With Serval, Mercor onboards thousands of contractors in minutes, automates over 50% of IT tickets, and enables teams to build automations with zero coding, all while operating 24/7 for a global workforce.
Dana Stocking
Head of IT, Mercor
Artemas Radik
Head of Team Platform, Mercor
Derek Shimozawa
Product Manager, Payments, Mercor
Dana Stocking
Head of IT, Mercor
Company name
Mercor
Industry
AI Expert Training Data
Company size
~250 FTE, 150 Contractors
30,000+ Expert Contractors
Mercor is a rapidly scaling AI company that partners with leading labs and enterprises to train frontier models using human expertise. Mercor internally supports 30,000+ external experts, 250 employees and 150 contractors. Their unique business model uses human expert training data to improve frontier models. As a result, they’re prolifically hiring these experts as their business grows, leading to ballooning manual onboarding tasks.
Dana Stocking joined Mercor as Head of IT five months ago, stepping into their fast-moving environment where employee support requests were handled exclusively through a Slack channel without a formal ticketing system. Among these requests were daily requests to onboard Mercor “experts” onto the proper Mercor systems - a highly repetitive task that required a number of slow, manual steps.
Other teams across Mercor were facing similar scaling challenges. For example, Artemas Radik (Head of Team Platform) was managing an increasingly high volume of database-access requests that required structured approvals. Derek Shimozawa leads payments product management, but without a reliable system to log and route contractor payment issues, higher volume made intake and triage a headache for his team. As Mercor grew, these manual processes increasingly became a bottleneck to scaling efficiently.
As Mercor scaled globally, their operational challenges multiplied across teams:
No proper ticketing infrastructure: The IT team was managing support exclusively through a Slack channel, leading to requests falling through the cracks without a systematic way to track or resolve issues.
Time zone challenges: With employees and contractors spread internationally, IT needed to provide support 24/7—impossible with a small team working local hours.
Manual contractor onboarding: The team was manually creating Google accounts for contractors in a separate domain, processing dozens of requests weekly without automation in place.
Manual operational workflows for database access: Engineers needed database access to answer questions like “Was this person paid?”, “Why is this number wrong?”, or “What broke?” but had to request access through database administrators (DBAs) manually to provision database access. These DBAs who manually checked permissions and ran the necessary queries. As Artemas described it: “You’d have to open a tool, check permissions, run the request, and then report back the results.”
When Dana joined Mercor, he had a unique opportunity: the company was scaling rapidly, and its internal IT systems were still early enough that he could rebuild core workflows from the ground up and choose tools designed for AI-native automation. He chose Serval because of its powerful automation capabilities.
After Dana rolled out Serval, teams across Mercor quickly saw the value and began implementing it within their own functions.
Today, Serval powers workflows across seven teams at Mercor:
IT: Faster, automated ticket resolution and 24/7 global support
Infrastructure: Automated database queries with approval-based workflows
Payments: Quicker ticket triage and easily-tracked payment investigations
Engineering: Self-service database access and automated categorization
Security: Stronger compliance through end-to-end platform auditability
HR: Smoother employee and contractor onboarding
Office Support: Higher deflection on common office inquiries via FAQ and policy-linked responses
What convinced Dana wasn't just Serval's AI capabilities, it was transparency. "As I was building the workflows in natural language, I could then go and see exactly what the code was in a little sidebar to make sure it was doing exactly what I wanted," he explained.
Derek's first impression centered on speed: "I was able to implement my first workflow in a couple of hours, demo that to the team, and within the day, we were already using it to automate triaging for tickets."
For Artemas, the value was strategic: it made the buy-vs.-build decision easy. Mercor could get the equivalent of an internal support agent without dedicating headcount or months of platform work to build automations. Serval doesn’t just execute workflows—it also suggests and helps generate the automations a team needs, drastically reducing the time from idea to working system. He described Serval as like having a “coworker” sitting next to you to support your team. Because it’s so easy to build automations, the setup effort is “near-zero from the start.” In his view, for most companies, “you might as well just buy [Serval] now.”
"Serval can grow with your team and make suggestions about how to handle those issues. The cost of having, effectively, a coworker to sit next to you and help support your team is near zero in setup from the start. Might as well just use Serval now"
Artemas Radik
Mercor's business model relies on a marketplace of 30,000+ external contractors who write training data. Historically, onboarding these contractors was a manual, high-volume process - Dana manually processed hundreds of contractor onboarding and access requests each week.
Now, it’s much faster to onboard contractors, because Mercor has automated onboarding, and expert managers can trigger the onboarding themselves - removing IT as the bottleneck. First, Mercor employees can initiate Mercor expert onboarding in Slack. Then, Serval checks that the requester has the right permissions to onboard a Mercor expert, creates a Google Workspace account (email) for the contractor, and applies a role-based access template for the core tools they’ll use in their job. As a result, to date, Mercor has onboarded 4,000+ contractors with this automation.
"We were able to onboard ~4,000 of our Mercor experts and it only took minutes"
Dana Stocking
Head of IT, Mercor
With team members and contractors distributed internationally, Mercor needed IT support that never sleeps. Serval now provides AI-powered assistance around the clock, responding immediately even when the team is offline.
“Using an AI bot to answer all of those questions for people allows me to answer things while we’re sleeping, get people way quicker responses than we would if we were waiting for a human to interact,” Dana said.
Engineers on revenue-driving teams needed database access to ship features and investigate issues, but manual approval processes created bottlenecks. Artemas automated the workflow while maintaining security controls. Engineers post SQL queries in Slack, Serval validates them, then routes queries to 14 database administrators who approve with a single click. What used to require multiple manual steps now happens in seconds—with complete audit trails of who requested what and why.
"We've been able to basically remove the need for any humans to be in the process except for that critical decision-making step"
Artemas Radik
What's Next
Payments automation
Derek's team is exploring using Serval to complete full payment investigations: gathering data from multiple sources, diagnosing issues, and even preparing automated payouts with engineer approval as the final step.
Advanced database operations
The infrastructure team plans to automate more complex database workflows and expand self-service access for engineers.
Broader IT automation
Dana is continuing to identify high-volume ticket types to automate, working toward even higher zero-touch resolution rates
More teams are experiencing how easy it is to set up Serval, because Serval can suggest new automations based on resolved tickets. Artemas’s team is quickly using the Suggestions feature more often to quickly build automations into their already robust set of Serval automations. That low friction translates into fast iteration in practice.
Derek also saw the impact firsthand: “I was able to write my first Serval workflow in a couple of hours, demo that to the team, get feedback really quickly, tweak it, and within the same day it was automated for us.”
What starts as a simple suggestion quickly becomes a working automation, embedded into how teams operate day to day. With Serval, automation isn’t a long-term project or a specialized skill – it’s a natural next step after solving a problem once. The result is less manual work, faster improvement cycles, and a system that quietly gets better as teams use it.