Do CFO / COO / CIO Become 1 Role?

This week we have the following CFOs answering your reader questions:

If your question is picked for the Mailbag, you’ll win a Mostly Metrics Yeti Rambler ($84 economic value, unlimited street cred).

Topics for today:

Now let’s get into it!

I don’t think these roles fully merge, but the overlap is legit and getting larger.

The CFO is increasingly the connective tissue between strategy, data, AI, and risk (especially when you think about customer data/privacy in an HR/payroll software business like mine). AI pushes this hard because finance sits closest to the source of truth and can better articulate the trade-offs.

If you want to be a CFO in ~5 years, the job is less about accounting mastery and more about judgment: capital allocation and understanding how AI changes messy data into decisions. You don’t need to code, but you do need to understand systems, incentives, and where the bottlenecks are. The lines between CFO and COO are definitely blurring.

I don’t think there’s a single, final version of the CFO role. Job titles can be misleading, and context really matters.

From my career - at Coda, Superhuman, and Google - the CFO title meant something different at each place. The industry matters too. For example, a software CFO needs to understand AI, data, and product speed, while a manufacturing CFO needs to deeply understand the movement of physical goods. The skills and work should match how the business creates value.

The other big variable is the management team you’re on. Organizations adapt to the people within them. I’ve played closer to product on some teams and closer to GTM on others. The CFO/COO/CIO blend is less about org charts and more about who’s best positioned to own a decision and the responsibility behind it.

Hemant Tenaja (CEO of General Catalyst) is an investor I’ve worked with closely for the past 10+ years. And at one point, he gave the advice to “play your game” - don’t play to the expectations of others.

Said differently, learn where you get energy and what you’re good at, develop that as a deep spike, and then join teams and companies where that creates alpha in the role.

Others may see a “CFO” title, but what matters is how you complement the management team, solve problems, and help build a winning company.

I think the CFO org stays intact (FP&A, Treasury, Controllership, Procurement). However, the clear lines of other organizations (COO/CIO/InfoSec) will continue to blur across the enterprise.  In one direction, the CFO org will play a more influential role across the enterprise to provide important insights with capital allocation, customer analytics, operations, risk management and compliance. And in the opposite direction, the Finance org will be heavily reliant on InfoSec teams for better data insights and more automated reporting.  

So, even with the blurred lines, Finance will remain a separate org, as it will still need to maintain independence and be the overall steward of data integrity across the organization.

This notion of independence will need to be maintained across strategy (likely in COO org), as Finance can serve as a "check" or independent assessment of capital allocation and would be best suited to ascertain buy, build, or partner analysis.  

What this means for an upcoming CFO... master financial management (table-stakes), understand how all the other teams interplay across the organization, and understand how enterprise data is best managed.

You can change definitions, but you can’t change reality. Rolling early expansion into “New ARR” might make sense internally if that’s how customers really ramp. But externally, be careful. Investors anchor hard on NRR as a signal of product strength. Juicing New ARR at the expense of NRR usually backfires.

My bias: keep external definitions clean and consistent, and run a separate internal view that shows “land + ramp.” If you do change definitions, do it once, explain it clearly, and never touch it again.

I’m a bit of a purist about this.

Your metrics should reflect how customers actually use your product, not just what looks good on a board slide.

In a land-and-expand motion, I’d anchor on which signal captures the most meaningful economic commitment for the product. If your motion is truly PLG-first and then upsell into a managed motion, I’d define New ARR as that very first credit card purchase. That’s when the customer decides, “This product is now part of my workflow.”

If your managed sales process is more independent, and you’d never close a deal without a salesperson, then I’d define New ARR as starting with the first enterprise contract. In that case, product-led growth is more about generating leads than driving revenue.

Those are the two extremes, but most companies fall somewhere in the middle. Smart investors will easily see through definition hacks - especially when they start modeling the business themselves. And at the end of the day, you want your definitions to reveal the underlying dynamics of your business, not obscure them or excessively inflate one aspect.

responses…

On R&D, I expect companies to spend roughly the same % of revenue, but ship more and faster. The winners won’t cut R&D - they’ll reinvest the productivity gains into speed and experimentation.

On close: we’re not at one-click month end, but AI already helps more than people think. The biggest wins aren’t fancy agents inside Excel - they’re upstream. Auto-recs, anomaly detection, variance explanations, draft commentary. If your data is clean (you put sh*t in you get sh*t out), AI saves a ton of time. If it’s messy, no agent is saving you.

