With AI enabling businesses to build their own software at the ‘speed of light’, concerns are rising over the future of traditional software. The concern is that companies will replace third-party vendors with custom-built applications. But the evidence doesn’t support this.
CEO of Basware.
What leaders are getting wrong with AI vs software
According to research, 44% of CFOs feel pressure to adopt AI. And 61% say they are funding “experiments” by CIOs. But these experiments are largely failing, which is why AI will not eat software.
The digital transformation officer at a large European company, who has spent around a million euros on internal AI-related projects in finance over the past year, recently told me he couldn't point to a single penny that the company has saved, earned, or helped the business in any way.
And he’s decided to pull the plug and bring in vendors who know how to use AI in their processes and workflows.
When to build vs buy
AI can’t simply be set free. To be trusted and effective, it requires clean data, strong governance and the right infrastructure. And it has to operate on accurate, integrated data across systems and transactions and work with human oversight.
So, the key questions to ask when considering how and where to use AI are: will it improve the company’s value proposition, and is there a case for owning a unique data set?
If so, building custom applications can drive differentiation and long-term value. Proprietary risk models are a good example of when in-house AI solutions can deliver ROI, with outputs linked directly to leadership scrutiny, and models improve over time as internal data compounds.
It does, however, require a higher upfront cost, the need for ongoing maintenance, and strong AI governance.
If not, the smarter move is to buy solutions from trusted, AI-powered software from vendors with deep expertise, proven outcomes, and established moats to deliver faster results without unnecessary risk. Buying offers more predictable returns and, when automating an efficient process, can avoid potential pitfalls.
In short, buy to accelerate, build to differentiate, but never do either blindly.
Why finance is a good testing ground
Finance functions offer one of the best tests to demonstrate the real value of AI tools. Many tasks in finance, particularly in areas like accounts payable, are still highly manual and repetitive. Especially when tasks are data-heavy and require pattern recognition, AI can scale contextual judgement well beyond manual teams.
These processes are easily measurable, as they are directly tied to business outcomes such as revenue protection, cost control, fraud detection, and cash flow management, making it far easier to prove the impact of AI investments.
AI is already helping automate many of these time-consuming tasks. When embedded within operational software, AI doesn’t just predict outcomes, it can execute tasks securely and at scale.
And when finance teams no longer spend hours on invoice capture, validation or exception routing, they can focus on higher-value, strategic work, from cash flow optimization to supplier strategy and compliance oversight.
How AI is already delivering ROI
We’ve reached a tipping point where boards and CEOs are done with AI experiments and expecting real results.
Organizations that leverage intelligent platforms that are embedded with AI and anchored in areas like finance where value can be proven and scaled rather than cobbling together point solutions or attempting to build their own can quickly deliver them.
And the research, bears this out. According to the data, overall ROI on AI has risen from 35% to 67% since 2024, and companies using third-party solutions already embedded with AI are seeing the strongest results, achieving an average ROI of 80%.
The focus when implementing AI in finance is disassociating it with experimentation and tying these projects directly to ROI. After AP, the top three agentic AI deployments that finance leaders are targeting is automating invoice capture and data entry, cash flow management, and scenario modelling and forecasting.
That’s the proving ground for CFOs and AI.
The clear message in all of this is that AI can deliver transformational results, but only when it is deployed with purpose and discipline. And that means embedding it directly into finance workflows, grounding it in trusted data, and governing it with clear policies. This is how AI moves from innovation to impact.
The AI is going to kill software narrative makes for good headlines, but we’ve heard this story before. Every major technological shift has sparked doomsday scenarios, and none have played out. AI is one of the most disruptive technologies we’ve ever seen. And it’s certainly going to change the way business gets done.
But it isn’t going result in the wholesale replacement of software. What it will do for companies that apply it properly is drive greater efficiency and productivity and more impactful outcomes.
We've featured the best AI website builder.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit