cost-of-work-index
Per-unit cost of 13 back-office tasks done by a person vs by software, in the US, Greece and Ukraine, with an official statistic behind every hourly wage. Bundled data, works offline.
README
The Cost of Work Index
Per-unit human labour cost vs. AI-employee cost across 13 back-office tasks in three markets (United States, Greece, Ukraine). A small, fully-sourced dataset published under CC BY 4.0.
What this measures
For each task ("enter a supplier invoice", "chase an overdue invoice", "screen one candidate") the dataset records two numbers:
- Human cost per unit: the published hourly labour cost for that market, divided by the human minutes the task takes.
- AI cost per unit: the advertised per-unit price of the matching Nika AI employee.
The multiple is the ratio of the two. A multiple above 1 means the AI employee is cheaper per unit; below 1 means a person is cheaper.
What's in the data
- 3 markets: United States (USD), Greece (EUR), Ukraine (UAH)
- 13 tasks across back-office functions (bookkeeping, scheduling, support, collections, quoting, candidate screening)
- 39 human-vs-AI cost comparisons (13 tasks × 3 markets)
- 3 hourly labour cost figures, one per market
The two kinds of number
The dataset is explicit about which figures are statistics and which are estimates:
- Cited: carries a publisher, a resolvable URL, a reference period, and a retrieval date. The three hourly labour costs are cited from BLS, Eurostat, and Derzhstat.
- Estimate: the human-minutes-per-unit figures are Nika's own estimates, labelled as such and documented with their written basis per task.
Sources
| # | Publisher | Figure | Period |
|---|---|---|---|
| 1 | U.S. Bureau of Labor Statistics | Bookkeeping median hourly wage ($23.66) | May 2024 |
| 2 | Eurostat | Greek whole-economy hourly labour cost (€16.70) | 2024 |
| 3 | Derzhstat (Ukraine) | Average monthly wage (UAH 196.78/h) | June 2026 |
Caveats (carried in the dataset itself)
humanMinutesPerUnitis a Nika estimate, not an official statistic. Its basis is stated per task inhumanMinutesBasis.- Hourly labour costs are not like-for-like across markets: the US and Ukraine figures are gross wage only, while the Greek figure is fully loaded and already includes employer contributions.
- For some tasks in some markets a person is cheaper per unit than the AI
employee. Those rows are real and are not filtered out. Check
multiple. - Employees marked
waitlistare not yet running. Their per-unit price is the advertised price, not a measured one.
License
CC BY 4.0: free to use and cite with attribution.
Citation
Nika (hirenika.com). The Cost of Work Index, version 2026.3, 2026-08-04. https://hirenika.com/cost-of-work
Files
data/cost-of-work-2026.3.json: full dataset (tasks, markets, sources, caveats)data/cost-of-work-2026.3.csv: tabular export
Resources
- The canonical page, with full methodology, per-row bases, and live figures: https://hirenika.com/cost-of-work
- Machine-readable REST API (no key, CORS open): https://hirenika.com/api/v1/cost-of-work
- Nika, the AI employees behind the per-unit pricing: https://hirenika.com
MCP server
The index ships as an MCP server so agents can query it directly:
npx -y nika-cost-of-work-mcp
Five tools: list all tasks, get per-unit cost for a task and market, estimate
monthly/annual cost from a volume, list the official sources, and export the
whole index as JSON. The dataset is bundled, so the server answers offline;
no API key is required. The server source lives in this repository
(src/, Dockerfile), and the npm package is published from the hirenika
repository where the data snapshot is generated, never hand-edited.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.