What Goldman Sachs Didn't Say - And Why That Matters
The coverage of yesterday's Goldman Sachs 10,000 Small Businesses AI for Growth day will tell you what Rishi Sunak thinks small business leaders should do about AI. I'm more interested in the question that not one of the speakers, including Sunak, asked.
I spent yesterday in Birmingham with 300 of the most commercially sharp SME leaders in the UK. Rishi Sunak on stage. Government advisors. Goldman partners. A genuinely impressive lineup, and a genuinely impressive room.
But an event designed to drive adoption has a structural incentive to keep the frame positive. And I came away thinking about everything that wasn't said.
The conversation had a shape to it. The shape of adoption. How to start, how to train your workforce, how to identify the right use cases, how to govern it responsibly - questions pointing in the same direction: forward, up, faster. The panel contributions from the alumni were honest and practical in a way that keynote events rarely manage, and the energy in the room was real. These are business owners who are leaning in, investing time, taking AI seriously. That matters. What was missing was the question of what forward actually leads to.
When Rishi Sunak mentioned the K-shaped economy - the risk that AI splits our economy into two diverging tracks - he framed it as large businesses pulling away from small ones. A real concern, and one the 10KSB programme is directly positioned to address. But nobody asked the harder version of the same question: what happens when these 300 businesses leverage AI to hoover up market share from competitors that can't adapt, and the workers displaced in that process - people who couldn't reskill fast enough - are the same people who buy from them? Multiply that dynamic across the country, and across the wider world, and the problem becomes easy to see. Yesterday, adoption was framed in isolation. Nobody framed it in societal terms.
The opening figure, quoted in the welcome remarks, showed that 22% of the room expected headcount to grow as a result of AI investment. That figure was cited as evidence of optimism. Nobody asked what the other 78% said.
The room itself was a selected sample - Goldman 10KSB alumni, self-selected for an AI day, travelled to Birmingham to be there. Even within that highly motivated group, the spectrum ran from daily ChatGPT users to people running fully integrated automation stacks. The distance between that room and the median SME owner-manager in this country is enormous, and the distance between the median SME owner-manager and the workforce they employ is enormous again.
This is the territory I'm spending most of my thinking time in at the moment, through the lens of Harshlight - a platform I'm building around a straightforward proposition: the businesses, investors, and policymakers who understand which economic scenario is actually materialising, and who act on that understanding before the signals become undeniable, will be disproportionately positioned for the decade ahead.
The scenarios matter because they lead to very different strategic decisions. In a scenario where AI-driven productivity genuinely creates new economic activity faster than it displaces existing roles, the adoption-first narrative is right and the job is execution. In a scenario where displacement runs ahead of rotation - where labour markets can't absorb the transition fast enough - the risk sits not just in the workforce but in the consumer base that underpins SME revenue. In a third scenario - what Daron Acemoglu's work points toward - the productivity gains from AI are real but captured almost entirely at the top, while the middle is hollowed out without replacement. New task creation, which drove re-employment through every previous wave of technological change, simply doesn't materialise at the scale needed. That isn't unevenness across a map. That's a structural ceiling on who the economy works for.
I've spent twelve years as MD of a manufacturing business in the South West. I understand exactly how powerful the productivity argument is, and I've lived through what happens when businesses fall behind technologically. The entrepreneurs in that Birmingham room are, as Sunak rightly said, the bridge between the economy and the communities they operate in. The question is whether the bridge holds if the community on one side is rendered unstable by the very technology the bridge depends on - and an unstable bridge is a risk to both sides, not just one.
That's not a comfortable question to ask in a room full of people who've just been told that adoption is everything. But in my experience, the conversations that aren't happening loudly enough are usually the ones that most need to be.