How SaaS leaders are actually pricing AI in 2025
As a finance leader, I’ve seen firsthand how AI breaks legacy pricing models. Seriously… I worked at a company where costs were unpredictable, usage was spiky, and billing systems weren’t built for real-time models. That’s why I found Metronome’s AI Pricing in Practice: 2025 Field Report so valuable. It’s packed with firsthand insights from pricing and finance leaders at top SaaS companies on how they’re handling AI costs, hybrid billing, and credit models that don’t confuse customers.
If you’re figuring out how to price and package AI features sustainably, this is the playbook to start with.
For the last three months I’ve been in the lab.
I’ve spoken to founders and execs who are supercharging the tools we love and use each day with AI - from ERPs to Close Management to Invoicing and Billing to Procurement - to figure out wtf is actually real and game changing.
The output is a comprehensive report, with market maps built for each distinct layer of the cfo tech stack.
We. Love. Market. Maps.
It’s like Gartner, without the librarian tone or the software-like price tag (my report is free AF)
I’m sending the report early to a select group of finance leaders. Sign up to be one of them.
(Screw it, here’s a teaser…what’s going on in Core Accounting Software???)
The report also covers: FP&A, Close management, Invoicing & Billing, Rev Rec, Tax, Cap Table, AP, Employee Expense Mgmt, Travel, and Procurement
Finance in 2030
👋 Hi, it's CJ and this is it… the final post in our series on agentic finance. Next week you’ll get the entire report we’ve been cooking up. You could call it The Final Countdown.
(“If you heard this song, on the road, in the playoffs, your season was about to end”)
Over the last two weeks, we covered when NOT to use AI (most of the time) and where it actually works (the messy middle).
This week? The future.
What finance looks like when agents are everywhere, what you need to prepare for NOW, and how to actually adopt this stuff without lighting money on fire.
(Plus the uncomfortable question we're all thinking: Will my job exist in 2030?)
Spoiler: Yes. But not if your name is Brad it looks different.
Picture this:
It's Monday morning, 2030. You wake up to a briefing from your lead finance agent:
"Whilst you were sleeping, I reconciled 847 transactions, flagged 3 compliance issues for your review, proposed 2 policy changes based on error rate drift, and found five business trends you should look into. Your 9am close sync is ready. I've pre-populated the variance analysis."
(Side note: I picture my agent has a British accent and says shit like Whilst")
You grab coffee (from Dunkin. yes. still exists amidst the AI in 2030). You review the exceptions (there are 4, not 847). You approve the policy changes. You spend 30 minutes reviewing the company’s five year forecast instead of 3 hours chasing down discrepancies and asking your CRO why he expensed a first class ticket to Jamaica.
This isn't science fiction.
Kinda reminds me of a Waymo
This is where we're headed. And the CFOs who start (just start something, anything) will have a massive advantage over those who don't. Because these efforts and process changes compound.
The Timeline: What to Expect and When
Let's get specific. Because "the future is coming" is useless without dates. Rumor has it that Michael Burry predicted 19 of the last 3 recessions. And Ned Stark said Winter was coming for 16 straight summers.
2025: The Shakeout Year
What's happening:
Hype vendors start to die (people get through renewal cycles and know what works)
Finance teams start to put out recs with more "engineer persona" roles
Reasoning models improve dramatically (passing CPA exams)
Early multi-agent systems prove value (Brex + Navan, Tropic + NetSuite)
What you should do:
Start your data cleanup NOW (seriously, this takes 6-12 months at large companies)
Pick ONE low-risk use case to pilot (invoice matching, duplicate detection)
Build a muscle for validating AI outputs
Start tracking ROI obsessively, and celebrating publicly inside the company when it works
Most of 2025 is learning. That's okay. The companies that figure it out in 2025 earn the right to scale it in 2026-2027. Embrace the slog.
2028: The Inflection Point
What's happening:
First public company with a 10-person accounting team (likely a hyperscaler with simple revenue model)
Self-healing finance systems go mainstream (agents propose policy changes when error rates drift)
Close processes highly automated (but human-in-the-loop remains critical)
CFO/COO roles merge; COO joins the dinosaurs in an ever lasting extinction
Ubiquitous automation of receivables, payables, and recon work
What the finance org looks like:
Accountants still on teams, but they're lean and excellent at describing desired end states via prompts.
Finance becomes orchestration and oversight, not transaction processing
We finally get to that “one pane of glass” we’ve visualized as nirvana for decades, able to both see and query performance data across multiple systems of record.