AI doesn’t fix broken finance hygiene - it just exposes it faster.

In the medium term, I think R&D spending as a percentage of revenue will stay about the same, but output will keep growing. Competition will drive this.

AI tools make every software development team more productive. You could take that to the bottom line, but in competitive markets, the real move is to reinvest in more experiments, shortened feedback loops between product, sales, and customers, and bolder roadmaps. These productivity gains will lead to way more software getting created.

To the finance angle specifically, IMO we’re still early. At Superhuman, we’ve been experimenting with a small set of packaged tools - for example, shifting from FloQast to Numeric - and general-purpose productivity (Claude, ChatGPT, etc.), but I haven’t yet seen the “vibe coding moment” for month-end close or Finance/Accounting processes generally.

That said, even if we get there, I’m betting that we will continue to have some form of human-in-the-loop process - perhaps shifting away from “black-pen” work (where you do it from scratch) and more “red-pen” work (reviewing agentic outputs).

Part 1: While I do expect the innovation and product release cycle to speed up, I also believe there will be some increase with cloud usage which may partially offset the ROI. Still early days on innovation, but likely the expectations across any market participant, in order to effectively compete, is to leverage AI tools. 

R&D % may come down initially (less incremental headcount required), but I believe this will likely have to increase as all companies will gravitate towards the new innovation paradigm, and eventually will require additional investment to generate a competitive advantage (whether that be lowest cost or product value). These added costs may come in the form of incremental cloud (likely COGS allocation) but may also find themselves with AI licensing (yes, I do expect these subscriptions to go up with the billions being poured into the sector) and a retrained middle management layer that will manage junior engineering and AI team members.  

Part 2: I have yet to find an AI tool that can effectively manage a close. Not to mention leveraging any AI tool within Finance requires clean, high-quality data that would minimize any errors with accounting.

That being said, automation plays a critical role with shortening our monthly close process. For us, revenue accounting estimates are usually conducted at month-end after aggregating all upstream data files (i.e. not enough rows in Excel) that have been received through the last day of the month and reconciling this data to our cash balances with our partner banks; this often spills across several days post close to synthesize the data and accrue for any estimates. We are now beginning to employ transaction level matching and cash reconciliation on a daily basis to truncate monthly close - the tool we employee is called Trintech.  

So automation can still play a huge role with speeding up the monthly close. However, I have yet to see AI deliver month-end financial statements, as producing financials still requires accounting judgement based on management decisioning within GAAP.

I think about R&D as Run, Win, Expand.

At scale, Run is usually ~40%, Win ~40%, Expand ~20% (give or take). If Expand gets too big without proof, returns fall off a cliff. If Run gets too big, velocity dies. The CFO’s job isn’t to starve Run - it’s to keep pushing it down over time so you can fund Win and Expand without wrecking margins.

I loosely follow a 70/20/10 rule that I’ve borrowed from my time around Google-style operators.

Protect the 70, rigorously curate the 20, and time-box the 10.

Ultimately, you want to get towards an effective balance were 50%-60% goes into keeping the lights on and 40%-50% placed on innovation. Many companies, particularly within FinTech, have amassed complex tech infrastructure and cumulated significant technical debt along the way; thus having to allocate 70%-80% of R&D with maintaining platforms. Not where you want to be.

CEO and CFO every meeting. Non-negotiable.

CFO provides continuity across performance, and capital - boards expect that. CRO should present when GTM, efficiency, or growth motion is changing, not just by default. Other leaders should present when there’s a decision, risk, or inflection point - not for visibility.

Boards don’t want airtime, they want the truth on what’s relevant at that moment. Simple test: if you’re not asking for input or alignment, you probably shouldn’t be presenting.

I don’t begin with a set schedule or speaker rotation. Instead, I focus on what the board meeting needs to accomplish. A good board meeting does three things, in this order:

Once you see it that way, who presents becomes a design choice rather than a template.

I typically see the CEO and CFO in all sessions (excluding CEO debriefing with the Board).  The CFO will present Financial snapshot of where the company is (performance and cash), current outlook, and known risks & opportunities. The COO may come in to present updates around key OKRs and discuss customer pipeline. 

Ultimately, the presenters are up to the CEO, but my recommendation is to keep the attendees consistent with an emphasis on transparency and objectivity in order to build trust and cultivate the right level of dialogue.    

Wishing you a define your expansion ARR the right way,

CJ