Ground Truth Advisory · The operational truth

Everyone is flat out. Nobody is productive.

Ask anyone in a dealership how the day went and you will hear the same three words. They are not lying. But "flat out" is not a measure of productivity — it is what happens when demand arrives faster than anyone can triage it. And it is precisely the state in which the highest-value work gets dropped: the pre-call, the qualification, the follow-up, the customer sitting alone.

Flat out is the symptom. It is then used as the excuse for the outcome it caused. The question worth paying for is how you turn flat out into productive — and the answer starts by admitting what "flat out" is actually costing you.

The method

Most AI programmes hand a dealership thirty-three use cases and call it a strategy. Nobody adopts thirty-three of anything. Three moves can be trained, tracked, and held in a Head of Business's head. The moves are the same in every department. Only the content changes.

MOVE 01

Capture

Stop losing what you already know. Every role in this building generates information that dies — in someone's head, on a notepad, in a conversation nobody wrote down.

The information already exists. You are simply not keeping it.

MOVE 02

Draft

Never start from blank. The blank page is where the delay lives — and in this business, delay is where the deal, the booking and the goodwill die.

Not "AI writes it." AI starts it. A human finishes it and owns it.

MOVE 03

See

Turn what you captured into something you can act on today — not something you read about in next month's pack, when it is already someone else's problem.

Most dealership reporting tells you what happened. This tells you what is happening.

AI does not fix a broken process. It exposes it — faster, at scale, and in front of the customer.
— Which is why the operating model gets read first, not last.

Worked example · Reception, 08:50

What "we need AI on the phones" actually means.

At any given moment, she is holding:

  • A caller she is trying to put through to a department that is not picking up — not because they are with a customer, but because they have decided that whatever is on their desk outranks a stranger on a phone.
  • Two more lines ringing, which she can see.
  • Three customers at the desk, standing five minutes, unacknowledged — because every colleague who walks past presumes they are waiting for a different department.
  • A transport driver with three new cars and a timed drop.
  • A service customer who needs to leave their car and get to work.

Then she has to take the message. And then decide: does the message go first, or the queue that is now standing in front of her?

The read

There is no AI product on the market that fixes this. Buy the smartest call handler in the world and she still has three people at the desk, a timed delivery, and a colleague who won't pick up.

Because this is not a technology gap. It is an operating model with no owner at the point where every customer arrives. No triage rule. No escalation. No definition of what outranks what. And a cultural agreement, never spoken aloud, that the phone is somebody else's problem.

What AI will do — and this is worth buying — is prove it. Ring counts. Extensions that don't answer. Abandonment. Time-at-desk before acknowledgement. It turns "reception is chaos" into a number with a name on it.

Then somebody has to be willing to say the quiet part out loud. That part is not a software licence. It is a leadership decision — and it is the one I am actually here for.

Worked example · The parts window, 08:15

Every department's failure, arriving at the same counter.

One adviser. Simultaneously:

  • The phone is ringing — almost certainly someone price-checking against a factor, who will order elsewhere the moment they are told "I'll call you back."
  • A customer at the trade counter, who came in person, specifically, for one thing.
  • A queue of technicians at the window. They are clocked on. Every minute they stand there chewing the fat is idle time the workshop is paying for — and it appears on no report anywhere.
  • The parts delivery that has to be booked in first thing, or the whole day runs behind it.
  • A job that was pre-picked badly, or ordered wrong — so the technician is back for a second time, and that car will not be ready today.
  • A sales executive who forgot to push the accessory line through, has just remembered, and the car goes out in three hours.

The read

Ask which of those comes first and you will get six different answers from six different managers, each of them certain, each of them protecting their own number.

And here is the one nobody asks: what pricing autonomy does that adviser actually have? If she cannot close the call on the call — genuine part, this warranty, in stock, collect today — she has to promise a call back. The order is gone. Not lost on price. Lost on speed, and on the confidence she was never given permission to have.

Then the chain. Wrong part to the job. Car not ready. Service carries the customer. CSI takes the hit. The failure began at the parts window and is recorded against a department that never touched it.

AI will not resolve a queue that has no priority rule. What it will do is cost it — idle technician minutes, abandoned calls, pre-pick failure rates, and the exact hour the day fell over. It will hand you the number.

Whether anyone is willing to act on that number is not a software question.

