The test drive is a blind spot in automotive sales. What could AI help us see?

By Simon Knocker

We can record when a customer enquires, how quickly we respond, which vehicle they view and whether they take a test drive.

But ask what happened during that drive and the picture often becomes much less clear.

What did the customer discover? Which concern remained unresolved? Did the sales executive connect the experience to what mattered to that individual? Did the customer return feeling more confident about their decision?

Too often, a rich and potentially decisive experience leaves behind little more than a completed activity in a lead management system.

That is a substantial gap in our understanding of the sales process—and an interesting place to challenge how we think about AI.

A perspective shaped by people, processes and technology

I have spent around 30 years working in automotive, much of it examining how dealership processes, technology and people combine to shape customer experience and business performance.

That work has included co-designing and building AudiCam, an early aftersales workshop video solution, supporting the introduction and deployment of lead management systems, and working on changes to sales journeys and retail operating models.

More recently, my work has extended into behavioural development, learning and coaching with sales executives, business managers, transaction managers and sales managers.

Across those experiences, one lesson keeps returning: the value of a system depends on what people do with it. Technology can prompt an action and record its completion. Understanding the quality of that action takes us much further.

The test drive brings that distinction into sharp focus.

The evidence says the experience matters

In research published in 2024, CDK Global surveyed 1,048 new car buyers. Some 78% said the test drive ultimately sold them on the vehicle they purchased, while 54% identified the opportunity to test drive as their leading reason for buying at a dealership rather than online. These are buyers’ reported perceptions, rather than proof that a particular test drive approach causes a sale. Nevertheless, they underline the importance customers attach to the experience. CDK Global

There is a strong UK signal too. Cazoo’s June 2026 Consumer Insight Panel, polling 2,008 car buying decision makers, found that 72% wanted to view, test drive and buy their next car at a dealership. A further 21% wanted to view and test drive at a dealership before purchasing online. Cazoo: Used Vehicle Marketplace for Dealers

My reading is that the physical experience remains central even as more of the surrounding journey becomes digital.

Which raises a question: if this moment matters so much, how deliberately are we designing it?

Two routes are a limited definition of personalisation

In my experience, a familiar answer when discussing test drives with dealerships is: “We have two routes.” Sometimes three.

Usually there is a shorter drive and a longer one, perhaps with a dual carriageway option.

There are sensible operational reasons for knowing the local roads, journey times and suitable places to demonstrate a vehicle. The limitation comes when those routes become the default definition of the experience.

Choosing the longer route because the customer has more time, or the dual carriageway because they want to try higher speeds, offers some choice. There is scope to go much further.

The best sales executives already do. They use what they have learned during qualification to shape the drive around the customer’s life, priorities and uncertainties. Their choice of route follows a purpose.

A customer concerned about moving into a larger car may need to experience narrow streets, visibility and parking. Someone who spends hours travelling for work may want to understand seat comfort, road noise and how settled the car feels on a sustained journey. A first-time EV buyer may need space to become comfortable with regenerative braking and a practical explanation of charging, delivered while safely parked.

Those customers could drive the same model and need quite different experiences.

A useful starting question is:

“What would you need to experience today to know whether this car works for you?”

Design experiences that can be joined together

Imagine a dealership developing a repertoire of experiences that can be combined into a coherent journey:

• A quiet stretch to build confidence with an unfamiliar vehicle.

• Urban roads to experience visibility, manoeuvrability and everyday driving.

• A suitable car park to try parking and assistance features.

• A dual carriageway section to assess refinement at cruising speed.

• An appropriate uneven road to understand ride comfort.

• A safe stopping point to explore controls or features without distraction.

• A practical check of access, luggage space or family requirements.

• Where feasible and agreed, a visit home to assess driveway access or garage fit.

The route becomes the means of connecting the experiences that matter. Local conditions, available time and safety still shape what is possible.

Consider two customers interested in the same SUV. One is worried about parking at home; the other wants a more comfortable motorway commute. Sending both around the same loop may demonstrate the vehicle. Designing around their different questions helps each judge whether it belongs in their life.

That is where good qualification should become visible in the experience.

Research also gives us reason to recognise existing strengths. In another report from its 2024 buyer study, CDK found that 94% of respondents said dealership staff assessed their needs, and 90% felt recommended vehicles met their wants and needs. CDK Global

The opportunity is to examine how consistently that understanding carries through into the drive—and how we help more people develop the judgement that the strongest sales executives already apply.

The most revealing conversation may be the one we barely capture

During a test drive, customers can move beyond the language of specifications.

A question about boot capacity becomes a conversation about a dog, a wheelchair or a family holiday. A question about range turns out to be concern about visiting an elderly parent. An interest in parking technology reflects anxiety about a tight space at home.

