For institutions

Any listed company in the world.
A full report in about thirty minutes.

Valuatum has built equity and credit analysis software for banks, brokers and asset managers since 2000. AI Equity Reports is that same modelling, on demand, for any of the world’s roughly 50,000 listed companies — and steered by your analysts rather than written around them.

Request a free pilot See a real revision Read a finished report
~50,000
Listed companies worldwide you can order a report on
~30 min
From request to a finished PDF
5 rounds
Revisions where your analyst steers the model
25 years
Of Valuatum models behind the numbers
Where it fits

Most of an analyst’s week is not analysis.

Pulling the financials, building the model, checking the peer set, laying out the tables — it has to be done, and none of it is where the judgment lives. The engine does that part in about half an hour, for any listed company, from a real financial model rather than a summary of the annual report.

What that buys back is your analysts’ time, and it goes where it is worth more: questioning an assumption, working out what the current price already assumes, making the calls a model cannot make. The same week produces deeper work, because none of it went on wrestling a spreadsheet into shape.

What arrives

A complete PDF: company value map with enterprise value split by segment, reverse valuation showing what the current price already assumes, bear / base / bull scenarios, risks and catalysts, full financial statements and forecasts, a rating and a target price.

Open a finished one and judge it yourself →

What it is built on

Valuatum's own financial models, the same ones behind ProFinder and CreditRisk, which credit institutions, stockbrokers, investment banks and asset managers across Europe have used for two decades. The AI writes the analysis; it does not invent the numbers.

The methodology in full →

What your analysts do

They argue with it. Every generation carries revision rounds in which your own analyst challenges assumptions, swaps the peer set, adds a channel check the model cannot see, and pushes the report toward the house view. The last word stays with a person.

The revision loop, step by step →

In practice

Six ways institutions actually use this.

Each of these is a real job someone on your team is doing today, by hand, on a deadline.

Asset managers

The base work is already done

Financials, peer set, segment split, DCF and multiples arrive built. Your analyst opens a finished argument instead of an empty spreadsheet, and spends the day on the part that decides the call.

The output is consistent across companies, which matters more than it sounds: comparing two names is only meaningful when both were valued the same way.

Portfolio managers

The first pass before a meeting

A company calls, a broker pitches, a name jumps on results. You have an afternoon. Generate the report, read the reverse valuation, and walk in knowing what the price already assumes and which assumption has to break for the story to work.

Research desks

A second opinion your analyst has to answer

Run a report on a company you already cover and compare it with the internal note. Where the two disagree is where the thesis is actually load-bearing, and the model states its assumptions explicitly, so the disagreement is specific rather than a matter of tone.

Heads of research

Training that produces something usable

A junior analyst who has to improve a generated report learns faster than one building a model from an empty sheet, because the argument comes first and the mechanics are already there.

The revision prompts are kept with the report, so you can see what they changed and why.

Corporate finance & IR

Your own company, read from outside

Generate the report on your own listed company, and on the peers you are compared with. It shows what a disciplined outside model concludes from public numbers alone — usually the fastest way to find which part of your equity story is not reaching the market.

Advisory & wealth

Background for a client conversation

Structured, dated background on any listed company a client asks about, without pulling an analyst off their book for a day.

Client-facing material, branded output or a wider distribution is a conversation we are happy to have — say so and we will put terms in place.

How it works

One generation, then as many rounds of argument as it takes.

The engine builds the report unattended. Everything after that is your analyst reading a draft and telling it what is wrong — the same loop they would run with a junior, except the rework takes minutes.

How a report is produced Five stages: choose the company, the engine builds the report, your analyst reads the draft and comments, a revision round re-runs the model with your analyst's instructions, and the final report is delivered. Stages three and four repeat, up to five rounds. 1 Choose Any listed company. Worldwide. New ones imported on request. 2 MINUTES 2 Engine builds Financials, peers, segments, DCF, multiples, the write-up. UNATTENDED, 20–40 MIN 3 You read it Your analyst reads the draft and writes down what is wrong with it. YOUR PEOPLE 4 Revision round Assumptions, peer set, emphasis, your own channel checks. MINUTES PER ROUND 5 Final report PDF, dated, with the revision history kept alongside it. YOURS TO KEEP Repeat until the report says what your analyst thinks Up to five rounds per generation. Every round re-runs the model, not just the prose.