The pattern

Reception, the parts window and service reception are the same problem wearing different clothes. Each sits at a junction where every department's demand arrives at once. Each is held by a person with no authority to decide what outranks what. Each absorbs failures that originate somewhere else and get recorded against someone else.

Every dealership in the country has these junctions, and not one of them has a written rule for what happens when all five things land together. It is managed by whoever is standing there, on instinct, badly, every single day — and then it is called a training issue, or a CSI issue, or a people issue. It is none of those. It is a design issue, and nobody designed it.

And every person standing at one of those junctions will tell you, honestly, that they are flat out. They are. That is exactly the problem — because flat out is the state in which the pre-call, the qualification and the follow-up are always the first things to go. The work that would have prevented tomorrow is sacrificed to survive today.

Then there is the other half, and it is not structural at all. Month end is not a surprise. The behaviours around it are not unknown. They are permitted — because the number arrives, and the cost of it lands on somebody who is not in the room when the number is celebrated.

Put AI on top of that and you will have measured the chaos beautifully.

Worked example · Service reception, 08:00 to checkout

One morning. And it started the afternoon before.

This is not a list of problems. It is a single chain, it begins the day before, and it runs until checkout. Follow it — and notice that every turn of it ends at the same person, who has no authority to fix a single link above her.

  1. Yesterday
    16:30

    The pre-call that didn't happen. Tomorrow's customers should have been called today — what to expect, how long, courtesy car, are they waiting, is it a lease, will they need authority. It is the first thing dropped when the department is busy. So the department is too busy today to make the calls that would make tomorrow less busy. Every question that should have been answered yesterday will now be asked in person, in a queue, at eight o'clock.

  2. 08:00

    The queue forms, and it is not one queue. A routine service — is she waiting, does she have a courtesy car, is it out front, is it charged, is there fuel in it, does she know how long? A lease vehicle whose driver does not know we cannot start until the leasing company authorises. A customer who has to make a train. A service plan holder who assumes it's all covered, and one who assumes the same thing and is wrong.

  3. 08:05

    "You're seeing Paul today. He's with a customer — take a seat." The customer has never met Paul. There is a free adviser standing in front of him. He is being told, politely, that our internal allocation matters more than his morning. He sits down. Nobody offers him a drink, because the 08:00 madness has already swallowed the person who would have.

  4. 08:20

    The third-visit customer, and the diagnosis that dies in your head. He is frustrated, and he is telling you exactly what is wrong with the car — when it happens, what it sounds like, what he was doing. You are listening while the queue builds behind him and the phone goes. You will remember it. You will not write it down. It reaches the technician as three words on a job card, and it will come back a fourth time.

  5. 08:40

    The keys weren't tagged properly. The technician starts the wrong car. The right car is with the valeter and cannot be touched. The job that was booked in has not begun.

  6. 10:00

    The waiting customer whose MOT was promised for ten. It has not been started. She is now standing in front of you, and she is right.

  7. 10:05

    Reception rings. Yesterday's customer has not had a call to say their car is ready. So they rang the switchboard. So reception is now ringing you — while you have a customer in front of you who was promised ten o'clock. Their call was never made because of the queue that formed at eight, because of the pre-calls that were never made yesterday afternoon.

  8. 14:00

    The video went out. Nobody followed it up. Red and amber work, identified by a technician, sent to a customer, and then nothing — because you were flat out with the morning you have just read. The work is not declined. It is simply never asked for.

  9. 16:30

    Checkout, and the same madness again. A queue, and a payment. Did we charge what she thought she was paying? The follow-up appointment does not get booked. The call to the third-visit customer, to check the fault is actually fixed, does not get made.

  10. +3 days

    The survey. Which measures the last three minutes of a chain that broke yesterday afternoon, in a quiet moment that was never quiet. And the score lands on the department that was carrying it, not the one that dropped it.

The read

There is no product you can buy that fixes this, because it is not eleven problems. It is one problem with eleven symptoms, and the root of it is a phone call that should have been made yesterday — one that nobody owns, that everybody agrees is important, and that is always the first thing dropped when the department is busy. Which it is. Because the calls didn't get made the day before that, either.

