AI hitting jobs is
just layer one.
In a services-heavy economy, labour displacement doesn't stay in labour. It pulls revenue out of professional-services firms, concentrates unemployment in services-heavy cities, softens consumer spending, drags office property, pressures tax receipts, breaks the graduate pipeline, and feeds back. Seven layers. At the 2035 peak of this model, up to £481B/yr of aggregate UK impact.
Each layer is triggered by the previous one. Not all seven need to land.
Scroll through the cascade, or tap a layer pill in the sticky nav to jump. Each layer compounds the one before; the model projects cumulative impact, not a sum of independent shocks.
- 01Revenue Evaporation in Professional Services2026–2028Extreme in exposed sectors
AI good-enough output destroys billable-hour models. Phase-transition demand drops of 10-100× cheaper; quarter before threshold revenue looks normal, quarter after it is gone. £40bn+ at risk across tax, recruitment, legal, consulting, audit.
- 02Concentrated White-Collar Unemployment2027–202925-45 age bracket, cities, university-educated
Entry-level hiring already frozen (graduate vacancies -44% YoY). Median UK savings £1,000 → 3 months unemployment = personal financial crisis. Unlike 2008, this falls on the exact demographic that funds middle-class consumption.
- 03Consumer Spending Collapse in Services-Heavy Cities2027–203015-20% GDP contraction in London
Every £1 of professional wages generates £2-3 of local economic activity. Hospitality, retail, services already contracting (tech sector -6.3%, first decline in a decade). London, Manchester, Edinburgh hit hardest.
- 04Commercial & Residential Property Correction2028–203120-40% value decline in professional belts
Office demand collapses as cognitive roles offshore or disappear. Outer zones (Reading, Guildford, St Albans) crash hardest. Premium central London holds as wealth concentrates. Banks' loan books take direct hits.
- 05Government Revenue Collapse2028–2032£100bn+ fiscal exposure
Income tax + NI + corporation tax all disrupted simultaneously. OBR forecasts assume 4% full employment; our projection: 15%+ cognitive-sector unemployment. Student loan book (£267bn, projected £500bn) becomes fiscal black hole.
- 06Education Pipeline Collapse2027–2030Professional degrees hardest hit
University value proposition breaks. Law, accounting, consulting applications decline as graduate destinations dry up. Russell Group maintains prestige; mid-tier institutions face existential pressure.
- 07Reflexive Feedback LoopsContinuous (2026→)Multiplied, not additive
Unemployment → spending ↓ → tax ↓ → govt support ↓ → unemployment ↑. Property ↓ → wealth ↓ → confidence ↓ → lending ↓ → property ↓. Each layer amplifies the others. Traditional forecasting models assume these are separable — they are not.
Compounded fiscal, property, consumer, and pension drag.
Stacked annual impact by channel. This is the shape of the cascade in £: the layers compound rather than substitute. 2035 peak £481B/yr, roughly 5% of UK GDP.
Where the first £41B goes.
Every row is a professional-services sub-sector whose billable-hour model AI undercuts. The “at risk” column is the midpoint of the disruption range applied to 2025 revenue.
| Sector | Current rev. | Disruption | At risk | Mechanism |
|---|---|---|---|---|
| Tax Advisory | £3.2B | 75–85% | £2.6B | Rules-based work, AI outperforms humans |
| Recruitment | £11.0B | 60–70% | £7.7B | Information arbitrage demolished |
| Mid-tier Consulting | £8.0B | 30–50% | £4.0B | Associates / analysts eliminated |
| Legal Services | £41.0B | 30–40% | £16.0B | Document review, research automated |
| Audit | £5.0B | 40–60% | £3.0B | Junior sampling → continuous AI monitoring |
| Architecture | £4.4B | 25–35% | £1.5B | Parametric design, compliance checking |
| Financial Advisory | £12.0B | 35–50% | £6.0B | Analysis commoditised, relationships persist |
| Total | £84.6B | £40.8B |
The signals aren’t projected. They’re arriving.
Where the cascade has already moved.
What the cascade lands on.
Applications per role nearly tripled in four years.
The supply-demand inversion is already severe. Since 2019-20, graduate positions have shrunk from ~21K to ~15.6K; applications have ballooned from 700K to 1.1M. Apps per role jumped from 33 to 71. That's a generational squeeze on access to early-career professional work.
| Cycle | Apps (K) | Roles (K) | Per role | YoY Δ |
|---|---|---|---|---|
| 2019-20 | 700 | 21 | 33 | – |
| 2020-21 | 580 | 17.5 | 33 | -12.0% |
| 2021-22 | 560 | 19.6 | 29 | +12.0% |
| 2022-23 | 850 | 19 | 45 | -3.0% |
| 2023-24 | 1200 | 17 | 70 | -10.5% |
| 2024-25 | 1100 | 15.6 | 71 | -8.0% |