Engineering the digital foundations: AO Group's 2035 view

Progress has always required infrastructure

Every major leap in human progress has depended on infrastructure.

Roads connected markets. Electricity transformed industry. Telecommunications collapsed distance. The internet connected people, businesses and information at a scale that would once have been unimaginable.

The next decade will be no different.

The technologies may change, but the underlying principle will not.

By 2035, the world will need more software. More data. More computing capacity. More storage. More connectivity. More digital infrastructure. And significantly more energy to power it all.

At AO Group, we believe the organisations that understand this relationship will be better positioned for what comes next.

Because there is no meaningful AI strategy without infrastructure.

There is no digital economy without software.

And there is no technological progress without solving real human needs.

The world will need more code, not less

There is an increasingly popular idea that artificial intelligence will dramatically reduce the need for software engineers because AI can now generate code.

We believe the opposite is more likely.

AI will make software development faster.

That does not necessarily mean the world will produce less software.

It may mean we produce considerably more of it.

When the cost of producing something falls, demand often expands.

Businesses that previously could not justify building specialised software may now be able to do so. Systems that remained unchanged for twenty years because replacing them was too expensive may finally be modernised. Manual processes that were previously too small to automate may become economically viable to digitise.

Entire categories of software may emerge that would previously have been uneconomical to create.

AI therefore doesn't eliminate the software opportunity.

It expands it.

AI changes engineering. It doesn't remove engineering.

AI-enabled development tools are already changing how software is created.

Engineers can explore codebases faster, generate repetitive components, accelerate testing, document systems and investigate problems with dramatically greater efficiency.

AO is incorporating these capabilities into how we approach software engineering, legacy modernisation, architecture and data migration.

But generating code is only part of engineering.

Someone still has to determine what should be built.

Someone has to understand the business problem.

Someone has to decide how systems should interact.

Someone has to understand security, data, architecture, resilience, performance, regulation and the consequences when something goes wrong.

The more powerful our tools become, the more important good engineering judgement becomes.

The legacy opportunity is enormous

Some of the world's most important organisations still depend on systems created decades ago.

Banks. Insurers. Governments. Telecommunications companies. Manufacturers. Logistics businesses. Healthcare organisations.

Many of these platforms continue operating because replacing them is difficult, expensive and risky.

The challenge is rarely simply old programming languages.

Legacy systems contain years — sometimes decades — of accumulated business knowledge.

Rules.

Exceptions.

Integrations.

Data structures.

Operational processes.

Institutional knowledge.

Modernising them requires understanding all of those relationships before deciding what should survive, what should change and what should disappear.

This is where we see AI-enabled engineering tools becoming particularly valuable.

They can help teams analyse large codebases, document dependencies, understand data structures and accelerate discovery.

Combined with model-driven architecture and experienced engineering teams, this can materially change the economics of legacy modernisation.

Data migration becomes an engineering discipline of its own

Every transformation eventually encounters data.

And moving data is rarely as simple as moving rows from one database to another.

Old organisations contain duplicate records, undocumented structures, inconsistent formats, missing relationships and years of historical decisions embedded inside their information.

AI can increasingly help identify patterns, map structures, propose transformations and highlight anomalies.

But again, automation is only part of the answer.

Data has meaning.

Understanding that meaning requires context.

The future of data migration will therefore combine intelligent automation with human validation, governance and engineering judgement.

Banking will become more behavioural

Financial services provides a useful example of how we see digital systems evolving.

Historically, banking systems have been exceptionally good at recording transactions.

Money arrived.

Money left.

A balance changed.

A payment occurred.

But a transaction tells us what happened.

Behaviour can help us understand why.

As data, analytics and AI mature, financial institutions will increasingly be able to understand patterns across customer behaviour and respond more intelligently.

That creates the possibility of banking experiences that are more contextual, personalised and proactive.

This is the thinking behind AO's work in behavioural banking: moving beyond systems that simply process transactions toward systems capable of understanding patterns of behaviour.

Done responsibly, this could fundamentally change how people interact with financial institutions.

Fraud has already moved to real time

The same digital infrastructure that creates opportunity also creates risk.

Payments are becoming instantaneous.

Commerce operates continuously.

Customers move between devices, channels and countries.

Fraudsters operate at machine speed.

Fraud prevention therefore cannot depend entirely on investigating what happened yesterday.

Increasingly, the decision must happen while the transaction is happening.

AO's work with technologies such as Tazama reflects this shift toward real-time fraud monitoring and detection.

The future of financial crime prevention will increasingly depend on systems capable of evaluating events, patterns and behaviour as they occur.

AI is ultimately an infrastructure problem too

AI often appears to be weightless.

A prompt goes in.

An answer comes back.

Behind that interaction sits an extraordinary amount of physical infrastructure.

Data centres.

Processors.

Storage.

Networks.

Cooling systems.

Power generation.

Transmission infrastructure.

Physical buildings.

And people.

As AI adoption grows, demand for all of these things grows with it.

The digital economy may appear virtual.

Its foundations are very physical.

More intelligence means more energy

This may become one of the defining engineering challenges of the next decade.

The world wants more computing power while simultaneously needing to transition toward cleaner and more resilient energy systems.

These objectives cannot be separated.

Data centres will need power.

Factories will need power.

Electric transport will need power.

Cities will need power.

AI will need power.

The technology industry therefore cannot think only about software.

Digital infrastructure and energy infrastructure are becoming increasingly connected.

The organisations building the future will have to understand both.

And yet human needs remain remarkably constant

For all our technological progress, something interesting remains true.

In 2035, people will still need to eat.

We will still need somewhere to sleep.

We will still wear clothes.

We will still need homes.

We will still move between places.

We will still need healthcare.

We will still communicate.

We will still build businesses and communities.

Technology changes how these needs are fulfilled, but rarely eliminates the needs themselves.

Agriculture becomes more automated.

Homes become smarter.

Retail becomes more personalised.

Transport becomes increasingly connected.

Healthcare becomes more data-driven.

Financial services become more intelligent.

Engineering improves the systems around human life.

The underlying human requirements remain surprisingly durable.

Technology should ultimately serve something

That distinction matters.

Progress cannot simply be measured by how much technology we create.

The better question is what that technology enables.

Does it make financial services safer?

Does it help businesses become more productive?

Does it reduce waste?

Does it make infrastructure more resilient?

Does it help people access services more easily?

Does it create opportunities that did not previously exist?

Technology is most powerful when it becomes almost invisible — quietly enabling people and organisations to do things better.

Our view of 2035

We do not believe the next decade will be defined by a single technology.

It will be defined by the convergence of many.

AI.

Software.

Data.

Cloud.

Cybersecurity.

Payments.

Energy.

Networks.

Automation.

Physical infrastructure.

And the engineering disciplines required to connect them.

The world will need more code.

More storage.

More computing.

More infrastructure.

More energy.

And perhaps most importantly, more people capable of understanding how these systems should work together.

The tools will become dramatically more powerful.

Our responsibility is to use them well.

To serve. To solve. To evolve.