What AI does here is real, and it is not automation. It captures the diagnosis the customer gave you while you were under pressure, so it reaches the technician in his words rather than in your memory. It drafts the pre-call so it actually goes. It follows the video up so the red work gets asked for. And it shows you, at 10am rather than at month end, that the day has already fallen over.

But it will only work if somebody first admits that the pre-call is not optional. That is not a licence. That is a leadership decision — and it is the one nobody makes, because admitting the pre-call matters means admitting how many mornings have gone exactly like this one — and how many of them were avoidable from a desk, the afternoon before.

Worked example · The Business Manager

A role that quietly stopped existing, while everyone kept reporting it.

What it was

A private office. The customer taken in, properly, and given thirty minutes. How are you looking to fund this? What works for you? What are you actually hoping to spend, and what will this car really require?

And those thirty minutes were load-bearing in two directions at once. While the customer was in that room, their part-exchange was being appraised properly and priced well. The customer wasn't waiting. The conversation created the time.

What it is

A desk. An office they rarely leave. A queue of sales executives huddled around it waiting for a quote — after they have already queued for the used car manager to price the car.

An offer sheet is printed. The exec carries the figures back to a customer who has been sitting alone, and who will never meet the person who priced the next four years of their life.

Ask ten Business Managers what the job is and you will get ten answers. That is not a recruitment problem. That is a role with no definition.

The questions nobody asks

The customer has been in the showroom twenty minutes. How do we know what they want? How do we know this quote is the right one? Have we listened, or have we started pricing because pricing is the thing we know how to do?

Nobody checks the appraisal against the actual car. We take the description and the photographs as given, and we find out later.

The customer who has come in to sign — do they come before or after the customer in the showroom who is still deciding? Nobody has ever written down the answer.

Every customer who took no product: was a single one of them re-solicited by a manager? When the exec presented the full pack, did the customer understand what the products were for — or did we spend longer selling the products than we spent selling the car?

And the pre-handover, which is where the joy gets killed: products set up in advance? Re-solicited on collection? Pre-called about the valet? Pre-called about the right ID? A pre-handover appointment to do the paperwork and take the payment, so that handover is nothing but the exciting part?

Almost always: no, to every one of them. And nobody is lying about why. They are flat out.

Worked example · Month end

The most predictable crisis in the business. It destroys us every time.

It arrives on the same date every month. It has never once been a surprise. And yet every month, from about the 25th, the operation abandons everything it has been trying to build and reverts to whatever will get the number over the line. That is not a capacity problem. It is a business that has decided not to prepare for a thing it knows is coming — because preparing would mean confronting the behaviours that make it necessary.

16:00, the last working day

The sales administrator is in tears. Her workload has tripled in an afternoon, and there are two hours left.

Every car has to be taxed. That is time-bound. It cannot slide. And it is not even the whole of it — the handovers have to be on the system, the registrations have to show, and every one of those sits with the same person while the clock runs down.

Around her: a certificate that was never chased. Accessories that were never fitted. A software update pushed at the last minute. Paperwork the customer could have signed days ago, if anyone had asked them to come in. A new car that wasn't supposed to land this month and just turned up — and an exec who needs it.

None of this was unknowable. All of it was avoidable. Every single month.

The quiet part

An executive has hit their target. They have just sold a car. They could get it into the month.

They would have — if they needed the number. They don't. So next month starts at one instead of zero, and the unit is quietly held.

The sales manager needs it. The exec says the customer would rather collect next week. And how hard does that actually get pushed back?

The month is not decided by demand. It is decided by where each executive personally sits against their own target.

The site's performance is the aggregate of eleven individual commission positions. Everybody in the building knows this. Nobody says it. And the administrator cries at four o'clock because of it.

What AI does, and where it stops

You could see this. In real time, on one screen: units sold but not delivered, by executive, against where that executive sits on their own target. Deliveries that wobbled out of the diary and the reason given. Every avoidable task — the certificate, the accessory, the signature, the tax — flagged with days to spare rather than hours.

The technology is trivial. The number would be on a Head of Business's screen by the 20th.

And then it is the same question as the handover alert, and the same answer. Will anyone act on it? Because acting on it means telling an executive who has hit target that the unit goes out this month regardless — which is a conversation about culture, incentive and courage, and not one that any software vendor is able to sell you.

That is why month end never improves. Not because it is hard. Because fixing it requires somebody to say the quiet part out loud, in a room, to people who would rather it stayed quiet.