Stage 3 is where the value is added

The draft is finished work, not a skeleton — it is exactly what a paying customer receives. Your analyst is not filling gaps, they are disagreeing with a complete argument, which is a much faster thing to do well.

Stage 4 takes instructions, not edits

Your analyst writes what should change and why: a different WACC, a peer set that reflects how the company actually competes, a risk the model underweighted, a channel check nobody outside the firm has. The model re-runs with that input and the numbers move with it.

The loop is bounded on purpose

Five rounds per generation. Each one is a full engine run rather than a text rewrite, so the cap is real work rather than a paywall. If you find five is not enough for the way your team works, tell us; the number is set by us and can move.

The reasoning is kept

Every instruction your analyst sent is stored with the report. Six months later, when the call is being reviewed, you can see what the model concluded, what your analyst changed, and on what grounds.

One boundary worth knowing before you pilot this. Instructions steer evidence and assumptions, not conclusions. Your analyst can argue for a lower terminal growth rate, a different peer set, or more weight on a risk, and the valuation will follow. They cannot instruct a rating into existence or overwrite reported financial data — the rating follows arithmetically from the target price. If you want a SELL, you have to make the case upstream, which is the only honest way to get there.

That is the mechanism. Here is one that actually ran, with the comment and both versions of the numbers →

Worked example

One comment, and a HOLD became a BUY.

This is a real revision from this pipeline, not an illustration. The engine produced a Stora Enso report on 24 August 2026. An analyst read it and wrote down what was wrong with it. The report was regenerated against those objections. Both versions, the comment and the differences between them are kept together, which is why it can be shown at this level of detail.

1

What the engine produced

24 August 2026
RecommendationHOLD
Target priceEUR 9.50
Share priceEUR 9.82
Implied−3.3%

A competent report. It valued the Swedish forest estate by discounting the reported fair value, and it was explicit about why:

“This report applies a 30.4% discount to the EUR 8.5 billion reported fair value, producing approximately EUR 5.92 billion for the forest part in the target SOTP. The discount explicitly covers appraisal-to-market uncertainty, tax and transaction leakage, protected acreage, minority and ownership complexity, listed-company liquidity, the 18-year wood-supply commitment and the possibility that debt is allocated to Bergslagets Skogar.” Original report, forest valuation
2

What the analyst said back

Written in Finnish, rendered here in English

Four objections. None of them asks for a different rating; every one of them attacks a piece of evidence or an assumption. That distinction is the whole method.

The discount was priced for risks that cannot happen

“The company's valuation as a whole — both today and after next year's split into two separate companies — is completely independent of how debt is allocated between the two companies. The total debt is taken from shareholders' pockets in exactly the same way. The current owners will initially own both companies outright, so it makes no difference in which ‘pocket' the debt sits.”

The same objection covered the supply contracts: Stora Enso already buys wood at market prices from private owners and from its own companies, so an 18-year commitment prices nothing new. And no tax effect is coming.

The peer set was measuring the wrong thing

“These peers should be genuinely pure investment companies — not forest companies such as SCA or Holmen that have other businesses, because that muddles the valuation of the investment assets. Use companies whose business consists almost entirely of investment assets and the cash flow they generate, where NAV is clearly reported and the share price can be compared with NAV. Nordic companies are preferable.”

One number looked simply wrong

“‘The Swedish estate's annual biological growth is approximately 1%.' I think this is seriously wrong. Sources indicate that biological forest assets grow by roughly 3–5% at the latitude in Sweden where Stora Enso's forests are located. It should be removed, or its source should be disclosed and the report should explain why other sources give different figures.”

The comparisons were not comparable

“Why are Metsä Board, Billerud, Smurfit Westrock and Mondi described using different types of figures — percentages for some, tonnes for others, and then EBITDA margins for Smurfit and Mondi? Why not report tonnes, market share or margins on a comparable basis for all of them?”
3

What the revised report said

25 August 2026
RecommendationBUY
Target priceEUR 12.50
Share priceEUR 9.82
Implied+27.3%

The rating was never instructed. It moved because the forest discount fell from 30.4% to 12.5% once the objections were accepted, and the rating follows the target price arithmetically. On the debt point the revised report says the opposite of the original:

“The demerger's allocation of debt between Stora Enso and Bergslagets Skogar is not an economic value driver for shareholders who initially own both companies. Combined enterprise value less combined net debt produces the same combined equity value whether debt is placed in the forest company, the industrial company or split between them.” Revised report, forest valuation
“The most relevant evidence for a post-demerger valuation discount comes from companies where reported NAV and market capitalization are observable, not integrated forest companies whose manufacturing earnings obscure asset pricing.” Revised report, on the new peer set

Line by line

Every row below is a difference the system recorded between the two versions, not a summary written afterwards.

