Buy vs Build: When "I Could Just Get AI to Do It" Becomes a Trap

Every quarter, CFO Techstack brings together the FLAC: the Finance Leaders Advisory Council, a small group of finance leaders from across the community who meet to compare notes on where the stack is heading.

They don't always have the answers. But they're usually asking the questions before the rest of us are.

This quarter's topic was AI. And one debate came up again and again: now that AI can build almost anything, when should you actually let it?

The seduction of "I could just build that"

AI has made it genuinely tempting to skip the purchase order and build in-house instead; a 3-way model, a deferred revenue schedule, a custom reporting tool. One finance leader put it plainly:

"If I can automate it using AI, why wouldn't I?"

It's a fair question. But several people in the room pushed back on it just as fast, pointing out that the real cost of a build isn't the build, it's everything after.

Maintenance is the bit nobody budgets for

One leader, weighing up building a bespoke customer invoicing tool versus buying something off the shelf, laid out exactly what "build" actually commits you to:

"Maintenance is a really big part of that. Who owns it, who maintains it, who understands the build, who - if they leave - who takes over it. Auditability is another one."

That's the pattern across the group: AI makes the building easy. It does nothing for the owning. A tool that took an afternoon to spin up still needs someone to understand it in eighteen months, when the person who built it has moved on and the edge cases start showing up.

Sometimes the smartest fix isn't a tool at all

The FLAC were also quick to highlight that there is an important third option: stop doing the task altogether.

One example that came up was a 90-store franchise running monthly COGS accrual journals, reversed every month, for amounts too small to matter.

Nobody had asked whether the work was necessary in the first place; AI had just made it faster to keep doing.

The group's general instinct: before automating a process, ask whether the process should exist at all.

Buy is often the boring, correct answer

Where a specialist tool has genuine depth; accounting complexity, tax logic, quality audit trails, deep integrations with your accounting system (the kind of thing that's easy to get wrong)... several in the room concluded buying still wins, even at a real dollar cost:

"It's actually just cheaper to pay for the tool than to build it ourselves and change everything all at once."

When build is the right call

The group wasn't anti-build. Where the need is genuinely specific to your business, and simple enough that someone other than the builder could pick it up, building can be the faster and better option.

Before you build ask

Does this task need to exist? If not, the cheapest build is no build.
Who owns it in 18 months? And who takes over if they leave?
Can it be audited, and is it secure? If your auditor or board asked how it works, could someone answer? Is it safe and secure?
What happens if it's wrong? If getting it wrong is expensive or embarrassing, lean towards a tool built by people who do this full-time.
Is it genuinely bespoke? If plenty of businesses have the same need, someone has probably already built it better.

___

In short, the FLAC's view was this: build when the use case is genuinely bespoke and you have the team to maintain it. Buy when your needs aren't that unique, and/or getting it wrong would be costly.

And before you do either, ask whether the task needs to exist.

‍

Have you built something you now regret, or bought something you wish you'd built? Come and share it in The Stack Exchange. We'd love to hear how you're making the call.

‍

COMMUNITY INSIGHTS

🎤 How I stacked it

Katrina Bignasca, Group CFO at Asseti
Winner: Finance Leader of the Year for 2026 🏆

Asseti is a real asset management and monitoring platform, designed to revolutionise how asset managers predict and operate property assets at scale. As Group CFO, Katrina is responsible for finance, payroll, ISO and support across sales, operations and processing.

The The stack has continued to evolve alongside the business. Asseti's processing centre in Nepal has joined the group, US operations have grown exponentially, and a new branch has opened in New Zealand, increasing the geographic footprint again. The biggest shift has been AI becoming a core part of the stack, alongside new tools for dynamic forecasting and real-time reporting, and the full rollout of platforms for international payroll, payments and FX.

Visit Asseti.

Here are the apps Asseti runs across Australia, the US, New Zealand and Nepal:

Xero is the core accounting platform at the heart of the stack, giving the group one central place for its financial data that every other tool connects to.
Mayday handles intercompany reconciliations and harmonises the chart of accounts, which matters most when acquiring businesses or bringing in legacy entities.
Joiin brings data together from multiple entities and Xero organisations for consolidated reporting, giving a group-level view without pulling each ledger separately.
Maxio runs subscriptions and SaaS reporting, backed by onboarding and support that continues well past implementation.
HubSpot tracks communications and automates workflows, not just for sales but for processes that run across other apps.
Rippling runs HRM across the whole group and payroll in Australia and the US, having fully replaced Employment Hero in a single group-wide switch.
Rigo HR handles native payroll in Nepal, while Rippling works on localising its payroll offering there.
xpna feeds live AR and AP data straight into forecasting, replacing a manual weekly update. Three-way forecasting is next.
Airwallex manages payments and FX across the group, now including the new New Zealand branch.
Pulse replaced Power BI entirely, giving real-time visibility over financial and operational data, with office dashboards in the works.
Claude has been embedded since March as the reporting and analysis layer, with every workflow and SOP documented in Notion alongside it.

