HSS Rebuilt Its Whole Operating Model. Now Nearly Half Its Orders Never Touch a Phone.

THE CRAFT

HSS spent nearly seventy years as a company you could touch. It began in 1957 as a single rental shop in West London — HSS is short for Hire Service Shops — and grew into one of Britain’s best-known equipment-rental businesses: the place a builder calls when the job needs a cement mixer, scaffolding, a generator, a floor sander for the weekend. All of that equipment lived in depots: local branch-yards, several hundred of them at the peak, full of machinery waiting to be rented out. The depots were the company. Owning them, stocking them, maintaining them, insuring them — that was the business, and had been for as long as anyone working there could remember.

Last year, HSS gave the depots away. Roughly 130 rental locations went to a buyer for a single pound. And the pound wasn’t the whole price — because the physical rental business had been losing money for years, HSS paid the buyer £26 million to take it on. The company wrote off £113 million on that side of itself in the same set of accounts. When a business is worth less than nothing — when it comes with more liabilities than value — you can’t sell it for a positive number. You pay someone to accept it. So HSS did.

What it kept was the part with no warehouse: the relationship with the customer, and the software it had quietly built to serve that customer without owning a single generator. HSS didn’t sell tools anymore. It ran a marketplace — a piece of software that takes a builder’s order and routes it to the best supplier in a network of hundreds — and it had written that software itself. The man who built the thing they kept, and who now runs the whole company, is Tom Shorten.

I want to be careful with this one, because on the surface it looks like a distressed British small-cap doing something desperate, and in some of the numbers it still does. But underneath is the cleanest example I’ve found this year of a discipline I keep telling you matters more than any tool you’ll be sold: they rebuilt how the work actually happens first, on a foundation they controlled, and only now — with a clean system underneath — are they adding the AI agents on top. Most companies run that order backwards. This brief is about the sequence, and about a leader so committed to it that he burned the boats to prove it.

THE OPERATOR

The situation

To understand what Shorten inherited, you have to understand what equipment rental actually is, because the economics are unforgiving in a way that shaped everything. A rental company buys expensive machinery, then rents it out in small daily and weekly increments and hopes to earn back the purchase price and the upkeep before the equipment wears out. It carries the cost of the fleet, the depots that store it, the vans that move it, the mechanics who fix it, and the debt that paid for all of it. The margins are thin in a good year. HSS did not have many good years. For most of its life as a public company it was heavily indebted and barely profitable, a proud old name running a structurally hard business.

By the numbers that matter, the physical side was in trouble. In its most recent full accounts, a fifteen-month reporting period, HSS reported a pre-tax loss of £130 million — the bulk of it a £113 million writedown on the value of the old rental business. That writedown is the whole story in one line: the company’s own accountants had concluded the depots-and-fleet operation was worth far less than what sat on the books. This was not a business you could fix by renting out a few more floor sanders.

But here is the thing Shorten had seen, and had been building against for years. He’d joined HSS in 2017 and built its online operation from scratch — the small digital rental business that took orders through a website instead of over a counter. And what he understood was that the customer never actually wanted the depot. The builder on the job site doesn’t care whether HSS owns the generator or sources it from someone down the road. The builder wants the right equipment, at the right price, delivered to the right entrance by the right hour, with one invoice that’s correct at the end. The depot was how HSS had always met that need. It was not the need itself.

Which raises the question that reorganizes the entire business once you ask it honestly: if the customer only wants the outcome, why own three hundred warehouses of depreciating steel to deliver it? Why not keep the customer, keep the ordering and the coordination and the billing — the parts the customer actually experiences — and let a network of suppliers own the physical equipment? That is the question Shorten’s marketplace was the answer to. And answering it meant HSS would have to stop being the thing it had been for generations.

The move

The move came in the right order, and the order is the whole lesson, so watch the sequence.

Shorten’s team did not start by buying an AI product and looking for somewhere to point it. They started by rebuilding how an order actually flows through the company, in software they wrote themselves. The center of it is a system they named Brenda — a matching engine that takes an incoming order and finds the right supplier for it out of a network of more than nine hundred, then manages the hand-off. Brenda has five faces onto the same machine: one for the customer placing the order, one for the supplier receiving it, one for the HSS colleague overseeing it, one for administration, and one for the engineers who maintain the system. It is not a chatbot bolted onto an old process. It is the operating model of the new company, expressed as code.