And it is not only sales

In aftersales, the money has already been spent. The race is to collect it.

A unit is either delivered or it isn't — binary. But in aftersales the work is already done. The hours are turned. The parts are consumed. The cost is sitting in your P&L. What is outstanding is only the invoice — and that is the single thing that decides whether any of it counted.

  • The WIP that will not get paid. It is sitting in work-in-progress rather than in provision because the adviser said the customer said they'd pay. That is the whole basis of it. When it doesn't come in, it lands in your accounts — carried on the strength of a conversation somebody half-remembers.
  • The warranty job you can no longer claim. A step was missed, weeks ago. Nobody knew. You find out at the point of invoicing, and now it goes to reject. The work was done. The cost was real. The recovery is gone.
  • The warranty WIP list, staring at you, every job waiting to be invoiced and out of your numbers.
  • The used car preps awaiting approval to invoice — done, but not counted.
  • The £3,500 repair that was collecting tonight, so parts and labour would land in the month. It is now collecting at the weekend.

And then the technician

His bonus is on hours invoiced, not hours turned. So the £3,500 job that is now collecting on Saturday does not invoice — which means it does not count — which means he misses bonus on work he has already done. The money has left the business. The person who earned it will not be paid for it, because a customer changed their collection day.

And on that same afternoon, we go to that same technician and ask him to drop everything to sort a sales problem. A PDI. A fault on a car that has to go out tonight. An accessory nobody fitted. So that sales gets the unit. So that sales gets paid.

He does it, because he is a professional and because you asked. And that hour almost certainly never reaches a job card — it is internal, or absorbed, or a favour. He has spent the last hour of the month generating somebody else's commission at the direct cost of his own.

He knows. He knew before we did. He will do it anyway — and he will be a little quieter next month.

The loop

Nobody in this chain is a villain. That is why it never gets fixed.

Every person in what follows is making a reasonable decision inside a constraint set by somebody else who was also making a reasonable decision. The aggregate is a business where everyone is flat out, nobody is productive, the best people leave — and it is nobody's fault.

  1. The budget arrives, higher than last year. The basic salary does not move. So the money you earned last year is now unreachable — not by working less, but arithmetically. Everybody on the floor does that sum in about ten minutes, and they do it correctly.

  2. The team quietly reprices itself to base. That is not laziness. It is an accurate response to a number they can already see they will not hit. The variable element has become fiction, and they have stopped chasing it.

  3. Into that room, we bring new processes and new requirements. More asked, on money they were already unhappy with, chasing a target they have already written off. Do the team give 110%, as you do? Some will. Most will not. And they are not wrong.

  4. The technician who lost his bonus to an invoicing date is now asked to come in on Saturday, or stay late, to push work through. Overtime pays more. He says no. His time was not respected, and he is waiting another month for money he has already earned. Why would he help us?

  5. He starts looking. And the Head of Business can see it coming — can name the person and roughly the month. They sympathise, because they have stood exactly there.

  6. They ask for the headcount. A person who would pay for themselves. The answer is no, because there is a firm line on overheads, set higher up the tree — by somebody who was once sat in the same chair, and who knows exactly what they are saying no to. They are not ignorant. They are constrained by a number that arrived from somewhere above them, too.

  7. The good one leaves. Experience, knowledge, brilliant with customers. No exit interview happens — not really — because everybody already knows why, and knowing changes nothing.

  8. We recruit. It takes months. We do not get like for like. The junction that was being held together by one person's competence and goodwill is now held by someone learning the job. Comebacks rise. The parts queue lengthens. Admin absorbs more. And the next good person starts looking.

And then the scorecard

It comes back as regretted attrition, and it is pointed directly at the Head of Business. The person they did not want to lose. The one they begged to stay. It lands on their review as a retention number, with the quiet implication that they did not lead well enough to keep them.

They did not leave the Head of Business. They left the business that would not back the Head of Business.

The head was asked for. It was refused. They watched it coming and had nothing to offer, because the one thing that would have kept the person had already been taken off the table by somebody else.

So the same decision causes the loss and then punishes the only person who argued against it. The constraint sits upstream. The consequence lands downstream. The metric points inward, at the chair — and away from the line that actually caused it.

