Harness is introducing AI Budgets, a new capability in its Cost Management Agent that turns AI cost policies into live limits.
It comes as companies struggle to keep AI spend in check. Last week, Accenture told employees it will now fund just $29 a month per person for Claude, after some staff spent up to $1,000 a month. Most organisations have an AI cost policy, but only 13% can tell whether it’s actually working.
AI Budgets flags runaway AI spend before it hits the invoice. Teams set weekly limits by team, developer, model or project. Users get a warning at 90%, and access pauses at 100% until the weekly reset or a one-click approval. It pulls spend from Cursor, OpenAI, Claude Enterprise, Amazon Bedrock, GitHub Copilot and others into one view.
Harness tested it on its own AI bill first. Average weekly spend per developer fell from $184 to $115, and cost per million tokens dropped from around $1.80 to $0.80 the week governance went live.
More info here: Introducing Harness AI Budgets
Related
This entry was posted on October 20, 2026 at 12:18 pm and is filed under Commentary with tags Harness. You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
Harness launches AI Budgets – a live limit on AI spend
Harness is introducing AI Budgets, a new capability in its Cost Management Agent that turns AI cost policies into live limits.
It comes as companies struggle to keep AI spend in check. Last week, Accenture told employees it will now fund just $29 a month per person for Claude, after some staff spent up to $1,000 a month. Most organisations have an AI cost policy, but only 13% can tell whether it’s actually working.
AI Budgets flags runaway AI spend before it hits the invoice. Teams set weekly limits by team, developer, model or project. Users get a warning at 90%, and access pauses at 100% until the weekly reset or a one-click approval. It pulls spend from Cursor, OpenAI, Claude Enterprise, Amazon Bedrock, GitHub Copilot and others into one view.
Harness tested it on its own AI bill first. Average weekly spend per developer fell from $184 to $115, and cost per million tokens dropped from around $1.80 to $0.80 the week governance went live.
More info here: Introducing Harness AI Budgets
Share this:
Like this:
Related
This entry was posted on October 20, 2026 at 12:18 pm and is filed under Commentary with tags Harness. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.