AI COST REPORTING

Follow AI spending
to the work behind it

GenieWorks connects usage to the task, person, model, and outcome. Follow an expensive run into its calls and activity, then decide what to improve.

Follow the work

01 · FIND THE WORK BEHIND THE TOTAL

Start with who, when, and what ran

The monthly view shows totals and changes over time. The ticket list makes the spending concrete: which task ran, who started it, which model it used, and its estimated cost.

Here, Maya’s APP-58 specification run stands out. Follow that run before deciding whether the answer is a different model, a different method, or justified investigation.

02 · INSPECT THE INVESTIGATION

Three full reads deserve a closer look

APP-58 used 9.52M tokens in repository investigation, 2.406M in specification, and 1.02M in acceptance review.

Inside the investigation, api-schema.json was read in full three times. A strong prompt can repeat this expensive discovery when the repository does not expose a useful, smaller contract surface.

The trace gives you a cause to investigate, not just a larger total.

APP-58 · Call 01 · Recorded file activity
  1. 1
    09:14:08 · Read api-schema.json

    738 KiB returned

  2. 2
    09:16:42 · Read the same file

    738 KiB returned

  3. 3
    09:19:11 · Read the same file again

    738 KiB returned

03 · FOLLOW THE EVIDENCE

From the monthly picture to the next experiment

The gallery keeps one dataset throughout: overview, APP-58, and its improvement proposal. Ticket rows are a subset of the period total; file bytes do not establish exact per-file tokens or cost.

04 · TEST THE IMPROVEMENT

Make the next comparable run more informative

The proposal is to select the relevant schema objects before reading the API contract. A line limit alone would not bound a single-line JSON file.

Try that change on equivalent work with the same model and scope. Compare input tokens and total usage, then check that the resulting specification still covers the required behavior.

Savings have not been measured in this example. Lower usage alone does not establish a better result.

A SHARED TEAM CAPABILITY

Keep the explanation with the work

The next teammate can inspect the initiator, run, calls, and evidence without reconstructing someone’s private session. The improvement becomes a recorded comparison the team can repeat.

Task attributionModel-call usageFile activityQuality-aware comparison

CONTINUE THE WORK

Choose the model with the work in view

Inspect model and deployment choices, or see how a clear specification enables a cheaper implementation model.

See the model handoff
Models & deployment

Silent visual walkthrough · Narration shown above