AI · Oct 19, 2026
Microsoft ships Guardian Agents, saying frontier models ‘can’t be trusted’Anthropic gives enterprise buyers an AI receipt, and a reason to stay
· Updated · Signal Desk newsroom
Analysis of reporting by TechCrunch. How we work

Anthropic’s new reporting shows managers what a team produced, what it cost and where workflows broke, plus pooled token budgets. With agentic coding spend up 10x this year, proof of value is overdue, but a dashboard inside the vendor’s own product also deepens lock-in.
Key takeaways
- New reports show a team’s output over a date range, its cost and where workflows broke down, with recommendations.
- Reports draw on internal analytics and employee surveys; Anthropic says it won’t train on the data by default.
- Pooled group budgets let companies manage token spending across teams.
- Jellyfish found agentic coding tool spend grew 10x since the start of the year, now about 2% of engineering cost.
- Jellyfish data shows Claude-based tools behind nearly a third of proposed code changes submitted for review.
What launched
Anthropic launched reporting tools that let managers see what a team produced over a date range, what it cost and where workflows broke down. The reports combine internal analytics with employee surveys and come with recommendations. Anthropic says it will not use the data for training by default, and it is adding pooled group budgets so companies can manage token spending across teams. TechCrunch described the offering as a kind of forward-deployed-engineering lite.
Why buyers want proof now
The biggest question in enterprise AI right now is less which model is smartest than whether the money is buying anything. A Jellyfish report cited by TechCrunch found spending on agentic coding tools has grown 10x since the start of the year, to roughly 2% of total engineering cost, while most of the Fortune 500 still are not seeing major gains from AI. Jellyfish’s chief scientist called it growth most CFOs did not plan for. When a line item grows that fast, finance asks for evidence; Anthropic would rather supply the answer than see a budget cut while a CFO looks for one.
Anthropic has its own numbers to protect. It told investors it recorded an operating profit in the second quarter and 900% net revenue retention, meaning the same customers spent nine times more year over year. Its IPO filing describes a land-and-expand playbook from API to apps (Anthropic’s retention strategy), and a free year of Claude Team for startups aims to make Claude the default early (startup offer).
The lock-in and the monitoring questions
These reports are also retention machinery. Jellyfish’s AI product lead described Anthropic’s position as incumbent weight: teams standardized workflows around Claude, and switching costs are not zero even when tools are similar. Jellyfish’s data, drawn from 276,000 of its customers, found Claude Code and other Anthropic-based tools behind nearly a third of proposed code changes. OpenAI’s Codex has gained ground since its Astra model arrived this summer, though Jellyfish says that has not slowed Anthropic’s adoption so far. A dashboard that lives inside the vendor’s product and measures value by the vendor’s definitions deepens that weight.
Usage analytics plus employee surveys is also a lot of visibility into how individuals work. The no-training pledge is welcome, but how managers use the reports matters more: a tool meant to find broken workflows can easily become a productivity scoreboard. Anthropic’s own study of its workplace found fewer junior hires and more time in meetings (Anthropic’s workforce), a reminder that measurement shapes work. Companies should set ground rules before switching reports on. More in our Anthropic coverage and AI coverage.
Who feels it
- Engineering leaders
- A cost-versus-output view of AI coding spend and pooled budgets across teams.
- CFOs
- A vendor-supplied ROI story, useful but worth checking against independent measures.
- Employees
- Analytics and surveys raise questions about how granular workplace monitoring becomes.
- Competing labs
- Matching model quality is not enough when the incumbent also measures the value.
What to watch
- Whether OpenAI and Google ship comparable spend-and-value reporting.
- How net retention holds once ROI data reaches finance teams.
- Any controversy over employee-level visibility.
- Further Jellyfish data on whether Codex narrows Anthropic’s lead.
FAQ
What do Anthropic’s new reports show?
What a team produced over a date range, what it cost, where workflows broke down, and recommendations, drawn from analytics and employee surveys.
Does Anthropic train on this data?
Anthropic says it will not use the data for training by default.
How fast is AI coding spend growing?
A Jellyfish report cited by TechCrunch found agentic coding tool spend grew 10x since the start of the year, to about 2% of engineering cost.
Companies: Anthropic
More on Anthropic
AI · Oct 13, 2026
After months off the frontier, Google ships Gemini 4 Argon with benchmark claims and aggressive pricingAI · Oct 20, 2026
Anthropic’s draft federal AI bill would let a judge halt a model release over mass-casualty riskAI · Oct 8, 2026
Anthropic’s new usage policy sharpens election rules, loosens voter targeting and bars sustained cruelty toward ClaudeAI · Nov 5, 2026
Perplexity Teammates bets that always-on work agents are worth $200 a month