Which is precisely why it happens again next year. Nobody ever traces the attrition back to the overhead decision, so the overhead decision looks correct on paper — every single time.

What it actually cost

The headcount line saved a salary. It cost the operation several times that salary in comebacks, lost hours, unbilled WIP, CSI, recruitment and the experience that walked out of the door.

And not a single pound of that will ever be reported as the consequence of saying no to a self-funding head. It will land as a CSI issue. A quality issue. A training issue. And as regretted attrition, on the scorecard of the one person who argued for the head in the first place. Each will be discussed separately, by people who cannot see that they are looking at the same decision from four different angles.

That is what makes it self-sealing. Every consequence is recorded somewhere that points away from the cause. So the cause is never examined, the decision is remade next year, and on paper it looks correct every single time.

This is why flat out is permanent, not temporary. It is not a busy season. It is an equilibrium the business actively maintains, one entirely reasonable decision at a time.

"They should spend a day in a dealership to see what it's really like." They have. They remember. And the budget still arrives the way it arrives — because the number came from above them, and they are as constrained as everyone else.

Which is the whole point. A closed system cannot be opened from inside it. Everyone can see it. Nobody is positioned to say it. And so it is never said.

The ceiling

Build the alert. Then watch what happens to it.

Suppose we build an AI diary that alerts the Sales Manager the moment a handover is added without the minimum notice.

It works. It fires. It is accurate every time.

Does the sales management team act on it — or turn a blind eye, because they need the registration for the month?

You already know the answer

They turn a blind eye. Not out of malice — out of arithmetic. The unit is worth more this month than the notice period is, and everybody in the building has silently agreed that admin will absorb the difference. She always has.

So the alert is not worthless. It is worse than worthless. It is evidence. It proves somebody knew, and proceeded anyway, and now it carries a timestamp and a name. You have not built a control. You have built a paper trail of a deliberate breach, and handed it to your auditor.

This is the ceiling on every AI programme in this industry, and almost nobody will say it to you. The technology is not the constraint. The willingness to act on what it shows you is.

Which is why I read the operating model before we buy anything. Not because the tools don't work — they do. Because a tool that surfaces a truth nobody intends to act on has only made the problem legible, and the exposure worse.

And then there is the rest of it. The congestion charges and the parking fines that arrive for a service car or a sales demo, and nobody was driving. Somebody was. Everyone knows roughly who. It lands on admin, because admin is where things land, and it is quietly written off — every month, in every dealership, for years.

That one is not a technology problem either. It never was.

By role

Same three moves. Eleven jobs.

This is what makes it a rollout rather than a menu. A Sales Executive and a Parts Adviser are doing the same three things — capturing what they know, drafting what they'd otherwise delay, and seeing what they'd otherwise miss. That is trainable across a network. Thirty-three use cases are not.

Sales Executive

Sales

Capture

The qualification that never gets written down. Voice-note the conversation after the customer leaves — who they are, what they actually said, what's driving the change. It lands in the CRM as a usable note, not "cust wants estate."

Draft

The follow-up you keep not sending. A personal, specific message built from that note — referencing the actual conversation, not a template. Ninety seconds instead of "I'll do it tomorrow."

See

Your own pipeline, honestly. Which of your live enquiries have gone cold and you haven't admitted it. Which are real. Where your commission actually is this month.

The truth it exposes: most "live" pipeline isn't live. It's hope, written down.

Sales Manager

Sales

Capture

The one-to-one that vanishes. Coaching conversations, deal reviews and objections raised on the floor — captured as they happen, so the pattern is visible by Friday rather than lost by Tuesday.

Draft

The morning meeting, prepared in four minutes. Yesterday's enquiries, today's deliveries, who's short, what needs saying. Written from your own data, not typed at 07:40.

See

Which exec is actually in trouble. Not the volume board — the leading signals. Enquiry-to-appointment, appointment-to-deal, and who has quietly stopped calling anyone back.

The truth it exposes: you knew who was struggling. You just couldn't prove it until the month had already gone.

Receptionist

Front of house · Every department, at once

Capture

The message, and who ignored it. She is already capturing — that isn't the problem. What's missing is the record of the call that rang out: which extension, how many times, and how long the customer held before she picked it back up and apologised for someone else.

Draft

A handover with enough in it to act on. So the message doesn't come back to her thirty minutes later as "what did they actually want?" — which is the second interruption, caused entirely by the first.

