# Why Buy vs. Build Is the Wrong Question for Financial Institutions in Banking's AI Era
_Published: 2026-10-14T00:00:00.000-04:00_

Will Jung breaks down why financial institutions should stop asking whether they should buy or build and start asking how to do both.

_By Will Jung, Chief Technology Officer at nCino_

For as long as I've worked in banking technology, large institutions have framed their biggest modernisation decisions as a choice: build it yourself or buy it from someone else. That framing has held banks back longer than most leaders realise, because it was never really the right question. The real question isn't build or buy. It's what to buy, and what to build. Getting that split right now is the difference between banks that can move at the pace AI demands and banks that can't.

The pull toward building in-house is understandable. [According to global research from NTT DATA](https://datacentre.solutions/news/63869/banks-build-their-own-technology-solutions), 61% of banks prefer to build some banking software internally rather than buy it from an external provider. Most of the time, that preference has less to do with genuine business advantage and more to do with decades of legacy dependency. When your core has been custom-built and re-built for fifty years, "build it ourselves" starts to feel like identity rather than strategy.

But AI has changed the calculation, and it's changed it quickly. Building effective banking AI today isn't a matter of hiring a few data scientists and connecting a language model. It requires a unified data foundation across your lines of business, models grounded in real banking workflows and regulatory requirements, explainability frameworks regulators will accept, and governance infrastructure that runs continuously, not once at launch. Each of those is a serious undertaking on its own. Together, they're a multi-year, multi-million-dollar programme with no guaranteed payoff: [research from BCG](https://www.bcg.com/press/24october2024-ai-adoption-in-2024-74-of-companies-struggle-to-achieve-and-scale-value) found that 74% of companies have yet to show tangible value from their use of AI, and only 26% have developed the necessary capabilities to move past proofs of concept.

That's the case for buying the foundation. It's also, increasingly, the case for building on top of it, because the alternative, treating "buy" as the end of the conversation, leaves just as much value on the table. The institutions getting this right aren't choosing between buy and build. They're doing both, deliberately, on two different halves of the problem.

What's worth buying is the part every bank needs and none of them should have to re-invent: the AI infrastructure and model lifecycle underneath it, the regulatory compliance and explainability layer that satisfies auditors, the governance and audit trails that prove a decision can be explained after the fact, and the ongoing work of keeping all of that current as both the technology and the regulations around it keep moving. None of that is where a bank actually competes. It's the cost of entry.

What's worth building is everything that makes one bank different from another: the credit policies and risk frameworks that encode how this institution actually thinks about risk, the workflow logic that matches how its bankers really work rather than how a generic system assumes they work, the data models that reflect its specific portfolio and customer base, and the new products, features, and market moves that are the actual competitive game.

**Building well isn't automatic, though.** It starts with knowing precisely what makes your bank different, not just assuming anything you didn't buy is worth building. Most banks can point to this on a slide, but far fewer can say precisely which parts of their credit process, their risk appetite, or their customer experience are the real source of competitive advantage. Building the wrong things well is still the wrong outcome. And that takes talent that's genuinely scarce right now: engineers who can build production-grade AI capability, not just experiment with it. The honest question every bank should ask itself is whether it can actually give its own engineers room to move at that pace, the way an organisation whose entire job is building this kind of capability can or whether that talent ends up spread thin across infrastructure work a platform could have handled instead.

That buy the foundation, build the advantage split is what lets a bank move with the speed of a smaller, less regulated competitor, without actually being one.

[Fifth Third Bank's experience](https://fortune.com/2024/12/04/fifth-third-cio-ai-buy-rather-than-build/) shows what that split is worth in practice. When the bank weighed building its own commercial loan origination system against buying one, it found that developing in-house would have taken roughly five times longer and cost roughly twice as much. That gap is only getting wider on the AI side of the ledger, where building production-grade data infrastructure, model governance, and compliance tooling from scratch now takes years banks in a genuine growth market don't have to spare. Institutions that buy this kind of foundation instead of building it typically see 20–30% productivity gains within 18 to 24 months, without ever carrying the upfront build cost.

