her work also includes the dark web topics and the methods cybercriminals use to target companies, such as social engineering or ransomware attacks, for example.
Subscribe to our blog updates for in-depth perspectives on cybersecurity.