These exchanges explain what the product means to the person.

A skilled sales executive hears them, explores them and adapts. But how much survives afterwards? How much informs the next conversation? How much becomes useful feedback for that executive, their manager or the wider business?

We may record that parking assistance was demonstrated. We may lose the reason it mattered.

This is where the potential richness of AI insight becomes particularly interesting.

AI in automotive sales: An innovation concept we have explored

Working with Join Business Consulting and DT-Squad, we have explored and prototyped a test drive tool as part of our innovation work.

The concept asks what could become possible if, with appropriate permission and safeguards, we could bring together the conversation during a drive, relevant journey context and the customer’s stated priorities.

https://video.wixstatic.com/video/d19103_2cee725daf2042a29bd7d1239e6e1a8f/1080p/mp4/file.mp4

Could AI help identify the questions asked, the answers given, the concerns left open and the moments worth revisiting?

Could it help an executive reflect on whether the experience matched the qualification? Could it support a more useful debrief and more relevant follow-up?

These are possibilities to investigate through careful trials. The prototype is a way of exploring the idea; it does not establish that the approach improves sales or customer satisfaction.

What interests me is the opportunity to make an otherwise fleeting experience available for reflection and learning.

For example, a customer might ask about charging on longer journeys, receive a general answer, then return to the subject later. A useful analysis could flag the repeated question and direct the executive to review the exchange.

That creates a practical next step: check understanding, establish what remains uncertain and provide a relevant answer.

A learning loop for the sales executive

Think about the feedback available to someone developing their test drive skills.

A manager can observe some showroom conversations. Accompanying every test drive is much harder. Coaching therefore often depends on the executive’s recollection and the eventual outcome.

AI could help create a more continuous learning loop.

An executive could review whether they explored a customer’s question fully, linked demonstrations to stated needs, allowed space for the customer to speak, or missed an opportunity to check understanding.

Over several drives, patterns might emerge. Perhaps the executive explains technology confidently but regularly moves on before the customer has understood. Perhaps qualification is thorough, yet the resulting drive rarely reflects it.

Those are useful coaching conversations because they concern specific behaviours that someone can practise.

The quality of the feedback matters. A ratio of speaking time is a prompt to investigate, rather than a universal measure of good selling. Silence might mean comfort, concentration or uncertainty. Context and human judgement remain essential.

Rich insight needs intellectual discipline

There is a temptation to stretch the idea further: analyse a face or a voice and declare exactly what a customer feels.

That deserves scrutiny.

A major review by Lisa Feldman Barrett and colleagues, published in 2019, found that facial movements do not provide sufficiently reliable and specific indicators of emotion across people and contexts to support simple diagnostic interpretations. PMC

A smile does not prove purchase intent. A quiet customer is not necessarily disengaged.

The useful discipline is to distinguish what happened, what someone explicitly said and what an AI system has inferred. A suggestion that an exchange needs attention can support a good question. Presenting an inferred emotional state as fact could undermine the very understanding we are trying to improve.

I would apply the same care to commercial insight. Repeated customer questions could help a retailer or manufacturer identify topics worth investigating, from feature explanations to product concerns. They would still reflect the customers, vehicles and circumstances captured, rather than automatically representing the whole market.

Personalisation includes how people want to experience the car

We should challenge the assumption that every customer benefits from the same level of accompaniment.

CDK’s test drive research reported that 81% of buyers who drove alone said the drive sold them on the vehicle, compared with 77% who drove with a salesperson. This comparison does not establish that solo drives are more effective: the customers and circumstances may differ. CDK Global

It does reinforce the need to discuss preferences.

Some customers value guidance. Others want privacy to think, concentrate or discuss the car with a partner. A tailored approach should accommodate those differences, including a straightforward choice about participation in any recording or analysis.

Trust belongs in the design from the beginning. Customers and colleagues should understand the purpose, what would be captured, who could access it and how long it would be retained. Any real-world trial would need appropriate privacy, employment and data protection review.

A tool intended to support better understanding has to earn people’s confidence.

What are we prepared to reconsider?

The immediate opportunity is already available to every dealership: ask what the customer needs to experience, shape the drive around it and use the debrief to establish what they learned.

AI opens a further question.

What might improve if we could learn systematically from the quality of those experiences? Could we make strong practice easier to share, coaching more specific and follow-up more relevant? Could we uncover customer questions that our existing measures barely register?

That is the challenge behind our exploration with Join Business Consulting and DT-Squad.

Automotive retail has spent years improving the visibility of its sales process. There is still considerable value to discover in understanding the experiences inside it.

Where else have we accepted a blind spot because we have become accustomed to measuring completion?

If that question sparks an idea—or you see opportunities to explore this thinking in your own business—I would love to speak.