See

The queue, and who is meant to own it. Three people at the desk, two lines ringing, a transport driver on a timed drop, and a service customer who needs to get to work. Every one of them is somebody's — and right now, none of them is anybody's.

The truth it exposes: the receptionist doesn't have a capture problem. She has a triage problem, no authority, and a colleague who has decided that whatever is on their desk outranks a customer on the phone. AI will not fix that. It will simply prove it — with a timestamp, an extension number, and a ring count. What you do with that is a leadership decision, not a technology one.

Service Adviser

Aftersales · The junction

Capture

The diagnosis that dies in your head. The third-visit customer is telling you exactly what's wrong — while the queue builds and the phone goes. You will remember it. You will not write it down. It reaches the technician as three words, and the car comes back a fourth time.

Draft

The pre-call that never gets made. Tomorrow's customers, called this afternoon. It is always the first thing dropped when today is busy — and it is the reason tomorrow will be busy too. Drafted from the booking, it goes out while there is still an afternoon to send it in.

See

That today has already fallen over — at 10am, not month end. Jobs not started. Videos sent and never followed up. Cars ready with no call made. While there is still a day left to save.

The truth it exposes: the red work isn't declined. It is never asked for. And the comeback isn't a technical failure — it is a customer's own diagnosis, given clearly, to a person who was too busy to write it down.

Aftersales Manager

Aftersales

Capture

The comeback, properly recorded. Every return job — what it was, why, who touched it first, what it cost. Not a note in a diary. A record you can total.

Draft

The difficult customer letter. The complaint response, the goodwill decision, the explanation to the OEM. Drafted in minutes with the tone right, so it goes out today instead of festering for a week.

See

The cost of comebacks, in pounds. Hours lost, parts written off, goodwill paid, and the retention you quietly forfeited. The number nobody has ever put on a board pack.

The truth it exposes: comebacks are the largest untracked cost in the department, and everybody has agreed not to measure them.

Parts Manager

Aftersales · Parts

Capture

The lost sale. Every time a part wasn't there and the customer went elsewhere — logged, with the reason. Obsolescence and stockouts are both visible only if someone writes them down.

Draft

The supplier and OEM conversation. The stock return case, the surcharge query, the order justification — built from your own numbers rather than from memory and irritation.

See

What's dying on the shelf. Ageing stock against actual demand, and the trade counter margin you're giving away to hold a customer who isn't worth it.

The truth it exposes: the stock you're proud of holding is usually the stock nobody wants.

Parts Adviser

Aftersales · Parts · The pinch point

Capture

The queue, and what it cost. Technicians standing at the window are clocked on. Every minute is idle time the workshop is paying for, and it appears in no KPI pack anywhere in the country. Capture the wait, and you have created a number that has never existed.

Draft

The quote that wins it on the call. Priced, justified and sent while they are still on the phone — genuine part, this warranty, in stock, collect today. Not "let me ring you back", which is a competitor's order in all but name.

See

Where the day actually went. Trade counter, phone, window, goods-in — and the pre-pick failures that sent a technician back a second time for a job that was already priced.

The truth it exposes: the parts window is where every other department's failure arrives, all at once, and it is the only place in the building where the person holding it has no authority to prioritise. Wrong part ordered to a job means the car isn't ready — which lands on service, which lands on CSI. The failure starts in parts and gets recorded against someone else.

Sales Administrator

The stop gap between process and audit failure

Capture

Every failure, and whose it was. Handover added to the diary with no notice. PDI line never pushed. Valet ticket missing. Day 1 not done, so a car has sat in the compound for three days already past prep. None of it is hers. All of it arrives at her desk.

Draft

The forty chases she is running at once. Outstanding finance, the missing V5, the private plate paperwork, the internal transfer the driver is already waiting for, the invoice the funder needs. Drafted from the deal file, so they go out — instead of waiting for a gap that never comes.

See

What will fail the audit, before it does. PDI sheets unsigned by workshop and sales manager. Handover documents never signed by the customer. Cars going out Saturday that nobody put in the diary. Live — not at the audit.

The truth it exposes: the business's entire compliance and audit exposure is being carried by one person, and not because it is her job. Because she cares, and she is the last one standing before it becomes a finding. She is also covering another site this week, because someone is on leave. That is not a process. It is a single point of failure with a pulse.

