Origin Technology

Know where your intelligence is going.

Origin shows how your organization’s context and AI budget move through the AI stack, so you can understand what the investment supports and where it can work harder.

The Gap

The bill leaves out the most important part.

The invoice does not show the work.

Vendor bills show tokens and seats. They don't show which workflows created the cost, and what you got in return.

Frontier models become the default.

Providers often default to their most capable models, even when the work doesn't need them.

Your context is scattered.

The instructions and workflows that make AI useful are spread across providers, making their value difficult to see.

The ROI question is coming.

Leadership wants to know what AI is supporting. Most teams still only have the bill.

How It Works

Open the black box of token economics.

See what is driving token spend and connect the cost back to the work.

Observe

Map AI consumption.

Track usage and estimated cost across prompts, sessions, models, providers, tools, agents, and employees on endpoints.

Range7d14d30dProviderallanthropicopenai
Now, against the prior half
Estimated spend · 30d 18.1%
$56.7K$2.05K/day now vs $1.73K/day over the prior 15d
Run-rate / month
$61.4Kprojected from the recent daily pace
API requests 74.2%
256K10,824/day now vs 6,213/day over the prior 15d
Cost per request 32.2%
$0.22$0.19 now vs $0.28 before
Saved by prompt cache 5.6%
$295K
Unusual and expensive

This total is a lower bound. 78% of attributed tokens are priced. 39.7B tokens are excluded because no model was recorded, or the per-call count was impossible.

Prompt-cache reuse fell from 96% to 45% of context$173K
Caching still saved $151K. Holding the earlier rate would have saved about $173K more.
claude-fable-5 costs 3.3× the blended rate, $1.72 per 1M tokens$22.6K
It carries $22.6K, 39.9% of spend. The blend is $0.53 per 1M.
Daily spend is up 18.1%, now $2.05K/day against $1.73K/day before$4.71K
$30.7K over the last 15 days, against $26K in the 15 before.
Attribute

Connect cost to the work behind it.

Cluster AI activity by team, topic, and workflow, so spend stops being a black box and turns into something you can explain.

What work the money went to/cluster-spend

Clusters are named from the work itself, so spend arrives already grouped by what it bought.

ClusterRequestsCost
Release readiness checks4,945$11.2K
Applying review updates3,327$7.48K
Worktree approvals1,836$4.13K
Automated notifications1,395$3.14K
Goal recap interactions1,131$2.54K
Coding state actions921$2.07K
Pushing commits to remote861$1.94K
Pull request status727$1.64K
Top 8 of 74 clusters · cost estimated from attributed tokens
Optimize

Find the waste.

See where frontier models are the default for routine work, and where the same problem is being solved more than once.

claude-fable-5$1.72 / 1M tokens · 3.3× the blend
Where this model was used$22.6K

Area is spend. Red is work a cheaper model could do. Hover to isolate a flow.

$3.33K went to routine lookups a cheaper model handles identically.

Prove

Defend the investment.

Answer the questions every AI program will face. What did we spend? Where did it go? What did it produce?

ModelToolProvider
Where the money goesclaude-opus-5 is 49.8% of $56.7K, across 14 models
claude-opus-5$28.2Kclaude-fable-5$22.6Kgpt-4o$2.99Kgemini-2.5-pro$1.2Kother providers$1.17Kunattributed$540
What changed+$4.71K
prior 15d $26K → last 15d $30.7K
claude-fable-5+$3.83K
gpt-4o+$956
claude-opus-4-7-$922
claude-opus-5+$567
gemini-2.5-pro+$100
other providers+$88
What it produced
1,284Pull requests opened
3,610Reviews completed
74Clusters of work
42Endpoints active

Origin turned our token telemetry into a savings plan.

It showed what was driving the bill, and found savings without moving important work to weaker models.

$1Mannual AI bill
$250Kfound in savings
Who It’s For

Connect AI spend to the wider business picture.

For CFOs

Understand what is driving AI spend and where the next dollar should go.

For CIOs

See the full picture of AI usage across approved tools, unapproved tools, models, and agents on endpoints.

For IT leaders

Compare which AI tools employees are actually using and where usage is spreading outside approved systems.

For operations leaders

Find duplicated work, inefficient workflows, and opportunities to standardize how teams use AI.

Know what your AI budget is buying.

Origin connects AI usage to the work behind it, so you can see what the investment supports and where to put the next dollar.