Building software for value

For years, the “buy versus build” debate followed a familiar pattern. For many mid-market organisations, buying off-the-shelf software was considered the lower-risk, more predictable route. Custom development was certainly a powerful tool for competitive differentiation, but it was often viewed as resource-intensive and difficult to scale.

 

Over the past year, that dynamic has begun to shift. We are entering what many describe as the era of “vibe coding”, where the distance between a business idea and a working technical solution has closed dramatically, thanks to AI-assisted development.

 

This does not mean that organisations are abandoning SaaS, far from it. Off-the-shelf platforms remain fundamental to modern IT estates, but they are no longer the automatic answer to every new requirement.

 

Building closer to where the work happens

Across the mid-market, organisations are increasingly choosing to build tailored AI agents and lightweight applications within the environments where their people already collaborate, and where business data is actively maintained.

 

Rather than exporting information into yet another standalone platform, teams are developing AI-driven capabilities inside their primary cloud and productivity ecosystems, the same environments that already house documents, communications and operational workflows.

 

In Microsoft-centric estates, for example, that often means extending capabilities within the existing stack rather than procuring an additional niche SaaS product. For many organisations, Microsoft 365 now has a ‘centre of gravity’ in terms of day-to-day document storage and teamwork through the use of Microsoft Teams and SharePoint.

 

Documents are continually created, edited and shared within Microsoft 365, and many organisational processes need to work with current data rather than a one-off snapshot. We are seeing a surge in organisations using technology like Copilot Studio and Microsoft Foundry agents to work with data in-place, handling tasks and automations that would have previously required a separate SaaS subscription.

 

The rationale is straightforward. When data is already governed, secured and continuously updated within a core platform, embedding AI there can reduce duplication, limit integration complexity and improve user adoption. Instead of asking employees to adapt to another rigid workflow, organisations can shape tools more closely around how work is actually done.

 

Standardised software remains the most efficient solution for many commodity processes, but it marks the end of SaaS as the unquestioned default.

 

 

From hype cycle to value cycle

Public discussion around AI often centres on market valuations and competitive arms races. As Sundar Pichai has observed, there is an element of “irrational exuberance” in parts of the sector. Yet within mid-market organisations, the conversation is noticeably more grounded.

 

Leaders are less concerned with headline-grabbing breakthroughs and more focused on measurable gains: productivity improvements, faster reporting cycles, enhanced cybersecurity monitoring and streamlined internal processes.

 

Most IT decision-makers now view AI as a positive force because they are seeing practical returns. An AI agent that eliminates hours of manual reconciliation or automates routine service queries delivers tangible value, regardless of what happens to technology stocks tomorrow.

 

 

Closing the strategy gap

If there is a challenge, it lies not in access to tools but in clarity of purpose. The capability to generate code or configure AI-driven workflows has accelerated rapidly. Deciding which problems are worth solving, and how those solutions align with long-term architecture, remains the harder task.

 

Many organisations still report a gap between AI’s potential and realised outcomes. The democratisation of development makes it easier to create solutions quickly, but without governance and strategic direction, speed can introduce complexity.

 

This is where IT leadership must evolve. The function is moving beyond operational support towards shaping priorities, setting guardrails and ensuring that experimentation translates into sustainable business value. AI initiatives need to be secure, maintainable and aligned with broader transformation goals, as well as technically impressive.

 

Increasing flexibilty

The one-size-fits-all enterprise stack is giving way to something more flexible. The buy-versus-build decision is becoming less binary and more contextual, less about budget alone and more about agility, integration, and control.

 

For UK leaders, the opportunity lies in being deliberate. Standard platforms will continue to drive efficiency and consistency. However, tailored AI agents and solutions, embedded within the environments where the business already operates, can provide the differentiation needed to succeed.

 

Organisations shouldn’t build everything to succeed, nor buy everything. It’s all about making conscious choices about where standardisation serves them best, and where intelligent, well-governed customisation can unlock the next wave of value.

 

---

 

Chris O’Brien is CTO of Advania UK

 

Main image courtesy of iStockPhoto.com and MTStock Studio