Anthropic’s New CFO Tools: Why Claude Is Getting Receipts

anthropic just cracked open the corporate billing ledger rolling out three tracking tools that show enterprise buyers exactly what their model spend is purchasing in episode 27 we watched claude duet and gpt six turn workplace agents into a corporate turf war today anthropic is trying to lock down that turf by arming chief financial officers with granular receipts the new features smart reports pooled budgets and in app surveys are now generally available for cloud enterprise users smart reports log which internal tasks burn tokens track friction points and identify repeatable workflows meanwhile pooled budgets let administrators set shared department caps alongside individual seat limits preventing heavy developers from stalling while casual accounts sit on idle cash why does

this matter right now anthropic just filed its s one prospectus reporting 900% net revenue retention and $16,300,000,000 in revenue for the first half of twenty twenty six but quarter on quarter growth cooled to 47% in the third quarter enterprise buyers are scrutinizing opaque ai line items anthropic is handing them the magnifying glass that threatens enterprise seat aggregators and third party observability startups whose whole business model is demystifying ai invoices at the same time the regulatory gloves came off in sacramento anthropic and openai just submitted competing safety blueprints to california governor gavin newsom following his september 18 executive order they agree frontier systems need external scrutiny but they violently disagree on who pulls the plug anthropic

wants the state attorney general empowered to seek court orders that suspend model deployment or internal development if independent evaluators find catastrophic risk anthropic argues for circuit breakers that isolate a specific rogue agent or cut off an individual client openai wants the state to keep the courts out of it in its filing openai pushed for amendments to senate bill 53 that would mandate developers build internal shutdown controls and impose an enforceable duty on labs to trigger them directly to openai an external court injunction is too slow when code is actively executing to anthropic trusting frontier labs to police their own kill switches is an unacceptable conflict of interest finally openai is pushing dot its new personal agent straight into your daily communications united states pro subscribers can now text their dot agents over imessage and rich communication services routing instructions photos and voice memos directly through partner infrastructure from linked any memory dot captures over text syncs back into chatgpt automatically openai wants consumer intimacy but anthropic is building enterprise infrastructure that companies cannot rip out

while open ai tries to make agents feel like your most helpful contact anthropic is convincing corporate finance that claude is the only line item they can actually measure and control that makes the enterprise spending fight less about raw intelligence and more about who controls the audit trail