A research and strategy agent for the questions executives actually ask. It reads industry knowledge spanning 24 countries and 1,000 companies as primary sources, establishes where you stand in numbers, and builds the outlook that follows. Every figure carries its source and its as-of date.
Ergane Agent is in closed beta. Access is by invitation code only.
Producing indicators is not the point. Ask the question you actually have to answer.
Where do we stand among our peers, right now?
Place the company inside the distribution of its category and its market, and read its position from revenue growth, margins, share-price momentum and market capitalisation. Counted, not sensed.
Will the same way of winning still work in five years?
Read demographics and macroeconomic indicators, forty years of temperature records, and the revenue growth of the covered companies against one another, and separate the tailwinds from the headwinds.
What did the winner next door change in order to win?
Pick the companies winning and losing against the same structural problem, and compare what differed across their corporate-event histories and their results.
Which styles and techniques are gaining attention now?
Read the design vocabulary against the news collection, and carry it through to the brands and the listed companies behind them.
Three outputs built straight from the knowledge. Each carries the reference date of its own series, because the series do not share one.
| Country | Companies | Market value | 1-year price | Revenue YoY | Operating margin |
|---|---|---|---|---|---|
| France | 13 | ¥142.9tn | -7.6% | -1.1% | 15.7% |
| United States | 105 | ¥135.9tn | +3.0% | +2.2% | 5.1% |
| Japan | 172 | ¥57.4tn | +7.8% | +4.4% | 5.1% |
| United Kingdom | 28 | ¥42.9tn | +12.7% | +1.0% | 7.6% |
| China | 145 | ¥39.4tn | -15.5% | +3.4% | 9.4% |
| Switzerland | 8 | ¥38.5tn | +25.7% | -1.1% | 3.7% |
| India | 87 | ¥37.6tn | -14.0% | +11.0% | 9.4% |
| Spain | 5 | ¥35.4tn | +22.0% | +1.8% | 12.0% |
| Germany | 11 | ¥13.8tn | -2.8% | +4.0% | 1.7% |
| Mexico | 7 | ¥12.5tn | +3.2% | +1.7% | 8.4% |
| South Korea | 99 | ¥9.9tn | -10.3% | +1.9% | 5.5% |
| Taiwan | 88 | ¥8.8tn | -9.5% | +4.9% | 4.3% |
Not written for this page. These already sit in the knowledge — one line of conclusion taken from each. Ask the agent the same question and it rebuilds the analysis from the same data.
Differences in price-to-book are largely explained by differences in return on equity. The layer that makes things trades below book — valued under what it would fetch broken up.
The median five-year return was -5.5%. The 198 companies that halved outnumber the 142 that doubled — a barbell of a five years.
Excess return in the twelve months after an appointment: -12.6%. In the twelve months before: -27.3%. They were falling before the change, not because of it.
Companies in the six shrinking markets win 30.8% of the time, level with the 30.2% in growing ones. A falling population is not a verdict — it redefines what winning looks like.
Months averaging above 28°C across the 24 cities rose from 268 to 314. The heat has started entering temperate cities that had none.
Apparel exports: China ×0.92 against Vietnam ×1.80. Yet China still exported 5.1 times what Vietnam did in 2023 — the shift is real, the replacement is not.
Attention is decaying across the board (median ×0.77), which makes anything above ×1.0 already a relative win. Only 27 of 149 entries grew.
101 of 595 companies do not earn from their core business. None of the global majors do — the losses live in the smaller middle.
The difference is not the model's cleverness. It is the ground it stands on. Nothing begins with collection when the question arrives — the gathering, the structuring and the dating are already done.
Ask a web search how many companies are losing money on their core business and it cannot answer: nothing defines what is being counted. Here the population is fixed at 1,000 companies across 24 countries, and it can be counted on the spot.
Comparing revenue across borders starts with reconciling 24 currencies. Those rates are in place for 200 months, and every company sits on one row of 49 columns. Aggregation starts at the answer, not at the plumbing.
Reading forty years of temperature against company results, or tracing where orders moved from trade statistics — those questions only exist because both sides are structured the same way and dated the same way.
Figures carry their reference date; conclusions carry their sources. The verification step on the reader's side disappears, and that is where the hours actually go.
as of August 2026
What used to take weeks to gather and days to reconcile now starts at the question. What reaches the executive is not a faster summary — it is material that can be decided on.
Asked what happens next, this agent does not paraphrase an article or a company's own plan. It measures against 1,000 companies and builds its own view — including the conditions under which that view breaks.
Place the subject inside the industry's distribution — same category, same market, same role in the value chain — and actually compute where it sits. No diagnosis, no proposal.
Where a covered report has already framed the same question, use its measure — what it measured with. The figures themselves are recalculated here rather than copied across.
Results, announcements and plans dated after the knowledge are filled in from the web, cited by URL and kept separate from what the knowledge says. Never the web's opinion as the answer.
A conclusion carries the data it was read from, and the falsifying condition: what would have to happen for this view to change. The company's own plan and this agent's assessment stay apart.
Asked what to actually do, it answers in three parts: the diagnosis in numbers, two to four options, and one recommendation — each option carrying the conditions it works under and the risk that kills it. It does not recommend buying or selling; that question is recast in business terms.
An outlook is not there to be right. It is there to show how it would break.
Knowledge as of August 2026
Where your growth and your valuation sit among peers around the world, with the figures a board paper needs and the sources standing behind them.
How mild winters and severe summers, now the norm, work through your category mix. Forty years of city temperatures read against company results.
Which styles, techniques and materials are gaining attention, traced from the vocabulary to the brands and the companies that carry them.
Requests you can send word for word — from a check before a meeting to a single chart. Log in and the rest of the list is on the screen.