What does digital development take in the age of AI? Three capabilities that matter more than coding

21 Aug 2026

Building digital business has changed, and a lot of organisations haven’t yet acted on the openings that change has created. AI has closed much of the gap between big companies and small ones. And the winner is no longer whoever writes the most code.

What decides success now is harder to pin down: knowing how to define what you want, setting clear rules for how AI gets used, and running a project that spans several suppliers at once. This piece looks at each of those three in turn.

Why building the thing is no longer the hard part

The hard part of digital development has moved from building to understanding. Pioneers have been saying this for years, of course, but it’s true across the board now. You used to describe a technical problem to an expert, who solved it with skills almost nobody else had. Those same tools and that same knowledge are everywhere today. The value isn’t in the code AI writes for you. It’s in having worked out how to apply that code to your own work.

So the thing that matters most is how precisely you define the goal and the requirements before anyone starts building. A badly defined goal can now be reached fast and, on the surface, cheaply. It’s still the wrong goal, and the bill lands later.

One thing hasn’t changed, though. The total cost of digital development hasn’t really dropped, even though producing code got cheaper. The money just moved. A smaller share goes to writing code, and a bigger share goes to working out what needs doing and keeping the result under control. Procurement works the same way. With so many ready-made solutions on the shelf, the buyer’s main job is knowing what to order and why.

Capability 1: You define before you build

Good digital development comes from being able to focus on what counts. Planning matters more than it used to, but it shouldn’t drag on. A sharp specification lets you move to building quickly. Once you’ve described the goal, the users and the constraints well enough, you don’t need a hundred-page document.

In practice this is three things. You audit where you stand before you buy anything new, meaning you take stock of your digital environment, how people use it and where the worst bottlenecks are. You write your development practices down so they don’t live in one person’s head. And you make sure everyone shares the same picture before each phase of development, not just at kickoff. That way the organisation knows exactly what it’s building. It’s worth more than clever code, and once the direction is clear the technical side follows.

It also helps to have someone outside assess where your digital development stands, because your own habits are hard to see from the inside. Sofokus has a service for this: Sofokus IT audit, which comes in two parts. The value audit asks whether the digital service actually delivers what it should. The technical audit goes under the hood: the system’s condition, its architecture, where it stands in its lifecycle, the way it’s built and maintained.

The work runs partly as workshops and partly as independent consulting, and it leaves you with a prioritised report for the development team plus a summary for management and decision-makers. Our client Suomen Hippos, for instance, had us audit the Heppa system, which around 150,000 people use every month. The report laid out development priorities and, at the same time, gave them material that made it easier to talk the matter through with management, funders and the authorities.

Capability 2: You steer AI with a practical AI handbook

AI has put things within reach of organisations of every size that used to need a development team of their own. Because it multiplies what a handful of people can produce, the size of the tech budget no longer settles the outcome. What settles it is who can steer the use of AI across the whole organisation.

An AI handbook is built for exactly that. It spells out, in plain terms, how AI is used in your processes, your procurement and your development. Think of it as a management tool that gathers up the questions many organisations haven’t had time to sit down with: where AI belongs, what data you can feed it, who’s responsible for what, how you check quality, and which risk category your use cases fall into under the EU AI Act.

The handbook has to be a living document. Instructions written once go stale and end up in a drawer. Knowledge that never gets written down doesn’t pass on reliably. It leaks away as people come and go, and it shifts a little every time someone repeats it. A living handbook grows with the organisation and builds a shared understanding, both inside the company and with its partners.

Capability 3: You can run a team of several suppliers

Building your digital development around a single supplier leaves you exposed. In reality there are nearly always several: one runs the customer portal, another the ERP, a third the integrations, a fourth the infrastructure. Whether the whole thing works comes down to how their work is coordinated and what ground rules everyone has agreed to. Seeing the full picture can’t be left to any one supplier.

In a partnership that works, each side brings expertise, data and insight to the table, not just billable hours. That takes information flowing both ways and a partner who understands the wider business. At its best, the supplier feeds you things you didn’t ask for: what the market is doing, which new technologies are worth a look, how you compare to others. And you open up your internal processes just enough to let the partner actually help.