Business Manager

Sales · F&I · A role that quietly stopped existing

Capture

Who is actually in the showroom, and with whom. Right now the BM often doesn't know. A quote request arrives from an exec whose customer has been in the building twenty minutes. How do we know what they want? We don't. We are quoting a car nobody qualified, against a part-exchange nobody checked.

Draft

The quoting, off the human. This is the administration that swallowed the role. Take it away and the person is released to do the job the role was actually invented for — the conversation, in a room, with a customer.

See

Every gap in the pre-handover. Products set up? Re-solicited? Valet agreed? Correct ID confirmed? Paperwork and payment done in advance? The answer is usually no to all five — and each one turns handover from a celebration into an admin appointment.

The truth it exposes: the role has been hollowed out and nobody noticed, because the penetration figure still gets reported and someone is still sat in the chair. It used to be a private office and thirty proper minutes — which also bought the appraisal time, so the customer wasn't left waiting. Now it is a desk, a queue of execs, an offer sheet printed, and the figures carried back to a customer who is sitting alone and has never met the person who priced their life for the next four years.

Used Car Manager

Used cars

Capture

The appraisal, honestly. Condition, damage, story — captured at the point of appraisal rather than reconstructed a fortnight later when the margin has already gone.

Draft

The advert that isn't the spec sheet. Written from the actual car, its actual condition and why someone would want this one. Live the day it's ready, not four days after.

See

The blocker tree, live. Every unit not yet retail-ready and exactly what is stopping it — PDI, paint, parts, photos. Days to market, per car, today.

The truth it exposes: time to market isn't slow because of the workshop. It's slow because nobody owns the day the car is standing still.

Head of Business

The chair · Full P&L

Capture

What your managers actually told you. Department reviews, commitments made, risks flagged — recorded, so the conversation in six weeks starts from what was said, not from what everyone now remembers.

Draft

The board pack, and the hard message. The commentary, the exception report, the difficult email to the OEM — drafted from your own numbers so your time goes on the judgement, not the typing.

See

The site, on one screen, today. Not last month's pack. Where the P&L is bleeding, which department is quietly missing, and what you would have found on a walk round if you'd had three spare hours.

The truth it exposes: by the time it reaches the monthly pack, the decision you needed to make has already been made for you.

How it lands

The three moves are the easy part. The rollout is the job.

Every dealer group in the country can buy the tools. Almost none of them will get adoption, because nobody owns the bit between the decision and the Tuesday morning when a Service Adviser has to actually use it while the phone is ringing.

  1. Expose Read the operating model before anything goes on top of it. Where are the junctions? Who is holding demand with no authority? Whose failure is being recorded against somebody else? If the answer is "we don't know", that is the finding.
  2. Decide what you will actually act on Before a single licence is bought. If the alert fires and you will not act on it, do not build the alert — you are only documenting the breach. This is the conversation nobody has, and it is the one that decides whether any of this works.
  3. One move, one department, one site Not eleven roles at once. Pick the move with the clearest pain — usually Capture — prove it somewhere it can be seen, and let the network hear about it from a peer rather than from a memo.
  4. Train the move, not the tool Tools change. The three moves don't. Teach people what they are doing and why, and the software becomes an implementation detail rather than a threat.
  5. Readiness that means something A gate that can actually stop a go-live. Otherwise it is a date with a signature under it, and everybody in the building already knows that.
  6. Own the regression The month after go-live is where the real design happens. Budget for it, staff it, and stay for it — because the issues that matter only exist once it is live.

Flat out to productive. That is the whole job. Not more hours, not more people, and not more software. The work that is currently being dropped — the pre-call, the qualification, the follow-up, the acknowledgement — is the work that would have prevented the day you are now surviving.

AI can buy that time back. But only in a business that has first decided what outranks what, and that is willing to act when it finds out.

What this is

This is what a Health Check finds. In your building, the details will be different. The shape will not.

Nothing above came from data. It came from asking people what their day is actually like — and being willing to write down what they said. Two to three days on site, every level, no managers in the room, and a written finding you can act on without me.

Most of what I find costs nothing to fix. Tag the keys. Write down what outranks what. Make the pre-call somebody's actual job. Being heard is free.

Ground Truth Advisory →