By Kristiyana Raycheva, Data Analytics & AI Consultant, Stanton House

For years, technical excellence in Data & Analytics was often measured by a person's ability to write complex SQL, build sophisticated data pipelines or solve coding challenges quickly and independently.

Today, that definition is starting to evolve.

AI-powered development tools are becoming embedded within everyday technical workflows. Data Engineers and Analytics Engineers can now use tools such as Claude, Copilot and other AI assistants to generate code, refactor existing solutions, document pipelines, troubleshoot issues and accelerate repetitive development tasks.

The conversation I am having with Data leaders, however, is rarely about whether AI can write code.

That question has largely been answered.

The more interesting question is this:

When AI can increasingly produce the first draft of the technical work, what becomes more valuable instead?

For Heads of Data, Analytics leaders and Chief Data Officers, the answer has important implications for hiring strategies, team design and technical assessment. For Data professionals themselves, it may determine which skills define long-term career success.

AI Can Write Code. That's No Longer the Interesting Question

In one recent conversation, a Head of Data Science at a media organisation described how their team was using Claude to improve coding and development efficiency. Their experience was that the tool provided useful assistance more than 70% of the time, helping accelerate routine work and reduce time spent on repetitive tasks.

That experience aligns with broader industry research. Microsoft's research into AI-assisted software development found that developers using GitHub Copilot completed coding tasks significantly faster than those working without AI assistance, while additional field experiments across Microsoft and Accenture identified measurable improvements in developer productivity.  

Yet what stood out from my conversation with that Data Science leader was not the productivity gain itself.

Their conclusion was not that coding expertise was becoming unnecessary. In fact, they emphasised exactly the opposite.

The team still needed strong foundational coding and mathematical knowledge because somebody had to decide whether the AI-generated output was correct, efficient and fit for purpose.

This is where much of the public discussion around AI misses the point.

The question is not whether AI can produce code.

The real question is whether the person using AI has the knowledge required to challenge, validate and improve what it produces.

The Skill Is Shifting from Creation to Judgement

Historically, one marker of a strong technical candidate was their ability to produce high-quality SQL or Python quickly and independently.

AI changes that equation.

A Data Engineer can now generate sophisticated code within seconds. However, generating code is only one part of the job. Someone still needs to determine whether the logic is correct, efficient and secure; whether it will perform at scale and fit the organisation’s architecture; whether it is using the right data; and, ultimately, whether the output makes sense in the context of the business problem.

In many respects, AI is increasing the value of technical judgement rather than reducing it.

A professional who can generate code but lacks the knowledge to interrogate it introduces risk.

A professional who understands the underlying principles, uses AI to accelerate execution and applies rigorous judgement to the output becomes considerably more productive.

This distinction is important because AI assistants do not remove the need for expertise. They amplify the impact of those who already have it.

McKinsey's research into AI-assisted software development found that generative AI can significantly accelerate activities such as code generation, documentation and refactoring, while also noting that productivity gains vary depending on task complexity and developer experience. The organisations that benefit most are those that combine new tools with strong technical capability and effective upskilling.

As AI becomes more capable, understanding what good code looks like may become just as important as being able to write it from scratch.

The Bigger Opportunity Is Moving Data Professionals Closer to the Business

While much of the AI discussion focuses on technical execution, I believe the larger opportunity sits elsewhere.

If AI reduces the time spent on repetitive development work, what should Data professionals do with the capacity it creates?

For many organisations, the answer should be greater commercial engagement.

The most valuable Data Engineers and Analytics Engineers are increasingly those who understand not only how to build something, but why it needs to be built in the first place.

That means understanding the commercial priorities behind the request.

A technically strong Data Engineer might successfully build exactly what a stakeholder asks for.

A more commercially minded Data Engineer may start by asking:

"What decision are you actually trying to make?"

That conversation might reveal that a completely different solution would deliver a better outcome.

Are We Still Assessing Data Talent for the Way They Worked Five Years Ago?

This shift creates an important hiring challenge.

One trend I am seeing among candidates is growing scepticism towards traditional live coding assessments. Some professionals question why recruitment processes prohibit the use of AI tools when those same tools form part of the everyday working environment they are being hired into.

This does not mean technical testing should disappear.

Employers still need confidence that candidates understand SQL, Python, data modelling, architecture and engineering fundamentals.

However, it may be time to rethink what technical excellence looks like during an interview process.

If AI will continue to be part of day-to-day development, employers should consider assessing a candidate's ability to work effectively with AI, not simply without it.

For example, instead of asking a candidate to produce code from scratch, organisations could provide AI-generated code and ask:

·       What is wrong with it?

·       What assumptions has it made?

·       What risks can you identify?

·       How would you validate the output?

·       Would you deploy it into production?

·       What additional context would you need?

Similarly, candidates could be presented with an ambiguous business problem and asked how they would gather requirements before proposing a solution.

These assessments measure judgement, problem-solving and communication, capabilities that increasingly reflect how modern Data teams operate.

The Best Data Candidate May Now Look Different

The hiring criteria that defined great technical talent five years ago may not be sufficient for the next five.

Historically, employers often prioritised:

Technical depth + years of experience + tool proficiency

Increasingly, the strongest candidates combine:

Technical foundations + AI fluency + critical judgement + problem-solving ability + commercial understanding + communication skills

This does not mean every Data Engineer needs to become a management consultant.

It means technical excellence increasingly includes understanding the context in which technology is being applied. For Heads of Data, that creates several important questions:

·       Are we over-indexing on technical tasks that AI can increasingly augment?

·       Which technical foundations remain genuinely essential?

·       Are we assessing candidates' ability to validate AI-generated work?

·       Are we testing problem-solving capability or simply memory?

·       Do our teams understand the commercial context behind their work?

·       Do our interview processes reflect how our teams actually operate?

The answers may shape the next generation of high-performing Data teams.

Final Thoughts

AI is not making great Data professionals less valuable.

What I am seeing instead is a shift in where their value is created.

As technical execution becomes faster and more accessible, the premium is moving towards judgement, validation, business understanding and communication. Strong technical foundations remain critical, but they are no longer the only differentiator.

If you're reviewing your Data, Analytics & AI hiring strategy, rethinking technical assessments or considering how AI is reshaping the skills your team needs, I'd be delighted to discuss what we're seeing across the market and share insight from ongoing conversations with Data leaders and candidates.

Sources:

1.       Microsoft Research, The Impact of AI on Developer Productivity: Evidence from GitHub Copilot (2023). [microsoft.com]

2.       Microsoft Research, The Effects of Generative AI on High-Skilled Work: Evidence from Three Field Experiments with Software Developers (2025). [microsoft.com]

3.       McKinsey & Company, Unleashing Developer Productivity with Generative AI (2023). [mckinsey.com]