2030: Finance as Control Tower
“I can see my ERP from here”
New roles emerge:
Policy engineers for SEC and technical accounting
Finance ops for AI (like data ops, or rev ops, but for financial agents)
Morning routine:
AI briefings on business trends identified overnight
Exception review (not transaction review)
What finance owns:
Oversight and governance
Exception handling
Short term, medium term, and long term planning
Explaining decisions to boards and auditors
What agents own:
Transaction processing
Multi step reconciliations
Data synthesis from disparate sources (CRM, ERP, emails, customer calls)
First-pass variance analysis
Compliance monitoring and alerting
Multi-Agent Ecosystems: The Real Destination
Here's what gets me excited (and a little nervous):
Agents will coordinate with each other across functions:
Treasury agents ↔ Procurement agents ↔ FP&A agents
Coordinating outcomes autonomously
It's already starting:
Brex and Navan agents interact today
Tropic agents will interact with NetSuite
Numeric's agents quarterback a cross-functional close
The big question: We talk about LLMs being open or closed ended models. Similarly, will finance move toward closed ecosystems (one vendor controls the stack) or open ecosystems (agents from different platforms coordinate)?
My bet is open and modular wins. But only if CFOs demand interoperability.
Don't get locked into proprietary agent ecosystems. The future is too uncertain to bet on one vendor owning everything.
Trade-offs to consider:
Closed systems: Tighter integration, potentially better performance
Open systems: Flexibility, innovation, no vendor lock-in
Choose wisely. (Or don't choose at all and let your vendor make the decision for you. That usually works out great. Not.)
The choice is yours.
I always thought Morpheus was just holding Dayquil on the left and Advil on the right
But the window is closing. The CFOs preparing in 2025 will be the ones dominating in 2028-2030.
Get the Full Report
This three-part series is just scratching the surface.
Next week, I'm releasing "The State of the Agentic Financial Stack" - a comprehensive report with market maps across the following layers of the CFO tool stack:
Core Accounting Software
Close Management Software
Invoicing and Billing
Revenue Recognition
Tax Software & Services
Cap Table Management
Accounts Payable
Employee Expense Management
Employee Travel
Procurement
FP&A Software
This isn't a hype or trends piece (that shit be for the birds). It's a tactical roadmap for how finance teams will work in the next 3-5 years, and who the hot vendors are today.
The report includes everything you need to make smart decisions about agentic finance tools:
When to adopt (and when to wait)
How to sequence implementation
Which vendors are solving real problems (and which are rebranded chatbots)
ROI frameworks that actually work
Risk mitigation strategies
And it's free.
Because I believe every CFO should have access to real, tactical information about the future of their function.
Run the Numbers Podcast
THE GROSS MARGIN EPISODE IS HERE!!!!
Gross margins, GPUs, and the future of finance… this one’s for all my metrics nerds.
I sat down with Sarah Wang, General Partner at Andreessen Horowitz, to talk about what happens when the traditional SaaS playbook collides with AI. Sarah shares:
How legacy benchmarks like payback period and burn multiple start to break down in a world where compute, not headcount, drives costs.
Why sky-high gross margins can actually be an orange flag
How finance leaders can think about resource allocation between engineers and GPUs
Why the most valuable finance teams today are deeply operational.
We also unpack what it’s like partnering with AI-native founders, the evolution of pricing models as LLM costs drop, and whether we’ll see a private trillion-dollar company anytime soon.
Mostly Growth Podcast
Companies are spreading their marketing budgets too thin these days, oscillating between experiments without enough “oomf” and channels without enough traction.
We also talk about the surprisingly durability of SEO, and the shocking lack of results when it comes to trying AI SDRs.
We discuss Limewire’s comeback, why CFOs should (or should not) own the analytics function, and founder driven linkedin marketing.
Finally, we roast startup swag.
Looking for Leverage Newsletter
Financial Due Diligence for Owner Operated Service Roll-Ups
Buying owner operated service businesses is a high-variance game. For every well-run operation, there are ten more that look fine on the surface but are quietly bleeding working capital, masking liabilities, or inflating performance through one-time heroics.
And here’s the twist: it’s usually not malicious. These aren’t public company CFOs. They're mechanics, franchisees, or families trying to retire.
Here’s what we’ll be covering:
The GAAP Gap: What does the business actually earn in profits each year?
Revenue Quality: Customer concentration and recurring vs transactional revenue
The Hidden CAPEX Bomb: Has equipment maintenance been deferred?
Normalizing Working Capital: How much gas are they leaving in the tank for day one?
Quote I’ve Been Pondering
“How do you know when you have sufficiently good colleagues? In my experience, when you do, you know. Which means if you're unsure, you probably don't.”
Hoping you don’t pay AI prices for RPA,
CJ