With the ordering rebuilt, they turned to the money. In the old world — and in most of this industry still — a rental invoice is a small nightmare, because a rental has no fixed length. A tool goes out on Tuesday and comes back whenever the job finishes, which might be Friday or might be three weeks Thursday. Every day the equipment is out, the amount owed changes, and the amount HSS owes its supplier changes with it. Traditionally you reconcile all of that by hand, after the fact, and the customer gets a “surprise invoice” weeks later that nobody can quite explain. HSS built a system that recalculates the full financial position of every live contract every single day — what the customer owes, what HSS owes the supplier, in step with reality, no reconciliation, no surprises. The company’s technology chief, Daniele Turi, describes it plainly: it tells them every day precisely what each customer will pay and what each supplier is owed. This is not artificial intelligence. It is careful, conventional software doing a job the old company did with spreadsheets and apologies — and it is exactly the sort of unglamorous plumbing that has to work before anything cleverer sits on top of it.

Then — and only then — comes the AI. HSS built its software so that its own engineers work almost entirely through AI-assisted tools; Shorten says the team writes essentially no code by hand and gets many times the output it used to, shipping in two-week cycles. And the genuinely new frontier, the agentic layer, is being built onto the clean foundation right now: agents that read the unstructured emails suppliers send — a proof-of-delivery note, a collection notice — and turn them into structured actions in the system automatically, so a small supplier who lives in his inbox needs no special integration to plug into the marketplace. Agents that chase a missing proof of delivery across the supplier, the customer, and the timeline without a person holding the thread. These are early. I’ll be honest below about how early. But notice the ground they’re being built on: a rebuilt ordering system and a rebuilt billing system that already work.

There was a human decision underneath the technical one, and it’s the part most companies get wrong. Shorten’s team built the new business on a separate track from the old one — a second tower, deliberately walled off from the depot business’s systems and habits, with “no drawbridge” back to the old way of doing things, moving people across as the new tower could hold them. They did it this way on purpose, because the fastest way to smother a new operating model is to make it negotiate with the old one for every decision. And there was a false start worth keeping: when they first stood the system up, they held the data back from their own people, nervous about what would happen if everyone could see everything. Turi’s team admits they were, in his word, scared. Then they opened it up, because a marketplace where the colleagues can’t see the market is just a slower version of the phone. The willingness to name that hesitation out loud is, to me, a better signal of a serious team than any of the growth numbers.

The last move was the one that made headlines and made the strategy real: HSS shed the depots entirely. It didn’t run the old model and the new one side by side forever, hedging. It sold the physical business, took the writedown, paid to be rid of it, and became the marketplace. A rival rental company, Speedy Hire, stepped in to supply equipment into the network and took a small stake alongside — a competitor putting its own capital behind the idea that this is where the business is going. The company changed its name to match what it now was.

The result

Start with the number the whole strategy turns on, because it’s the one that tells you the bet is landing. When Shorten started building the digital side, about 5% of HSS orders came through the website; the rest came by phone, email, and counter. Today, by his account, 40 to 45% of all orders come straight through the software, with the remainder by phone, email, and chat. That is not a marginal channel improvement. That is close to half of a whole company’s order flow moving, in a few years, from a person taking down details to a system that takes the order, prices it, finds the supplier, and starts the contract on its own. In an industry where, by Shorten’s estimate, fewer than 5% of orders across the sector go through software at all, HSS has moved its own to nearly half. The rest of the industry is still answering the phone.

The shape of the business changed underneath that number, too. HSS now runs a network of more than nine hundred onboarded suppliers and tens of thousands of product lines, serving thousands of business accounts a month without owning the equipment behind them. The mix of what it sells shifted hard toward the asset-light side — non-rental revenue, the marketplace-style business, grew from a sixth of the total to better than a quarter in a single year. And the most recent trading update reported group revenue up more than 15% year over year as the new supply arrangement ramped. The marketplace, the part with no warehouse, is doing the growing.

Now the honest take.

This is not a turnaround you’d hold up as finished, and I won’t pretend otherwise. HSS is a penny stock — the shares trade around two pence, and earlier this year they fell about 15% in a day when the company warned that the new supply ramp was running slower than hoped. To get through the transition it took on fresh financing this summer whose convertible piece, if it fully converts, would dilute existing shareholders by roughly 60% and hand the new lender the majority of the company. Revenue is shrinking on the top line as the divested depot business leaves it, and profitability for the year ahead is guided barely above breakeven. If you want to tell the pessimistic story — a distressed small-cap that sold its heritage and mortgaged its future to buy time — the raw material is all there, and an honest brief has to put it on the table.

