Insights

Research

The research behind the standard.

Field notes, adoption data, and honest analysis on where AI actually pays off — and where it quietly costs you. Written from live client engagements, not from a press release.

Flagship report

22%

of organizations have not yet put AI into a single function — and the gap is widening every quarter.

Strategy9 min read

The 22% problem: what happens to the businesses that wait

Seventy-eight percent of organizations already run AI in at least one function. This report is about the other 22% — who they are, what waiting is quietly costing them, and the narrow window in which catching up is still cheap.

Every adoption curve has a long tail, and AI's tail is now the story. When roughly four in five organizations already run AI somewhere in the business, the remaining fifth are no longer 'early' in anything — they are late, and the gap compounds monthly. The cost of waiting rarely shows up as a single line item. It shows up as slower quotes, thinner margins, staff burning hours on work a competitor automated last quarter, and customers who quietly route to the faster vendor. Our field data across healthcare, finance, and retail engagements points to the same shape: the businesses that move first don't win because their models are better — they win because they started structuring their data, their processes, and their judgment eighteen months earlier. This report maps who the 22% are, why the honest math of falling behind is steeper than it looks, and the scoped, weeks-not-quarters way back in.

YK

Yisrael Krug

Principal, ELSIAA Intelligence Desk

Request the report →

The Numbers

Adoption is no longer a question of if.

Sources below · Updated July 2026

All industries

78%

Finance

91%

Marketing

71%

Healthcare

66%

Retail

63%

Manufacturing

55%

Read across: 78% of organizations already run AI in at least one business function. The industry lines below it aren't outliers — they're the baseline your customers now compare you against.

Sources

  • 78% — McKinsey & Company, The State of AI, global survey of organizations using AI in at least one business function.
  • 66% — American Medical Association, Augmented Intelligence Research, physician AI use survey.
  • 91% — NVIDIA, State of AI in Financial Services, firms deploying or assessing AI.
  • 71% — Salesforce, State of Marketing, teams using generative AI in their workflow.
  • 63% — NVIDIA, State of AI in Retail & CPG, retailers attributing revenue to AI.
  • 55% — Deloitte / MAPI manufacturing AI adoption studies, AI in production operations.

Figures are drawn from the most recent published editions of each survey; exact percentages vary by edition and cohort.

From the Desk

Field notes & analysis.

9 pieces · 6 categories

Healthcare5 min read

Why healthcare is AI's most demanding — and most rewarding — frontier

Two-thirds of physicians now use AI tools. The winners aren't the flashiest models — they're the best-integrated ones.

Read →
Finance5 min read

Finance quietly became the most AI-saturated industry on earth

Over nine in ten firms now deploy or assess AI. The lesson for everyone else is about data discipline.

Read →
Design4 min read

Generative AI ended the content bottleneck. Now taste is the moat.

When everyone can produce infinite adequate content, adequate content becomes worthless. Judgment is the scarce input.

Read →
Automation6 min read

The honest math of automation ROI

The most reliable returns hide in the repetitive middle — the twenty-minute task that happens ten times a day.

Read →
Strategy5 min read

How to choose an AI partner without getting burned

Three questions that expose depth, and one rule that protects you: start small, measure fast.

Read →
Retail4 min read

Pricing, recommendations, forecasting: where retail's 63% shows up

Retailers attributing revenue lift to AI rarely started with a moonshot. They started with three unglamorous loops.

Read →
Automation3 min read

Automating a broken process just produces mistakes faster

Map the process, fix the process, then automate the fixed process. Skip step two and 'AI' becomes your scapegoat.

Read →
Design5 min read

The interface is the product: designing AI people actually trust

Trust in an AI system is earned in the seams — how it shows its work, and how gracefully it hands control back.

Read →
Finance6 min read

Beyond fraud: the second wave of financial AI

Detection was the beachhead. Underwriting, regulatory reporting, and client service are where the frontier moved.

Read →

Get the research

The next report, before it's public.

We publish our field data and analysis to a short list first. No noise — a few emails a quarter, each one worth reading.

Continues on our contact page — no spam, unsubscribe anytime.