**Regulation is making this decision harder to postpone, not easier.** In Australia, APRA's operational risk standard, [CPS 230](https://www.apra.gov.au/standards/cps-230), has been in effect since July 2025, and as of July 2026 the transition period for existing arrangements has closed. The standard now applies in full, with no legacy exceptions left to lean on. That means every regulated institution has to demonstrate not just that its own systems are sound, but that its critical operations and material service providers are too, across the board. A bank trying to build and maintain that governance layer in-house is signing up to rebuild it every time the standard evolves. A bank that buys a foundation built by an organisation whose only job is keeping that layer current gets to spend that valuable energy and time elsewhere.

**The cost that's harder to see on a budget line is talent.** Software engineers are expensive to hire under the best of circumstances and the [Harvard Business Review notes](https://hbr.org/2021/12/when-should-your-company-develop-its-own-software) that recruiting fees alone typically run 15% to 30% of a new engineer's first-year salary, before you've paid them a cent. AI and data science talent commands an even steeper premium, and BCG's research suggests the harder problem isn't even the hiring: 70% of the real challenges in building AI capability are people and process issues, not technology ones. And while AI has made coding easier and faster, good engineers who understand design, security principles and scalable infrastructure are even more important, and scarce, than ever before. Every engineer a bank puts on infrastructure and compliance work that a platform could already handle is an engineer not working on the credit policy, the workflow, or the customer experience that's actually theirs to build, and in a market where competitors are moving fast, that's capacity a bank can't afford to misallocate.

None of this is an argument against building. It's an argument for building the right things. Buy the foundation that every bank needs and would spend years re-inventing. Build the parts that make your institution genuinely yours. That split is the strategy: the one that lets a bank stay compliant, stay agile, and stay in the race against competitors who've already stopped trying to do everything themselves.

_This is a bigger decision than a single article can cover fully. For the complete framework, including how to weigh speed, resourcing, and control tradeoffs, and the data behind the numbers cited here, read our white paper _[_Buy and Build: A Strategic Guide for Financial Institutions_](https://word-edit.officeapps.live.com/we/wordeditorframe.aspx?new=1&ui=en-US&rs=en-US&wdenableroaming=1&mscc=1&hid=269E39A2-2002-F000-4C72-AE777F4ED42E.0&uih=sharepointcom&wdlcid=en-US&jsapi=1&jsapiver=v2&corrid=4444a07f-eb8d-c532-9824-ffacec1c4732&usid=4444a07f-eb8d-c532-9824-ffacec1c4732&newsession=1&sftc=1&hfto=1788983216034.7&uihit=docaspx&muv=1&ats=PairwiseBroker&cac=1&sams=1&mtf=1&sfp=1&sdp=1&hch=1&hwfh=1&wopisrc=https%3A%2F%2Fncino.sharepoint.com%2Fsites%2Fmarketing.all%2F_vti_bin%2Fwopi.ashx%2Ffiles%2Fbfa0788f02d44e8884cf3f239c23c3de&dchat=1&sc=%7B%22pmo%22%3A%22https%3A%2F%2Fncino.sharepoint.com%22%2C%22pmshare%22%3Atrue%7D&wdorigin=DocLib&ctp=LeastProtected&rct=Normal&wdhostclicktime=1788983215897&afdflight=68&csiro=1&wdredirectionreason=Unified_SingleFlush#)_._

Will Jung serves as nCino's Chief Technology Officer, responsible for leading the company's engineering, architecture, and AI platform strategy across a global organization. With a career built equally in banking and technology, he translates deep financial services expertise into intelligent solutions for the agentic banking era. Prior to becoming CTO, Will served as nCino's General Manager of Product in Australia, and before joining nCino spent 15 years at Macquarie Group leading product and technology across retail banking, business banking, and client engagement.

---

[View sitemap](https://www.ncino.com/api/markdown/sitemap)