Building a Public Perks Calculator at Doist

What they built

Nadia and a small cross-functional squad at Doist built a public-facing Perks Calculator - a custom, open-source web tool that lets current employees and candidates:

• See every perk Doist offers, what it covers, and its exact monetary value
• Input a salary figure to calculate total annual compensation (salary + perks)
• Understand each perk category: Starter Pack, Booster (work-oriented), Wellness, and Connection budgets
• Access the tool via job descriptions, candidate email templates, and the company handbook

A Kit Assist feature (AI-powered Q&A bot) is in development to let anyone ask questions about perks directly within the tool, async-friendly and available 24/7.

Why they built it

• Volume of repetitive queries - the People and Finance teams (3 people each) were fielding daily/monthly questions about perk eligibility, approval status, and remaining balances
• No single source of truth - perk information was scattered across multiple handbook pages
• Low perk utilisation - employees weren't spending their full allowances, partly due to confusion about what was covered
• Candidate transparency - wanted prospective hires to understand the full value of the total package, not just salary
• Cost visibility - the build process itself revealed the true annual cost of perks per employee, useful for forecasting as a bootstrapped, profitable company
• Async-first culture - with ~100 people across 40 countries and minimal meetings, self-serve information is essential

Tools used

• GitHub - custom repository hosting the tool; the team had to learn pull requests as non-engineers
• Claude Code - used to fill engineering skill gaps during the build
• Outline - company handbook where the tool is linked; identified as a searchability gap to fix in the next iteration
• Ramp - expense management platform used alongside the calculator for approvals
• (Kit Assist - AI Q&A layer, in development at time of recording)

How you could build this yourself

1. Audit your current perks documentation - identify every place perk information lives and note where employees get confused or ask repetitive questions.
2. Define your perk categories clearly - separate budgets by purpose (e.g. work setup, wellness, connection) before building; Doist discovered mid-build that combining work and wellness into one budget was limiting both.
3. Calculate the true annual value of each perk - include full utilisation in your forecast, not a discounted percentage; this surfaces real cost-per-employee data.
4. Assemble a small cross-functional squad - include at least one person outside the People team; Doist benefited significantly from a CX team member with developing engineering skills.
5. Set up a GitHub repository - use Claude Code or a similar AI coding assistant to handle the engineering where your team has gaps; expect to learn basic pull request workflows.
6. Build a salary input + total comp calculator - even a simple field where someone enters their monthly salary and sees perks added on top dramatically increases perceived value.
7. Connect it to your handbook as the SSOT (Single source of truth) - link from job descriptions, candidate email templates, and your careers page; flag searchability between your handbook tool and the GitHub repo as a known friction point to resolve.
8. Move perk approval queries to a public channel — reduces DM volume, creates a visible record of what gets approved, and helps others learn by example.
9. Plan for an AI Q&A layer - once the core tool is stable, add a bot that can answer eligibility questions async; log the questions it receives to inform future iterations.
10. Treat it as evergreen, not a one-time build - budget values, perk categories, and examples will change; design for easy updates from day one.

Source

Built by Nadia Vatalidis, Head of People at Doist. Presented as part of an Open Org "Shipped by HR" session.