Build vs. Buy in the Age of AI

For years, buying software was the safest and fastest choice. SaaS reduced technical risk, saved development time, and gave teams access to proven functionality without having to build and maintain it themselves.

In many cases, that is still true, but AI is changing part of the calculation. That doesn't mean every organisation should suddenly start building everything it needs; that would be expensive and unnecessary.

The shift is more specific: AI is making it faster and more affordable to build tailored software around the systems, data, and processes an organisation already owns.

That changes the build-versus-buy discussion. The question is not just whether to build or buy, but which workflows are important enough to own.

Why the default is changing

SaaS solved real problems. Someone else built the software, your team configured it, the product improved over time. The trade-off was control, you adapted your processes to the tool, accepted the platform's roadmap, and paid per user or per event. For many years, that trade-off made sense.

AI-assisted development is shifting it. It does not remove the need for strong engineering, architecture, security, and business judgment; these still matter. But it helps experienced teams move faster: prototyping more quickly, automating repetitive tasks, building useful internal tools with less effort than before.

That matters because many SaaS tools are not replacing a core system. They are sitting on top of data and infrastructure the organisation already owns, reporting layers, activation workflows, internal tools, customer segmentation. When your data already lives in your cloud environment, it is worth asking whether every extra platform still adds enough value.

The SaaS renewal moment

This question becomes relevant when a SaaS contract is up for renewal.

A platform that made sense three years ago may now cost significantly more. The organisation may also have more data in the cloud, a more mature internal team, and better tools to build tailored solutions. The question becomes very practical: does this platform still deliver enough value for what it costs?

For some tools, clearly yes.

A good SaaS platform brings functionality, support, compliance, and integrations that would be difficult to recreate. But for tools that mainly act as a layer above data and logic the organisation already owns, the answer is less obvious.

The question is whether you are paying for unique value or simply for a wrapper around a foundation you already built.

Where ownership starts to matter

This does not apply to every system. Most organisations should not rebuild their finance system, HR platform, or CRM just because AI makes development easier. Standard software is still the right choice for standard processes.

The opportunity is more relevant where the workflow is close to your data, your business rules, or your competitive advantage:

• Analytics and reporting

• Customer data and activation

• Internal workflow tools

• Decision support and operational dashboards

• AI-driven recommendations and business-specific automation

These are the places where generic tools start to feel restrictive, where the business wants definitions that match how it actually works, data teams want logic that is transparent and maintainable, and commercial teams want tools that support their process instead of forcing workarounds.

Example: analytics and BI

A retail organisation had two years of transaction and customer data in BigQuery, with a mature data team managing pipelines and models. On top of that, they were paying €180k per year for a BI platform that was primarily acting as a visualisation layer and requiring them to maintain business logic in two places: the warehouse and the tool.

When the contract came up for renewal, they asked a different question: what do we actually need here?

The answer was a trusted semantic layer, clear metric definitions, and simple interfaces for the right teams. They moved metric definitions into dbt as versioned, governed code. Dashboards were rebuilt on a lighter tooling layer sitting directly above the warehouse. The €180k contract was not renewed.

The benefit was not only lower cost. It was more control over the logic that drives business decisions, logic now maintained in one place, by one team.

Example: customer data and activation

A B2C company was running a CDP on top of a cloud data warehouse where all customer data already lived. The platform was doing audience segmentation and activation, but it pulled data copies out of the warehouse, applied its own logic, and synced results back. The marketing team regularly found that segments did not match what the data team saw in the warehouse. Resolving discrepancies took days.

At renewal, the team rebuilt audience logic as warehouse-native models. Customer definitions, segment rules, and suppression logic all lived in the warehouse as versioned SQL. Activation syncs ran directly from there to downstream channels.

The marketing team got fresher segments. The data team had one place to maintain logic. Governance questions that used to require support tickets became self-service.

This will not be the right move for every organisation. But for this company, the CDP was adding complexity on top of a foundation they already owned and controlled.

Buy, build, or own

The answer is not to replace SaaS with custom software everywhere. That creates a different problem: too many custom tools, too much maintenance, too little focus.

The better approach is selective. Buy where the process is standard. Use SaaS where the platform creates clear, distinct value. Build where the workflow is specific, strategic, and close to your data. Own the logic where control, flexibility, and governance matter.

A practical place to start is your renewal calendar.

For each contract coming up in the next 12 to 18 months, ask:

• Does this platform solve a problem we cannot solve better ourselves?

• Does it create unique value, or mainly sit on top of data we already own?

• Are we adapting our process to the tool, or does the tool genuinely fit how we work?

AI has changed what small, experienced teams can build. The organisations that benefit most will not be the ones that replace every tool. They will be the ones that know where standard software is enough and where owning the workflow creates a real advantage.

Before your next renewal, it is worth asking whether you are paying for software you genuinely need, or for a workflow you could now own.