AI in analytics: Where it creates value and where human expertise still wins
Joe Dreyer, Data Transformation & Enablement Specialist at PBT Group
Artificial intelligence (AI) has changed how organisations work with analytics. It can analyse large volumes of information quickly, support faster reporting, generate useful starting points, and help teams explore patterns. Properly used, it can strengthen analytical capability and reduce the time between a question being asked and an insight being available.
That does not make AI magic. It does not remove the need for data quality, sound judgement, governance, or human accountability. In analytics, the risk is not only that AI may be wrong. The greater risk is that its output can sound convincing enough to be trusted before it has been properly questioned.
Analytics has always depended on interpretation. A dashboard, report, model, or data product does not create business value on its own. Value comes from whether the insight is accurate, whether the context is understood, and whether the people using it know what decision it should support. AI can improve parts of that process, but it cannot own the meaning behind the work.
AI usually adds value when the task is clear, the data is reliable, and the input and output can be tightly defined. It can help with repetitive analysis, anomaly detection, summarising information, identifying patterns, and generating options. In these areas, AI can act as an assistant to the analytics team, giving skilled people more room to focus on interpretation and validation.
Analytical thinking still matters
The problem starts when organisations treat AI as a replacement for analytical thinking. In strategic, financial, regulatory, or high-risk environments, output cannot simply be accepted because it was produced quickly or because previous responses looked accurate. Conditions change, data changes, and models can produce misleading answers, which is why review, auditability, and governance remain essential.
One common mistake is applying AI to work that still depends heavily on memory, meaning, context, and perspective. AI can process information, but it does not understand an organisation in the way experienced people do. It does not know the history behind a decision or the consequences of acting on the wrong signal unless those factors are made available to it.
Another risk is the weakening of critical thinking. If teams accept AI-generated outputs without interrogating them, they may lose the habit of asking better questions. The quality of the answer depends on the quality of the question, the framing of the problem, and the knowledge used to test whether the result makes sense.
Human-driven process
Human expertise, therefore, becomes more important, not less. Analytics teams still need technical competence, but they will also need curiosity, business understanding, communication skills, and the confidence to challenge an output that appears plausible. People need to know how to work with AI without allowing it to dictate the work.
Prompting is part of this shift, although it should not be reduced to a technical trick. Knowing how to ask the right question reflects the user’s understanding of the business problem. A vague request will produce a generic answer. A better question, shaped by domain knowledge, gives AI a stronger frame.
Trust is a cultural issue
For organisations, the issue is also cultural. AI adoption depends on trust, and trust depends on whether people understand why the technology is being used, how it affects their roles, and where accountability sits. If employees believe AI is being used to replace them or capture their knowledge without a clear future for their own development, adoption will suffer.
This is especially relevant when businesses consider junior roles. Removing junior people from the learning chain may appear efficient in the short term, but it can weaken organisational knowledge over time. Senior expertise develops through exposure, review, mistakes, and mentorship. If AI bypasses that path, organisations may have fewer people able to judge whether the output is right.
Accountability cannot be delegated to AI. A system can produce an answer, but it cannot carry responsibility for the decision that follows. People remain accountable for how AI is designed, governed, interpreted, and used. That becomes more important as AI moves closer to analytics-led decision-making.
Combining strengths
The real opportunity is not to place AI above human expertise, but to combine it with organisational intelligence. That means aligning people, process, systems, data, and knowledge so that AI has the right context and humans remain close enough to guide, question, and validate its output.
AI will continue to reshape analytics, but its value will depend on how carefully organisations apply it. It is useful where it accelerates work, expands analytical reach, and supports better questions. It creates noise where it is trusted without context, used without governance, or allowed to weaken judgement. The organisations that benefit most will use AI to strengthen human expertise rather than replace the thinking that makes analytics valuable.
