The economics of work, human and AI

What every process costs, in people and in AI.

WorkGraph joins human capacity and AI compute on the same processes, and shows whether it works.

workgraph / process economics · annual

Demo data

Process	Human · AI	Volume	Cost / unit	Confidence
Lead management Augmented	€720k €84k	62,000	€15.16	Medium confidence
Invoice processing Experimental	€610k €6k	48,000	€14.56	High confidence
Account research Assistive	€420k €58k	9,400	€53.40	Medium confidence
Month-end close No AI	€380k €3k	12	€34.4k	High confidence

Invoice processing costs €610k in people for €6k of AI. Lead management gets the largest AI budget. Is that the right call?

01The problem

You can see what AI costs. Not what it’s worth.

Dashboards report seats and tokens. None say which process the compute served, or whether it improved.

What you know

Demo data

AI spend

≠

What you don’t

AI value

02Three products

Where does human capacity go?

Processes, activities, roles. FTE, loaded cost, unit cost.

Where does AI compute go?

Assistants, applications, agents. Cost attributed to processes.

What does it return?

The operating statement of every process, tracked to the P&L.

We analyse work, not people. Aggregated by process and role, nothing installed on devices.Privacy

03Use cases

For leadership teams in companies of 250 to a few thousand people, investing in AI and asked to show what it returns.

CEO · COO
“Where does the team’s time go?”
A validated work graph with FTE, cost and unit cost per process.
CFO
“What are we getting back for AI spend?”
Spend attributed to processes, value tracked to the P&L.
CIO · Chief AI Officer
“Which compute serves which process?”
Tokens and cost attributed, with a stated confidence level.
PE operating partner
“Where is the AI upside, and how much reaches EBITDA?”
A comparable diagnostic of one process, repeatable across a portfolio.

Test it on one process.

A cross-functional process or a business unit. Read-only, nothing installed on devices.