Finout ads

Forecasting, allocation, unit economics: the FinOps basics were built for cloud, and the real question now is how much of that still works when the spend is AI. Some of it holds up well, and some of it breaks in ways that aren’t obvious until finance starts asking what the spend is actually producing, which a count of chatbot conversations won’t answer. Whether your team is just starting to track AI spend or already reporting it, join Victoria Levy (Member of the FinOps TAC) and Elly Rauch (Finout) live on November 4. They’ll walk through which parts of the FinOps Framework carry over to AI, which ones quietly break, and what to bring to finance instead of another dashboard. Save your seat here, and if you can’t make it live, register anyway and we’ll send you the recording: https://lnkd.in/eSJm7QjW Wednesday, November 4, 2026 | 12 PM ET | 9 AM PT

Your FinOps stack works fine until AI spend enters the picture. Then allocation breaks, shared costs get messy, and finance and engineering stop agreeing on the numbers.

FinOps keeps evolving, but we're still measuring the wrong things. Most teams track adoption by user counts or automation speed. Neither tells you if anyone actually owns their cloud spend. For us, cost visibility lived with a handful of analysts who became gatekeepers of the bill. More dashboards weren't going to fix that. So we changed the question: What % of our cloud spend is monitored by at least one engineer from each team, every month? Once that became the KPI, things moved: ✅ 14% of spend shifted from unmonitored to actively managed ✅ 31% more engineering teams started tracking their own costs ✅ One team responsible for 34% of total spend grew their adoption by 160% ✅ 41% MoM jump in platform users across the org FinOps stopped being an accounting exercise and became part of how engineers build. Analysts went from gatekeepers to enablers. Doing this with Finout in our corner made the shift possible. Full story here: https://lnkd.in/d9xeEASY Curious how others are thinking about this. What's actually moved the needle for your org beyond cost savings? Einav Shuker Uri Argov

Anthropic's new AI gateway is genuinely great. It's also not ready to run across 5,000 engineers on Monday. Both are true. That's the whole story. So I wrote up the honest version- the good, the bad, and who should actually adopt it now. The good is real: → Governance at the choke point, not an agent on every laptop. → Usage streams out on open standards, into tooling you already run. → Spend finally tied to a person and a team, not one aggregated model bill. The catch is just as real: → It's day one. You'd be running brand-new critical infrastructure yourself. → It governs one tool. Not Cursor, not Copilot, not your Bedrock bill. → A per-vendor gateway is a feature. It is not a cost strategy. My take: Pilot it now if you run Claude Code at volume and already live in OpenTelemetry. Wait if you need turnkey reliability across every AI tool from day one. The pattern is production-ready before the product is. That's exactly what early adoption is. Full breakdown- the good, the bad, and who should wait- in the comments. Are you piloting this, or waiting for it to mature? #FinOps #GenAI #AIInfrastructure #LLMOps #CloudCost

Ask any FinOps lead how much their company spends on AI. Confident answer. Then ask for the bill — the line items won't add up to the number. That gap is the problem. AI cost now lives across four distinct layers. Different pricing. Different owners. Different blind spots. → Cloud AI services — Bedrock, Vertex, Azure OpenAI. Your cloud bill sees this. → Direct AI contracts — Anthropic, OpenAI direct. Seven-figure commitments your cost tools weren't built for. → AI-native tools — Cursor, Claude Code, Copilot. Uber burned its entire 2026 AI budget on these in four months. → Embedded AI inside SaaS you already pay for — buried in vendor invoices finance thought were flat. Most cost platforms see one or two. Almost none see all four. 78% of IT leaders reported unexpected SaaS charges last year. Up from 65%. The spend is moving where visibility is thinnest. The old playbook hasn't caught up. Which layer does your team see the least? #FinOps #GenAI #AIInfrastructure #CloudCost #LLMOps

Our CISO has been screaming about Shadow AI for two years. Unauthorized ChatGPT logins. Customer data in prompts. Compliance migraines. Here's what nobody's saying: Shadow AI is also a CFO problem now. And it's growing faster. Look at the AI spend layer almost nobody tracks — features quietly added to SaaS apps you already pay for: → M365 Copilot at $30 per user. On top of your existing seat. → Slack AI at $10. Notion AI at $10. Zoom Companion bundled in. → Salesforce Agentforce — usage-based, billed in actions. → Atlassian Rovo. ServiceNow Now Assist. Adobe Firefly credits. Average enterprise runs 300+ SaaS apps. Large enterprises run 1,000+. Gartner says 40% will have AI agents built in by end of 2026. Do the math. A third of your portfolio adding AI at $10–$30 per user per month = a parallel AI bill in the thousands per employee per year. Invisible to every cost tool you own. Department head approves the renewal. AI line item never had a budget. Your CISO has been screaming for two years. Your CFO is about to find out it has a price tag. Part 2 of the 4-Layer AI Bill series Which SaaS line is hiding the biggest AI add-on in your portfolio? #FinOps #ShadowAI #GenAI #AIInfrastructure #tokenomics #AIOps #FinOpsx #CFO