How TechWolf built a job architecture using its own AI agent

[ TL;DR]

TechWolf grew fast—and fast growth meant new challenges: clarity on roles, internal mobility, and fair pay. Instead of spending months manually mapping a job architecture, we used our own AI Agent to build one in just two weeks.

What we did:

• Identified 24 unique roles across manager and IC tracks
  
  
• Used AI to define skills and job levels
  
  
• Partnered with team leads for validation
  
  
• Built five-level skill frameworks per role
  
  
• Delivered everything in a format ready for action
  
  

The result: Clear expectations, better development conversations, and a system that updates with the business—thanks to AI.

Why TechWolf needed a job architecture

By early 2025, TechWolf had crossed the 100-employee mark. We had talented people in 10 departments, but no standard way to describe roles, guide promotions, or ensure fair compensation.

We knew what we needed:

• A clear structure for growth
  
  
• Consistent expectations for every role
  
  
• A fair and future-proof compensation model
  
  

Our VP of People was staring down a challenge many HR teams face—traditional job architectures are slow to build and quick to go out of date. Meanwhile, our product team had just launched a beta version of the TechWolf Agent that helps customers do exactly this. So we decided to become our own beta customer.

‍

How we built it — In just 5 steps

1. Identified the roles

We started with roles that had multiple employees, which gave us 24 unique jobs across IC and manager tracks. Each role would have five levels—providing clarity for both progression and pay.

2. Fed the data to our agent

We combined internal documents with market labor data. Then, the TechWolf Agent inferred 5–10 core skills for each role.

3. Validated with leaders

Next, team leads reviewed and refined the skill lists. We kept only the most important ones.

4. Generated job grading

Our AI mapped out five job levels from most junior to senior for each selected skill, helping define what “growth” looks like in every role.

5. Finalized with human input

Using the AI-generated draft, each department head fine-tuned their job architecture. Everything was delivered in a structured Excel doc that teams could use immediately.

‍

Our job architecture is key in TechWolf's scaling journey. Being able to use our own AI Agent not only helped us get to a great result -- it got us there in just two weeks.
Jeroen, CTO TechWolf

‍

The Result

• Week 1: Skills inferred and validated
  
  
• Week 2: Frameworks finalized and rolled out
  
  

That’s it. In a matter of weeks, every TechWolf team member had a clear view of:

• What’s expected of them
  
  
• How they can grow
  
  
• How their role impacts compensation
  
  

And the best part? As the business evolves, our Agent keeps the architecture up to date. Changes take hours, not months.

‍

I’ve done job architectures before. They’re usually long, frustrating, and the output gets outdated quickly. With TechWolf, it took two weeks. You’re not just getting speed—you’re in control. The AI makes suggestions, you make the decisions. Plus, with real-time market data, we now have a future-proof, living framework. This is what job architecture will look like 10 years from now. We just did it today.
Giovanni Lotti, VP People at TechWolf

‍

A closer look: Engineering framework example

To show you what this looks like in practice, here’s a snapshot from our engineering department—the largest in the company.
‍

Each role includes:

• A list of 5–10 core skills
  
  
• Job grading from 1 (junior) to 5 (senior)
  
  
• Clear distinctions between IC and manager tracks
  
  

This framework isn’t just for HR. Engineers now use it in performance reviews, career development conversations, and team planning.

‍

‍

Ready to Try It Yourself?

‍

Want to see how your org can move from “outdated and unclear” to “living and transparent”?
Join our early adopter program and build your job architecture with AI.