See Asseti's full stack.

If you want to showcase your stack contact jack.thiel@getmayday.com

‍

GIVE ME HOPE

Reader's question:

"We are expanding into a second country for the first time. What do I need to have in place before we do, and what do most companies get wrong at this stage?"

Harriet's answer:

The honest answer is that it depends heavily on the country. At Arbolus we managed multi-entity accounting across the UK, Spain, India and the US, and each of those brought its own quirks. No two jurisdictions throw up the same set of problems.

Start by finding a reputable local advisor in the target country, not necessarily a big consulting brand. Plenty of smaller local firms do excellent work at a fraction of the cost, and in my experience the same firm that helps you incorporate can often carry on to handle accounting, tax and payroll long term. And if you're looking for suggestions, ask the community - The Stack Exchange is a great resource.

Once you've found an advisor, ask them for a summary of what you actually need to know: legal structure options, how corporation and withholding tax work locally, staff costs and payroll tax, and any permanent establishment risk. Depending on what the entity will actually be doing, a subsidiary may not even be the right structure. A branch or an employer of record can sometimes do the job with far less complexity.

Where most companies come unstuck is statutory compliance. Some countries are unforgiving about deadlines. Miss a payroll tax payment in Spain, for example, and you'll be fined regardless of how new you are to the market. Being unfamiliar with local rules is not a defence. Sit down with your local accountants early and understand the particularities of how things work there. You don't need to do the bookkeeping yourself, but you do need to understand it well enough to review what comes back to you. Otherwise you're signing off on numbers you can't actually interrogate.

Transfer pricing is the other piece people underestimate. How much rigour you need depends on your stage. A high-level study from a local or HQ-country firm is often enough early on. As the business grows, that usually needs to become more detailed, which takes more time and money, so it's worth planning for rather than reacting to.

The common mistake is moving fast on the commercial opportunity while doing too little diligence on the practical realities that follow. It's a genuinely useful problem to work through, and one that will teach you more about how businesses actually operate across borders than almost anything else in the finance seat.

Have a question you’d like answered? Submit your questions here.

‍

EVENT

"Year End Close: Community Drinks" 🍻🌎

Year end is close, and we're celebrating it in person, around the world!

If you call any of these cities home, register and come along. We'd love to catch up properly.

Adelaide, Thursday 3 December: Book your spot
London, Thursday 3 December: Book your spot
Melbourne, Wednesday 9 December: Book your spot [sponsored]
Sydney, Thursday 10 December: Book your spot [sponsored]
Perth, Thursday 10 December: Book your spot
Cape Town, date TBC: Register your interest

Note, spots are limited so please only register if you intend to attend.

‍

THE STACK EXCHANGE

🧵 This week's top threads, from The Stack Exchange

The Stack Exchange is the place to connect with peers, stay on top of the latest apps and industry news, and work through the challenges that don't have an easy answer.

Want to join the conversation? Sign up free here.

‍

NEW IN THE WORLD OF CFO SOFTWARE

🗞️ News from the stack-o-sphere

Stripe acquires Parafin to expand revenue opportunities and help small businesses grow. Read more
Spendesk launched their "Top CFO Tools Report 2026". Read more
Dext launches payments in Canada, giving businesses and advisers one place to manage supplier and employee payments. Read more

‍

WHAT THE DATA SAYS

📊 Stat of the week

IBM surveyed 1,500 CFOs globally and found that 62% say their role has expanded into enterprise technology or AI strategy leadership. Only 6% of CFOs say their finance function is “transformation-ready” for AI.

Read more in IBM's CFO Study here.

‍

MORE OF THE GOOD STUFF

And lastly, our top picks!

🎧 Podcast: Why Pegasystems is betting on work, not tokens | Ken Stillwell, CFO, Pegasystems. Listen here.

📝 Article: You need a new CFO | Andreessen Horowitz. Read here.

😆 Joke: The sales forecast and my kid's Christmas list have a lot in common. Both are ambitious, neither comes with a budget, and somehow I'm expected to deliver both by December.

‍

Why not forward this newsletter to someone you think would enjoy it?

‍