ItemBeforeAfter
RecommendationHOLDBUY
Target priceEUR 9.50EUR 12.50
Implied upside−3.3%+27.3%
Forest discount30.4%12.5%
Forest value in the bridgeEUR 5.92bnEUR 7.4375bn
Forest peer setSCA and Holmen, mixed with operating-business metricsPure NAV reporters: Investor AB 12.4%, Industrivärden 9% five-year average, Lundbergföretagen 13.5%, with HAL Trust and GBL as wide-discount cases
Biological growthAn unqualified 1% a yearA 1.0–4.3% evidence range including Stora Enso's own reported 3.3%; the 1% figure treated as a possible net-growth measure
Debt allocationListed as a possible value leakEUR 3.718bn consolidated net debt deducted once; allocation between two initially shareholder-owned entities treated as value-neutral
Industrial divisionsValued on EBITDAValued on modelled EBIT
Bear / base / bullEUR 9.31 / 9.51 / 9.97EUR 12.43 / 12.54 / 12.87
The reports themselves Original PDF 53 pp · 24 Aug Revised PDF 59 pp · 25 Aug

Both are the complete reports as they were produced, not extracts. Every figure in the table above can be checked against them.

Working together

Three ways in. Start with the first one.

They are not exclusive, and the first one costs nothing, so there is little reason to start anywhere else.

Option 1 · No cost

A limited pilot

Tell us who you are and how much research you do. We agree a number of generations your team can run internally, at no charge and with no card, on companies you choose.

  • Full reports, not a demo version
  • Revision rounds included
  • Internal use, named seats for your team
  • No card, no auto-renewal, no obligation
Request pilot access
Option 2 · Per report

Buy at list price

No agreement needed. A ready report from the catalogue is €20; a freshly generated one on any covered company is €50, delivered in about half an hour.

  • €20 ready report, €50 fresh report
  • Any listed company, imported on request if new
  • API access and volume terms on request
  • Client-facing and branded output on request
See pricing
Option 3 · You earn

Put your analysts on the platform

Your analysts publish their reports under their own names, priced by them, with half of every sale paid back. It costs nothing to join and builds a public, dated track record for the people who write it.

  • 50 % of every sale to the analyst
  • The analyst sets the price
  • Free membership, roughly an hour per report
  • Published under a real name, dated
How publishing works

Option 3 has a condition worth stating plainly: published means public. If your compliance function cannot let named staff publish research outside the firm, take options 1 and 2 and leave option 3 alone. The analyst terms are written down in full →

Free pilot

Request limited access for your team.

We size the pilot to the organisation, so we need to know a little about yours. Four facts are enough: who you are, where to find you, how much equity research you produce in a year, and how many analysts produce it.

1

You send the form

It reaches a person at Valuatum, not a queue. Nothing is charged and no account is created by sending it.

2

We check the organisation and reply

Usually within two working days. We confirm the number of generations, the seats, and any coverage we cannot provide before you spend time on it.

3

Your team starts generating

Seats are set up in the member area. Your analysts pick companies, run the revision loop, and keep the reports internally.

Prefer email? contact26@valuatum.com reaches the same people.

Before you ask

What this does not do.

Worth reading now rather than in week two of a pilot.

Banks and insurers are not what the models are built for

Enterprise value, DCF on free cash flow and EV multiples describe an operating business, not a balance-sheet one. Ask before you build a pilot around financials, and we will tell you honestly what a report on one is worth.

It is not investment advice

The reports are research material. They carry a rating and a target price because a valuation has to conclude something, but nobody at Valuatum knows your mandate, your holding period, or your client. The decision, and the responsibility for it, stays with your analyst.

Data is as good as what is published

The model reads reported financials. It cannot see an unannounced contract, a management change nobody has filed, or what a competitor said to your salespeople last week. That is precisely the part your analysts add in the revision rounds, and the reason the loop exists at all.

Start with the companies you have been meaning to look at.

A pilot costs nothing and tells you more in an afternoon than a call will.

Request pilot access Read a finished report first

Questions, or a procurement process to get through? contact26@valuatum.com