Spending on AI is easy. Knowing what it bought is not.

WHAT HAPPENED

Anthropic launched reporting tools for token spend

Managers can see what a team produced over a date range, what it cost, and where workflows broke down. The reports also recommend improvements. Anthropic is adding pooled group budgets too. TechCrunch calls it a clear attempt to show ROI.

THE NUMBER

10x

Spend growth on agentic coding tools since the start of the year

Jellyfish: now about 2% of total engineering cost. Most CFOs didn't plan for it.

THE REAL PROBLEM IN COMPANIES

What I see in companies: spend is tracked, outcomes are not

1. 01The bill shows up in finance. The result never shows up anywhere.
2. 02Teams adopt a tool, but nobody owns the question of what it delivered.
3. 03Scaling starts before anyone knows if the first workflow paid off.

I'VE ALREADY SOLVED THIS

AskMoses.AI, the SaaS I built end to end and run as a contractor

1. 0117 organizations in the US signed up. Only 6 actually use it.
2. 0211 of 17 never activated: onboarding needed a CRM and a rubric before any value.
3. 03Script Intelligence mines winning CRM calls into a weekly script. 88% gross margin.

MY TAKE

Launching is not the same as being used.

A dashboard of tokens spent tells you activity, not value. Tie each AI workflow to a business outcome and measure what it delivers per dollar before you scale it. That is how you tell real adoption from shelfware.