When several suppliers are involved, someone has to hold the overview even when someone else does the building. That’s how the work is set up for one of our clients, the Federation of Finnish Police Organisations. A consultant has given them steady support defining what the operation needs, spotting those needs in the first place, and running the tender for the membership register, where data-security requirements went into the procurement criteria from the start. Some of the systems were built by another supplier, some by Sofokus.

Using AI across several suppliers needs shared rules too: what data may go into the tools, how the output gets checked, who owns quality. Those belong in the AI handbook above, which applies to every supplier equally.

It’s worth looking for partners inside your own sector, too

The give-and-take doesn’t stop at your digital supplier. What we’ve seen is that the best results come when organisations that don’t compete with each other team up. Take two district-heating companies in different parts of Finland: they do the same work under the same obligations, but neither threatens the other, because the business is local. When that’s the case, joint specification, joint procurement and a shared development partner all become possible, and they bring a scale neither company could reach alone.

The Salo region shows it plainly. Non-recyclable waste is turned into energy and then into heat for homes, with three regional players each holding one link in the chain. Lounais-Suomen Jätehuolto collects the waste, Lounavoima turns it into energy, and our client Salon Kaukolämpö distributes the heat to homes and businesses. Lounais-Suomen Jätehuolto and Salon Kaukolämpö own Lounavoima together, and it’s that shared ownership, plus a genuinely open mind about technical solutions, that lets the three of them pull off what any one of them would find far harder alone.

Digital development follows the same logic. When players along a value chain build something together, say a shared approach to AI or to digital customer service, they get access to expertise and scale that would be too hard to reach on their own.

Where do I start? 

The natural place to start is by taking stock of where you are. From there the work branches into four stages, and the first three take only a few weeks each.

1. The audit sets out where you stand: what systems you run, what shape they’re in, how you develop them, and what eats the most time.

2. The specification sets out what you’re building and why, in enough detail to start building and refine as you go.

3. The AI handbook sets out the rules for AI: what you use it for, what data is in play, who’s responsible, and how those rules reach every supplier.

4. The multi-year partnership covers building and maintenance, extra specialists when you need them, and coordinating a multi-supplier setup.

The fourth stage only starts once the first three have led to decisions. And if the whole thing feels like a lot, it’s enough to begin with the audit and decide on the rest afterwards.

FAQ 

What is an AI Handbook? 

An AI handbook is a written, regularly updated description of how an organisation uses AI: in which cases, with what data, under whose responsibility and by what rules. Management tool, not technical manual. The idea is to make using AI a controlled, shared way of working.

Should digital development start with the specification?

Usually yes, as long as it doesn’t run on forever. Describing the goal, the users and the constraints precisely enough heads off the most expensive mistakes, because even a badly defined goal can now be built fast and cheap. A good specification is written so building can start early and sharpen as the project moves.

How does a digital partnership help organisations that don’t compete?

Through scale and speed. When you’re not competing, you can share a specification, share procurement and share a development partner. An online service built jointly, for example, costs less and lands sooner than the same thing run as several separate projects.

What does the Sofokus IT audit involve? 

Two parts. The value audit looks at the user experience and whether the service delivers benefits you can measure. The technical audit weighs how the system works, its architecture, where it sits in its lifecycle and how it’s developed. Out of it you get a prioritised report for the development team and a summary for management, and that summary earns its keep in funding and procurement conversations too.

How do you run a project with several suppliers?

By agreeing the ground rules and naming someone to hold the big picture. In practice that means agreed interfaces and clear lines of responsibility, a shared understanding of the architecture, and rules for how AI is used and how its output is reviewed. The coordinator is best placed when it isn’t itself building every part.

Want to know where to focus your digital development next?

An audit is a good place to start. It weighs up the state of your systems and your development practices, and points to what gives you the most for your money. You come away with a prioritised set of recommendations to take forward with the development team, management and investors. From there it’s a natural step to specification, AI rules and supplier coordination. You can also kick things off by bringing your whole value chain along. Get in touch, and we’ll take a look at where you stand together.