What that story leaves out is the sequence and the conviction. The marketplace wasn’t a panic move dreamed up when the depots started sinking; Shorten had been building it for years before the sale, which means the company shed the old business because the new one was ready to carry it, not the other way around. And when the refinancing was done and the shares were on the floor, Shorten went into the market and bought more of them with his own money. A chief executive buying a penny stock nobody likes, in the middle of a dilution, is not a man managing a story. That’s a man who thinks the thing is worth more than the market has noticed.

I’ll also be straight about the AI specifically, because it’s the reason for the sequence lesson and I don’t want to oversell it. The agentic layer — the email-reading agents, the proof-of-delivery chasers — is being built now, and there is no clean, audited number yet showing what those agents saved. The realized proof in this story is the rebuilt operating model: the marketplace that moved half the orders, the billing system that tells the truth every day, the shift to an asset-light shape that’s growing while the old business shrinks. The AI agents are the next chapter, laid on a foundation that already holds. Which, if you’ve read this far, is precisely the point.

The Craft of AI read

Here is what I’d take from this if I ran any business where the customer wants an outcome and doesn’t care how you produce it behind the counter — which is most businesses.

Every AI pitch landing in your inbox this quarter leads with a thing to buy. An agent. A number of copilot seats. Some platform with a clever name and .ai on the end. And it arrives before anyone selling it has sat with the people who actually do your work and understood, at the level of the order and the invoice and the hand-off, how the work truly happens. The tool comes first; the understanding is supposed to catch up later, if it ever does. That inversion is the single most expensive mistake in corporate AI right now, and HSS is a photograph of the opposite discipline. Shorten’s team understood the work down to the daily arithmetic of a variable-length rental, rebuilt the ordering and the billing around that truth in software they controlled, and are adding AI agents on top only now that there’s a clean surface to add them to. AI laid over a broken process just makes the brokenness run faster. AI laid over a rebuilt one compounds. HSS did the rebuild first.

It’s the same order Dave Peacock ran at Advantage Solutions — three unglamorous years consolidating systems and cleaning data before letting any AI touch the work — and it keeps being the order that separates the AI programs that pay from the ones that stall. The technology is different, the industry is different, the drama is different. The sequence is identical. Understand the work, rebuild the work, then multiply the work. In that order, every time.

There’s a second thing here, and it’s the harder one to copy, so I want to name it honestly rather than turn it into a slogan. HSS took the most extreme version of this move available: it didn’t just build the new system, it sold the old business the new system replaced. You almost certainly don’t need to go that far, and I wouldn’t advise most leaders to. You don’t have to sell your depots. The transferable lesson isn’t the drama of the disposal — it’s the order of operations, and the willingness to build the new way on its own track, walled off from the old one’s gravity, until it’s strong enough to carry the load. HSS burned its boats because its old business was a genuine money-loser it was better off without. Your version is probably quieter: build the rebuilt workflow beside the current one, prove it, and let it take the volume over as it earns it. Same sequence, less fire.

Things to consider

• The customer wants the outcome, not your machinery. HSS’s whole reinvention started with one honest observation: the builder never wanted the depot, only what the depot delivered. Ask the same question of your own business. Which of the things you own, staff, and maintain does the customer actually value — and which are just the way you’ve always produced the thing they value? The gap between those two is where a rebuild lives. Owning the means of delivery feels like the business right up until someone proves it isn’t.

• Rebuild the workflow before you buy the intelligence. The most useful system HSS built — the one that recalculates every contract’s money every day — isn’t AI at all. It’s careful, ordinary software doing a job the old company did with spreadsheets and apologies. If your data and your core processes can’t yet tell you the truth on their own, pointing an AI agent at them will only multiply the confusion. This year’s AI budget may be better spent making the work knowable than making it “intelligent.” The intelligence goes on last, onto a surface that already holds.

