The Constant and the Current
Lessons from our team on bringing products to market while the technology underneath keeps moving. We focus on the part of your job that does not change, leveraging the technology that does.
- 06
Who Controls the Materials Inside an Electric Vehicle
A handful of refiners, licensing desks and export quotas decide which electric vehicles get built, and every buyer, builder and backer in the supply chain now plans around them.
Read the issue - 05
Tariff Watch: What Tariffs Have Done to the Electric Vehicle Market
Two years of stacked tariffs have redrawn where electric vehicles and batteries are built, what automakers pay and how Chinese makers reach buyers. We trace the numbers, the history and the moves that follow.
Read the issue - 04
How Fast New Battery Technology Reaches the Road
New battery chemistries take years to travel from laboratory to car. We trace lithium-ion, LFP, solid-state and sodium-ion through dated evidence and show what learning curves and physics tell builders about timing.
Read the issue - 03
Where Clean Technology Capital Is Going, and Why Projects Are Delayed
US factory plans are being cancelled while global energy transition spending sets records. We follow the money from Michigan to Indonesia and Valencia, and explain what the shift means for builders, buyers and investors.
Read the issue - 02
One in Four: Where the Electric Car Market Really Stands
The world bought more than 20 million electric cars last year, and the US went its own way. Where the market stands in October 2026, how the numbers are made, and why the headlines rarely agree.
Read the issue - 01
The Constant and the Current
The founding issue: why the job of winning a customer holds steady while the technology underneath it moves, what every earlier wave of tools teaches about this one, and how our own companies prove the point.
Read the issue
The research library
The research behind our AI search work, published by LBOX.com. Every figure carries its source and its date, and the library grows as the engines change.
Where the clicks went
Fewer searches end in a click, and the visit now arrives later as a search for the brand the answer named. These pages show the evidence and how to measure it on your own site.
- 01
The Second-Search Hypothesis: Where the Lost Clicks Go
The buying decision moved into the AI answer, and the visit moved to a second step that goes to the brand the answer named. Six claims, with the evidence and status of each, and the journey mapped before and after.
- 02
Where the Clicks Went: The Commercial Guide to the Great Decoupling
The buyer now decides inside the AI answer and acts in a second step: a search for the chosen brand on Google, Amazon or the brand site. The five-step journey with evidence at each step, and what AI visibility is now worth.
- 03
The Great Decoupling: The Evidence
Who documented the decoupling and with what data: every credible click-loss study with its sample, method and limits, and the reporting events that distort the picture.
- 04
How to Measure the Clicks AI Answers Take From Your Site
AI answers in Google mean fewer searches end in a click. This guide explains the two ways to size that loss, why clean date windows and a comparison group matter, and how to read the result honestly.
How the engines choose
How a buyer's question becomes a shortlist of brands, and what each engine reads before it answers, drawn from the research digests LBOX.com keeps as the studies arrive.
- 01
How AI Search Decides Who It Recommends
How a question becomes the engine's own searches, why the shortlist is set from the model's memory before any page is fetched, and the measures of what a brand is up against.
- 02
Does Brand Strength Win AI Recommendations? What the Evidence Shows
Famous brands are often missing from AI answers. A review of 2026 research finds that search interest, third-party mentions and buyer-matched positioning predict recommendation far better than survey brand equity or ad spend.
- 03
What Makes an AI Engine Cite a Page: The First Controlled Evidence
A 252,000-trial experiment tests eighteen content factors one at a time. Price, a current date and relevance decide first citation, formatting does nothing, and related studies show where AI answers can be fooled.
- 04
Why Smaller Brands Go Missing in AI Answers: A 37,000-Run Audit
A large 2026 audit of AI recommendations found that many specialist and regional brands never appear at all, that leaders win less often than expected, and that identical questions return different brands half the time.
- 05
Where AI Models Think Your Brand Is From, and Why It Changes Rankings
A FAccT 2026 study found that AI models misjudge a brand's country of origin about one time in five, and that a stated user location can move who ranks first by up to 59%.
- 06
How AI Engines Choose Which Brands to Recommend
A controlled study across three AI models shows famous brands win every tie, a tiny quality edge breaks that advantage, and brands that ignore AI optimization can vanish once rivals adopt it.
Reading the numbers
What each measurement in this field is taken over, where measurement breaks, and what to say when someone challenges a figure.
- 01
Reading and Defending Our Numbers
What every number we report is measured over, what it is valid for, and what to say when someone challenges it. For anyone presenting AI search measurement to leadership, a client or a sceptic.
- 02
What AI Recommendations Do, How to Measure Them and How They Get Fooled
Recent studies on whether an AI recommendation changes what buyers do, why most confidence intervals for AI answers are too narrow, and how a single polluted web page can push a fake brand into an answer.
The 90-day playbooks
Week-by-week operating models for a 90-day AI search program, written for an agency, an in-house brand team, and a brand and agency working together.
- 01
The 90-Day AEO Delivery Playbook
The week-by-week operating model for a bounded engagement: twelve sprints across foundation, execution and optimization, roles, deliverables, phase-gate exit criteria and the warning signs to watch.
- 02
The 90-Day AEO Brand Playbook: An In-House Operating Model
The same ninety days for a brand running AI search itself: in-house roles, an internal business case, biweekly and monthly reporting, and production assets that build lasting capability.
- 03
The 90-Day AEO Partnership Playbook: Brand and Agency
The twelve-week spine with every deliverable split between what the agency produces and what the brand provides and approves, a shared workspace, written service levels and joint sign-off.
Glossary
Every term we use, defined once, with where it came from. Use it to settle what a number means before you act on it.