Insight » Beyond the prompt: keeping judgement at the heart of consulting

Beyond the prompt: keeping judgement at the heart of consulting

Date published: March 2025

When news broke recently in Australia that a consultant-produced report contained AI-generated sections with factual errors, the consulting world took notice. The report, prepared for a government department, included AI invented references and misleading citations – a by-product of an artificial intelligence tool that sounded confident but lacked understanding.

The incident was more than a technical error. It was a warning – a reminder that technology’s outputs, however polished, are only as sound as the human judgement applied to them. In an age where words can be generated in seconds, the line between efficiency and accuracy, between assistance and abdication, has never been thinner.

What we mean by ‘AI’ today

Much of what we call AI is, in practice, machine learning – large language models trained on vast data sets drawn from the public internet. These models learn patterns, not truth. Their behaviour is shaped by the information they ingest and the prompts or documents we feed them at use-time. If those data streams contain inaccuracies, gaps or deliberate misinformation, the outputs tend to reflect that.

Unlike experienced practitioners, models do not apply scepticism to their inputs. They do not pause to ask whether a source is credible, whether a trend is spurious, or whether a convenient ‘fact’ clashes with lived experience. People do this instinctively – judgement and education teach us to probe, triangulate and, when needed, to doubt. Models predict; humans interrogate. That distinction matters.

The seduction of certainty

AI has become astonishingly capable. It can write, summarise and simulate with fluency that was unimaginable only a few years ago. For busy professionals, the allure is obvious – the promise of time saved, patterns revealed, insights accelerated.

But AI’s fluency can be deceptive. The technology does not know what will land well in a community workshop, a stakeholder meeting or a boardroom. It cannot read a mood, anticipate a backlash, or understand how a message might resonate differently in London than in Leeds, or in Sydney than in Singapore.

These systems do not know; they predict. They assemble words that appear right, not necessarily those that are right. The result can be confident nonsense – persuasive, well-structured, and completely wrong.

For consulting – a profession built on evidence, interpretation and trust – that distinction matters profoundly. When we accept AI’s outputs uncritically, we risk replacing discernment with convenience, and reflection with automation.

Why consulting cannot be automated

At Hood & Woolf, we see AI as a powerful assistant. It can help us analyse data, spot themes and streamline processes. It can challenge us to think differently. But it cannot read a room, sense nuance or understand the emotional dynamics that shape effective communication.

Consulting, especially in communications and engagement, is as much about empathy and instinct as it is about intellect and information. It requires human judgement– the ability to recognise nuance, to interpret emotion, to weigh the unspoken against the explicit. These are skills that no algorithm possesses.

AI can tell you that 60% of stakeholders expressed ‘concern’, but it cannot tell you whether that concern stems from fatigue, fear or frustration. It can draft a statement, but it cannot sense when silence says more than words.

AI can replicate text but not trust. It can mimic insight but not meaning. It has no memory of what it feels like to stand before a sceptical audience, to rebuild confidence after failure, or to persuade through shared understanding.

A shared responsibility

This is not a rejection of technology, but a call for prudence. The responsibility does not rest only with consultants, but with all of us who use or rely on AI-generated work. The risk lies less in the tool itself and more in the quality of the data that trains it – and in our willingness to believe those outputs unthinkingly.

Blind trust in machine outputs – however elegant or efficient – erodes the very quality that makes professional judgement valuable: the ability to question, to test, to doubt.

AI has no sense of cultural nuance, institutional memory or the political landscape that often shapes public and organisational communication. It might not tell you that a phrase was used in a failed campaign ten years ago, or that a particular narrative will fall flat with your stakeholders. It cannot warn you when something sounds right but is not wise.

Those are human insights – hard-earned, intuitive and irreplaceable.

In the end, AI is becoming an indispensable collaborator, but it is not a substitute for human reasoning. Our role is not to resist its use, but to ensure that every use is guided by awareness, context and care.

Keeping humanity at the centre

The future of consulting will undoubtedly continue to be shaped by technology – but its soul will remain human. Success must not be measured by how quickly we can generate words and analyse data points, but by how deeply we can understand and use them.

As this new era unfolds, we have an opportunity to show what responsible, human-led consulting looks like: thoughtful, bespoke and rooted in listening.

Technology may accelerate the journey, but only people can give it direction – and ensure it leads somewhere meaningful.

Because in the end, no matter how sophisticated the system, only humans can provide true judgement.

Loading...