The Briefing
Signal over noise.
Signal over noise — a curated digest for busy founders and operators.
Last reviewed 4 July 2026 · Reviewed quarterly. No predictions — only what published research supports.
AI, used properly
Most owners don't need more tools — they need fewer, better-chosen ones. What the adoption research actually shows:
Start with two or three tools, not ten.
Survey data found owners spent an average of $2,340 on AI subscriptions in 2025, yet roughly a third of those tools sat unused within 90 days. Pick tools that remove a named weekly pain point, master them, then expand.
Measure time saved, not vibes.
Work out your effective hourly rate, multiply by hours saved per week, and compare against the tool's monthly cost. Under 3x return — reconsider it. Over 5x — find the next bottleneck.
Start where the before-and-after is measurable.
Across business types, the first AI win is almost always customer-facing communication or administrative documentation: high volume, repetitive structure, easy to compare.
Ask two questions before you subscribe.
What happens to my data, and can I export it if I leave? Vendors confident you'll stay make leaving easy.
Never publish AI output unreviewed.
Roughly three quarters of small business owners agree audiences now gut-check information with people they trust — a human voice is a competitive asset, not overhead.
What AI can't replace
The research consensus is complement, not substitute. Where humans keep the edge:
The EPOCH capabilities.
MIT Sloan research identifies five capability groups AI struggles to replicate — Empathy, Presence, Opinion and judgement, Creativity, and Hope — and finds human-intensive tasks are strong candidates for augmentation rather than automation.
Crisis handling and live judgement.
A 2026 analysis of 84 occupations found crisis intervention the single hardest skill to automate; de-escalation changes second by second in ways machines can't read.
Relationships built over years.
Client-relationship cultivation appears in nearly half of all occupations studied — far more than any purely technical skill — and roughly halves automation risk.
Accountability where mistakes are costly.
In medicine, law and finance, responsibility for the decision stays human even when the analysis is machine-assisted.
The rehiring pattern.
Gartner research suggests around half of companies that cut jobs citing AI will rehire for similar functions by 2027, often in retitled roles blending human expertise with AI. Jobs are changing shape faster than they are disappearing.
Where growth is concentrating
Figures below are analyst projections, not guarantees — but the direction of travel is consistent across sources:
AI infrastructure.
Hyperscale data centre services is ranked the fastest-growing US industry of 2026 at close to 29% revenue growth (IBISWorld), and the AI sector overall is projected to compound at roughly 28% a year through 2030.
Digital health.
Estimated at around $199 billion in 2025 and forecast to approach $574 billion by 2030 — remote monitoring and AI-assisted care are driving it.
Cybersecurity.
Growing at roughly 12% a year towards a projected $300+ billion market by 2027, pushed by cloud adoption and AI-powered attacks meeting AI-native defence.
Clean energy.
Global low-carbon investment reached about $2.1 trillion in 2024 (BloombergNEF) — now a mainstream capital allocation, not a niche.
Vertical software.
Analysts consistently point to industry-specific software — AI-enabled diagnostics, logistics dispatch, compliance tooling — as the most accessible opportunity for smaller builders, where domain knowledge beats platform scale.
Innovations that are already real
No crystal ball — everything here is deployed or in commercial use today:
Post-quantum encryption in your pocket.
Lattice-based cryptography — designed to withstand future quantum attacks — already protects Apple's iMessage, with Android support planned.
Sodium-ion batteries.
Made from abundant materials, cheaper and safer than lithium, and now moving into commercial deployment led by CATL, the world's largest battery maker.
Driverless taxis at scale.
Autonomous vehicles already operate around the clock in several major US and Chinese cities.
Precision fermentation.
Animal-free egg and whey proteins are being produced commercially in fermentation tanks — the same approach is reaching cosmetics and pharmaceutical compounds.
AI coding assistants.
Named among MIT Technology Review's 2026 breakthrough technologies — mainstream enough that the caveat has shifted from "does it work?" to "review what it produces".
Personalised gene editing.
The first bespoke gene-editing treatment was administered to a single infant patient, with clinical trials now planned — medicine tailored to one person's genome has moved from theory to practice.
A short list of places worth your time
Fewer, better sources beat an overflowing feed. Our shortlist:
MIT Technology Review — 10 Breakthrough Technologies.
One annual list, 25 years running, honest about its own misses.
World Economic Forum — emerging technologies and Future of Jobs reports.
Free, rigorous, and the standard reference on how work is changing.
IBISWorld industry trend rankings.
Fast, data-backed answers on which industries are actually growing.
Perplexity.
Sourced answers for business research and competitor analysis — every claim linked, so you can check it.
Zapier.
Plain-English automation between the apps you already use; automate one painful manual workflow before adding any new tool.
Notion.
One place for SOPs, meeting notes and project tracking — businesses scale on systems, not memory.
Fathom.
Free AI meeting notes, so decisions and action items stop evaporating after calls.