Real Insight in an Increasingly AI World

Why human contact and experience matter

We are increasingly automating both sides of the customer relationship.

Organisations are using AI to talk to customers through bots, automated marketing and generated content.

At the same time, the insight industry is increasingly using AI to listen to customers through automated moderation, translation, analysis, self-service research tools and synthetic data.

There are obvious benefits.

AI can make interactions faster, research cheaper and analysis more scalable. It can remove repetitive tasks and make sophisticated tools available to organisations that previously could not afford them.

The issue is not whether we should use AI. We should. The question is whether, as we automate more of the customer relationship and insight tools, we risk becoming less connected with the real customer and the insights that drive competitive advantage.

Genuine customer understanding is rarely found only in the obvious answer. There is still enormous value in sitting opposite someone and listening. Not simply recording the words they use, but noticing hesitation, tone and emotion. Seeing their reaction to an idea. Recognising when what they say does not quite match how they appear to feel.

There is value in observing customers interacting with a product or discussing a brand.

And there is particular value in being able to follow an unexpected response.

A good researcher can decide that the next question in the discussion guide is suddenly less important than understanding what a participant has just said.

That is often where research becomes insight.

We see this constantly in fieldwork across Latin America. A researcher sitting across from a small business owner in São Paulo or Bogotá notices the pause before an answer, not just the answer itself. That pause is often where the real finding lives, and no transcript or AI captures it.

We also see this in market sizing and market attractiveness analysis. AI can analyse vast amounts of data quickly, identify patterns and generate multiple market models in hours rather than weeks. But the quality of the answer still depends on the assumptions behind it and whether they reflect commercial reality. In one illustrative example, an AI-only assessment identified 100,000 potential users spending £2,000 a year and produced a perfectly logical market estimate of £200 million. Human validation then challenged the assumptions: only 35% actually met the target criteria, only half had the necessary budget authority, and realistic adoption was likely to reach around 20% within five years. The commercially achievable opportunity was therefore closer to £70 million. The AI calculation was not necessarily wrong; it identified the theoretical market. Human judgement helped identify the realistic one. That distinction can fundamentally change an investment or market-entry decision.

Efficiency versus discovery

AI is extremely powerful when we already know the question.

Summarise these interviews. Translate these responses. Identify the dominant themes. Compare these customer groups.

Those tasks can increasingly be completed faster and at lower cost.

But some of the greatest value from research comes from discovering that we were asking the wrong question in the first place.

That is a different challenge.

We had exactly this moment with a client recently. The client had one week to get answers and, understandably, turned to AI first. It gave them everything already published about their category, fast. What it couldn't give them was the one answer specific to their business, because that answer didn't exist anywhere yet. It only showed up once someone sat down with a real expert and asked. There is less value in knowing what everyone else knows and real value in uncovering something new. Here we also need to recognise that urgency also plays a part and can be counterproductive, especially when research becomes a tick box exercise. Organisational culture would be a whole new topic, suffice to say the attitude of the commissioning research organisation plays a significant part in the value they gain, recognise and act on.   

Competitive advantage often comes from recognising something that does not fit the established pattern:

·        A small change in customer behaviour.

·        An unexpected reaction.

·        A contradiction.

·        A need customers struggle to articulate.

·        A customer behaving differently from the segment we thought we understood.

These things may look statistically insignificant.

Commercially, they may be anything but.

What gets lost?

Translation provides a simple but effective example.

AI translation is now remarkably useful. It can provide an immediate understanding of research conducted in another language at a fraction of the traditional cost.

For an initial review, that can be enormously valuable.

But translation in research is not simply about getting the words technically correct.

• Was the participant enthusiastic or merely polite?

• Was something said ironically?

• Did a phrase carry a cultural meaning that does not translate directly?

• Was the participant uncomfortable with the question?

This is especially true across a region like Latin America, where Spanish in Mexico City is not Spanish in Buenos Aires, and Portuguese in São Paulo is not Portuguese in Lisbon. A phrase that reads as playful in one market can read as disrespectful in another. Translation gets the words right. It takes someone who has actually worked in that specific market to know whether the meaning survived the trip.

Most of the translation may be perfectly adequate, but the part that was missed or misunderstood can be critical.

This becomes even more important when we consider tools like AI moderation.

Moderation is not simply the delivery of a predetermined series of questions.

Good moderators probe. They challenge. They recognise hesitation. They change direction. They notice when something does not quite make sense.

Most importantly, they recognise when something unexpected deserves to be explored.

AI moderation will undoubtedly improve, and its advantages in terms of scale, speed and cost are significant. But there is a risk that increasingly efficient moderation produces increasingly efficient answers to the questions we already thought were important.

The same issue applies to AI analysis.

Automated analysis can process enormous volumes of information, identify common themes and summarise findings in minutes.

That is genuinely powerful. But again, insight is not simply the most frequently occurring theme. Sometimes the most commercially valuable finding is the outlier.

The contradictory comment. The customer whose behaviour does not fit.

The apparently minor observation that, combined with market experience, points towards something much bigger.

The valuable insight may sit in the 15% that is missed.

It might even sit in the 2%.

And that 2% could potentially be the source of competitive advantage.

Before we are shouted down with cries of 'AI models can be trained to do all of these tasks well'. Yes, they can, but more often it is the generic models that are being used and relied on.

A challenge for the insight industry

This creates a clear challenge for research and insight professionals.

If AI can help design questionnaires, conduct interviews, translate responses and analyse findings faster and more cheaply, then simply executing the research process is no longer enough to demonstrate value.

Our value increasingly has to come from somewhere else. From judgement. From curiosity. From understanding context. From recognising when something unexpected matters. From challenging an apparently obvious conclusion.

And, crucially, from connecting customer evidence to the commercial decisions an organisation needs to make.

In that sense, AI should raise the standard expected from insight professionals rather than lower it.

We should use technology to remove the work that does not require human judgement, while protecting the parts of the process where human judgement genuinely adds value.

A

t Advantage and AG3, this is exactly the balance we try to hold for clients: let AI carry the volume and the speed, and keep a real person in the room for the commercial understanding to recognise where the research changes a business decision.

That balance may well change as AI develops.

But for now, we remain convinced there is a very important place for direct human interaction with real customers and real world commercial acumen.

And in an increasingly automated world, that may become more valuable, not less.

A shared perspective

This article was prepared jointly by Advantage Market Intelligence and AG3 Consulting, drawing on many years of experience working with clients, customers and research participants across the UK, Europe and Latin America. While our businesses operate in different markets and bring different perspectives, we share the same view: AI is creating enormous opportunities for the insight industry, but its greatest value comes when it is combined with human curiosity, commercial judgement and genuine engagement with customers. The future of insight is unlikely to be human or AI. We believe the strongest results will come from knowing where each adds the most value.

About Advantage Market Intelligence

Advantage Market Intelligence is a UK-based, senior-led market research and market intelligence consultancy helping organisations make better growth and investment decisions. With more than 30 years' experience in research and market intelligence, Advantage combines customer research, market analysis, competitor intelligence and commercial judgement to help clients understand markets, assess opportunities and turn evidence into practical decisions.

Website: https://www.advantagemi.co.uk/Email: kevin@advantagemi.co.uk

About AG3 Consulting

AG3 Consulting is an independent full-service market research and competitive intelligence agency headquartered in Brazil, with local teams supporting research across Latin America. Established for more than a decade, AG3 works with international clients on both B2B and B2C research, including market entry, customer experience, product testing, brand research and store audits.

Website: https://www.ag3consulting.online/Email: geisa@ag3consulting.com.brTel: +55 48 99616 2060