Subversive Ventures
← Uncommon Sense
Uncommon Sense

The Shovel and the Teaspoon

31 Jul 2026•David Ashenden

You would not put sugar in your tea with a shovel. Nor would you dig the garden with a teaspoon. Both are tools, both work in the strict sense that a shovel can lift sugar and a teaspoon can shift soil, and both would leave you looking like a man who has lost an argument with his own kitchen. This is the whole argument about AI in business, and almost nobody selling it has noticed.

We have spent two years being told the great task is getting people to try the stuff. Put a chatbot on every desk, announce it at the all-staff, count the logins, wait for the productivity to bloom. And the small automations are real. Killing the invoice chase, drafting the dull email, tidying the spreadsheet nobody wanted to touch. Teaspoon jobs. Sugar in the tea. Useful, and fine. But we have gone past the point where trying it is the interesting part, because the tool is everywhere now, free, in every hand on the street. And a tool everyone owns has stopped being an advantage. When the whole village has the same shovel, the shovel is worth nothing. What is left is the person holding it.

Uncommon sense (for busy people)

  • A shovel and a teaspoon are both tools. Knowing which the job wants is the skill.
  • When everyone owns the same tool, the tool stops being the edge.
  • AI is a very good tool. It is still just a tool.
  • The technology changes weekly. People have not changed since the Romans.
  • A coach sells you belief. An expert tells you which year of the engine to avoid.
  • Primark and Savile Row use the same wool. You are paying for the fit.

I wrote before that in a human business the caring was never the product, the proof was. This is the same shape in a different coat. The tool was never the product. What you do with it is. AI is a very good tool, better than most, and I use it every day. It is still a tool. It does not know which job it is for. You do, or you don't.

Here is the bit the AI discussion mostly steps over. The technology changes every week. People do not. The way a customer decides, trusts, hesitates, and finally buys has not shifted since a Roman stood in a market haggling over a fish. A new tool does not rewrite the human. It just hands the human a faster way to do the thing they were always going to do. None of which makes the human the automatic winner, mind. Downing tools to do the whole job by hand is its own daftness, like turning down the shovel to dig the flower bed with your fingers. The trick is the right tool for the job, held by someone who knows which job it is. Which is why expertise, knowing what good actually looks like, knowing which job wants which tool, is the one thing that climbs in value while the tool itself gets cheaper and more common by the month.

Which brings me, with some reluctance, to the business coach. And let me be fair here, because plenty of them are the real thing, scarred and sharp and worth every penny. The trouble is not the good ones. The trouble is that the trade is unregulated, so a bloke who finished a book last month can print himself a certificate over a wet weekend and stand on the same shelf, behind the same confident website, as someone who has spent twenty years doing the job. Nobody inspects the shelf, so hiring one is a spin of the roulette wheel, and you do not learn which pocket the ball fell into until the money is gone. The good ones sell you more than belief. The weekend merchants sell belief alone, encouragement, asking you what you already half know in a warmer voice until you feel brave enough to act on it, the manifesto that writes itself over a good coffee. But belief is not experience. And experience, the kind that tells you which year of the engine to avoid, you cannot read your way to. You live it, over years, some of it joyous, some of it dull, some of it painful enough to still ache when it rains. It conditions you to the situations you have already walked into and the ones that walked into you. A textbook cannot hand that over, and neither can a comforting article on Forbes. What they sell you is the plan. And no plan survives contact with the enemy, as a Prussian general worked out long before the first chatbot.

It turns out more than 80% of AI projects fall over, which is roughly twice the failure rate of ordinary IT. Twice. And the technology did not get harder, it got easier. A shiny new tool just flatters everyone into skipping the dull question of what it is actually for. They bought the shovel and went looking for the sugar.

Think about a suit. A Primark two-piece and a Huntsman bespoke are made of roughly the same things. Wool, thread, a lining, some buttons. The gap between them is about thirty quid against four and a half grand, and that gap is not the cloth. It is the fit. Someone measured you. Someone who has cut a thousand shoulders clocked that yours sits lower on the left and built the thing around it. The cheap suit is the tool applied to nobody in particular. The tailored one is the tool applied by expertise to your actual body. AI hands every business the Primark suit for free. The valuable part, the part worth paying a specialist for, is the tailoring, the know-how that makes a generic tool fit your specific, awkward, particular company.

And cost-effective is the phrase that matters here, because the cheap suit only looks cheap until you count the two you replace every year and the room you lost the moment you walked into it. The tailored one costs more on the day and far less across its life. Expertise runs on exactly that maths. It is not the expensive option. It is the one that stops you spending four hundred grand teaching a chatbot to do the wrong job beautifully.

Which points at something duller than the AI conversation usually allows for. The people are the only part of this whole business that does not get cheaper and more common every six months. Everything else on the invoice is depreciating while you read this. The person holding it is not.

So before the next licence, or the next pilot, sit with the questions the vendor would rather you skipped over. Which of your problems are teaspoon problems, the sugar in the tea, and which actually need someone who can dig. Have you honestly told the two apart, or did you buy the shovel because everyone else on the estate had one. When the whole sector is running the same model by the new year, what do you still have that they don't. And who in your building already knows what good looks like, the person who can see the job before they reach for the tool. Are you listening to them, or to whoever logged in the most this quarter.

The shovel is magnificent. Genuinely, buy the shovel. Just keep it out of the sugar bowl.

Telling the shovel jobs from the teaspoon jobs, and being handy with both, is roughly what I do at subversive.ventures.

Further reading

← More Uncommon SenseGet new pieces by email