When someone asks "how’s business?", they’re really asking "how are your finances?"

Every business decision is a financial decision. Yet too many high-leverage teams spend their weeks stuck in reverse: chasing lost receipts, policing SaaS subscriptions, and grinding through a close that drags into next month.

That’s why I use Brex.

It’s an intelligent finance platform with AI workflows that absorb the manual maintenance: corporate cards, real-time policy enforcement at the swipe, and continuous reconciliation. Close takes days, not weeks.

Stop playing defense. Join Anthropic, DoorDash, and Coinbase, and take back your momentum at brex.com/metrics. 

My state of the Agentic Finance Report is about to drop. We cover 12 layers of the financial tech stack, from core accounting systems to FP&A to Tax to procurement.

If you want a copy of the report before the release date, sign up below. It’s free. It’s comprehensive. And it also like almost killed me to write lol. Holy market mapz.

We now have AI costs that behave very similarly to cloud costs, but show up in every department. It's pervasive.

If we harken back to the 2010s when cloud burst onto the scene, CFOs ceded a lot of that power to their CTOs to control the capital that was allocated to our AWS, Azure, and GCP bills.

The majority of it lived in our Cost of Goods Sold, and a small portion could be allocated (if you argued convincingly enough) to R&D. Either way, you could tie it back to what the CTO's org was doing. You didn't have cloud costs showing up in sales (unless you're Anthropic and claim in your S1 that all free usage should go here), and you definitely didn't have them showing up on the finance team.

It's like the electrical wiring of the house.

That's why the acquisition of OpenRouter by Stripe and the launch of Router by Ramp are so critical. Stripe agreed to buy OpenRouter for like a bajillion dollars, which routes token usage across +400 models from more than 80 providers. And Ramp was one of more than 10 companies, Cursor included, that launched their own routers in the same fourish hours.

The TL;DR on routers: A router is basically Kayak for models. If you want to fly to Cancun, you don't need to be tied to Delta, and you get to decide whether the layover is worth the savings.

But you only buy a router if you think you can save money by using it. As Peter Pezaris of Proxon asked me:

How do you know?

I just put out a video on how to audit your Claude bill for savings, and the number of plan permutations is endless. One CFO told me five people on a single team were on Claude Max plans. Meanwhile, the company-wide usage pool was running 45% over for the month. And a few others on the same team were paying for $20 OpenAI plans. Even within one vendor, what data you can see depends on whether you're on a team, enterprise, or personal plan and how it's priced.

And then you hear these horror stories. There is this one company, a company that we all use in our daily lives, honestly the company kinda sucks, so I don’t feel that bad, and an engineer spun up an experiment and spent $400,000 on tokens in three days before anyone noticed and turned it off. Shucks! Maybe that’s why they increase the price on my household indiscriminately each year!

So routing is one step, but first and foremost we need observability to be invented all over again. When I was in cyber, we were always benchmarking ourselves against New Relic, the OG of cloud observability. Peter and his cofounder Claudio Pinkus sold their last company to New Relic, and now they're building Proxon to do the same thing for AI. It attributes spend to vendors, models, teams, workflows, and owners, then compares the investment with the results it supports.

When Peter showed me the demo, my first reaction was "holy shit." I had seen the light. A flashlight that goes down to the department level and gives a measure of ROI.

Also, the CTO can't carry this one alone, for three reasons.

Justin Etkin, the CEO of Tropic, calls this always-on procurement. Something I didn't fully appreciate before speaking to him is the downstream impact of how often Anthropic changes its model and what it does for your own forecasting. It makes procurement a much hotter topic than the sleepy function it used to be.

He extended the analogy to hardware companies, which manage the direct goods that go into cost of revenue completely differently than the rest of their OpEx, and hire procurement early to get that cost dialed. Software companies never had to bother, because gross margins were so high there was plenty of cake to go around. With AI eating into cost of revenue, the stakes of being a 30% gross margin business vs a 50%, 60% one are HUGE.

As a CFO, throughout all the convos I have with finance leaders, I'm viscerally feeling the tension between AI as something we pay for vs being part of the product we offer.

What does that mean? You need someone with a financial background, who also understands usage-based pricing, to wrap their arms around it.

Julio Martínez, CEO of Abacum, a Built for NetSuite certified planning vendor, sees it the same way:

The CTO still decides which models are worth it for engineering and which are overkill, so it's a partnership. But finance must protect and hold two jobs: metering and managing what the company spends on AI, and figuring out how to fairly pass those costs on to customers as more of the product runs on inference instead of cloud, humans, or software.

This is too big of a cost for CFOs to cede this time around. You need to step up and play a bigger role.

In this report we cover new and existing players across the following layers of the tech stack:

My goal is to make you more educated buyers of technology. That required me to interview more than 30 people and individually research 145 companies across the 12 layers we cover

A big shout out to the following partners and thought leaders who made this report economically viable to spend months researching:

I should probably go get some sunlight now. After +120 days of working on this, my Lenovo skin tan is turning green. But in the meantime, sign up to get the report early! It’s dope.

Sid Kumar has ran GTM teams responsible for strategic planning at companies like Hubspot, AWS, and now Databricks. This is the most tactical podcast I’ve recorded on planning for usage based pricing companies. Learn from one of the best at one of the best UBP companies.

Wishing you a procurement software that does good procuring,

CJ