• Build the new way on its own track. HSS built its marketplace as a separate tower, deliberately cut off from the old business’s systems and habits, “no drawbridge.” The reason matters: a new operating model that has to negotiate with the old one for every decision dies of a thousand compromises. If you’re rebuilding how a core workflow runs, wall the new version off from the incumbent’s gravity long enough for it to stand on its own — then let it take the load. Most rebuilds fail not on the technology but on being strangled by the process they were meant to replace.

• Watch what your leaders do with their own money. In the middle of a dilution, with the shares on the floor, Tom Shorten bought more of them himself. You can’t put that in a pitch deck and you can’t fake it. When you’re evaluating whether the people running a transformation — yours or a partner’s — actually believe in it, look past the slides to where they’ve put their own capital, their own reputation, their own exit. Conviction that costs the person nothing is worth exactly what it cost them.

• The honest version of “we’re behind” is a starting line, not a verdict. By HSS’s own estimate, fewer than one in twenty orders across its entire industry go through software at all. A leader in that industry could read that as cover to keep waiting — everyone’s behind, so why rush? Shorten read it as the opening: while the rest of the sector answers the phone, he moved nearly half his own orders onto software and built the network everyone else will eventually need. If your industry is genuinely early, that is not a reason to relax. It is the last cheap moment to move first.

THE WORKBENCH

Do this tomorrow

This one takes an hour, a whiteboard, and a willingness to hear an uncomfortable answer.

Separate what the customer values from what you happen to own. Draw two columns. In the first, list the things your customer actually experiences and pays for — the outcome, delivered right, billed correctly, on time. In the second, list the major assets and activities you own, staff, and maintain to produce that outcome: the warehouses, the fleet, the field team, the systems, the departments. Now draw lines between them. For every item in the second column, ask one question out loud: does the customer value this, or is it just how we’ve always produced what they value? Be honest about the ones with no line running to the first column. Those are the candidates — not necessarily to sell, but to rethink. HSS ran this exercise and concluded the customer never wanted the depot.

Then price the truth in the billing. Pick your single most variable, most-disputed transaction — the invoice that generates the most “why am I being charged this?” calls, the reconciliation that eats the most of someone’s week. Count two things: how many hours a month your people spend reconciling it by hand, and how many of those disputes end in a credit, a write-off, or a lost customer. Multiply the hours by loaded cost, add the credits, and write down the annual figure. That number is what it costs you every year to run a process that can’t tell you the truth on its own — and it’s the number that tells you whether your next technology dollar should buy an AI agent or should first buy a system that gets the invoice right the first time. HSS built the second thing before it reached for the first. Most of that annual figure is sitting in your business today, this quarter, in a workflow nobody has ever put a price on.

THE QUESTION

Tom Shorten didn’t buy an AI product and go looking for a problem it could solve. He did the slow, unglamorous thing first — understood how an order and an invoice actually move through his company, rebuilt both in software he controlled, walled the new business off from the old one’s habits, and proved it moved nearly half his orders before he ever reached for an agent. Then he sold the part of the company that couldn’t be rebuilt, and bought the shares of the part that could.

So here’s the question, and it’s worth thirty honest seconds. If you pointed AI at your most important workflow tomorrow morning — the order, the claim, the schedule, the invoice, whatever carries your business — what would it find underneath: a rebuilt process that already tells the truth, or the same tangle of spreadsheets, phone calls, and surprise invoices you’ve been meaning to fix for years? Because the AI doesn’t fix the tangle. It just runs it faster, and hands you the same mess at a higher speed and a higher bill.

Shorten’s answer, when he started, was the tangle. The whole story is what he did about it before he let the AI anywhere near it.

---

Want this done for you?

You just ran the two-column version with a whiteboard and an hour.

My Ground-Up Workshop is the rigorous version of the same exercise, run on your business with the small group of your people who do the work. Every AI strategy you’ve been sold starts with software. This one starts with the people who do the work.

One to two days in person, then a synthesis week, and you walk out with a plan: a ground-truth map of how your core workflow actually runs, the one where a rebuild would release real operating margin, a target operating model with AI built in from the ground up, and a 30-60-90 day plan your own team can start on. I do a small number of these a quarter, for $20,000.

I’m an investor in geniant. For more than 25 years, our craftspeople have done one thing: understand how work actually happens, design how it should happen, and build software that works the way people do. When a workshop gives you the roadmap to execute or turns into a build, they’re who does it — one senior-led team, layered on top of the systems you already run, value in weeks rather than months.

Book a discovery call →

— Grant K. Baldwin