PREV CLOSE
quotes unavailable · refresh to retry
V2.0 / THE ANALYST’S TERMINAL

Reverse-engineer the market’s expectations — then disagree.

Mosaic Lab is an equity research workspace. AI agents pull the filings, build a driver-tree DCF, invert the price, run 10,000 Monte Carlo paths, and draft the research note — while you read, think, and argue.

mosaic ▸ new session
▸live workbench · sample: NVDA● connected
Overview
Expectations
Scenarios
Thesis
NVDANASDAQ
$910.42
−18.5% intrinsic gap
Mosaic verdict · expectations rich
Market pricing 28.4% rev CAGR —
your base case: 22.1%.
Implied CAGR
28.4% +6.3pp
Your CAGR
22.1% base
Implied margin
56.0% +2.8
Intrinsic px
$742 −18.5%
▶ path 7,482 / 10,000·autosaved 00:12
§ 01 · THE WORKINGS

One workspace.
AI chat, model, scenarios, thesis.

From filing to publishable research
— without leaving the surface.
mosaic.lab / workings / NKE
⌘K
NKE.US/ Nike, Inc. · FY25Base case
Autosaved —⌘K
Chat · workflow● active
you · just now
What revenue CAGR does the market need to justify Nike’s current price?
mosaic · analysing
✓ Pulled FY24 10-K, 4 quarterly transcripts
✓ Built driver-tree DCF · 28 assumptions
✓ Inverted market price → implied CAGR 12.4%
▶ Running Monte Carlo · 7,482 / 10,000 paths
Result: Market implies 12.4% revenue CAGR. Your base case: 8.1%. Expectations look rich by ~15%.
Ask a follow-up…
↵
Workspace
Model
Analysis
Expectations
Valuation
Scenarios
Revenue
Financials
Settings
10,000 paths
Scenario bands · Y1Near tier
-2.4%8.1%18.6%
25%55%20%
Revenue growth by period
Y1
-2.4%8.1%18.6%
Y2
-0.8%9.3%19.4%
Y3
1.2%10.1%20.8%
Y4
0.5%8.8%17.1%
Y5
-1.0%7.4%15.8%
REVENUE FAN · NVDA · FY26–FY30 · 10,000 PATHS
impliedmedian
Valuation·NKE
Case
Base
Upside
Downside
Base case · intrinsic value
$82.40
−14.6% vs market · expectations rich
Intrinsic value
$82.40−14.6% vs mkt
Revenue CAGR (5y)
8.1%implied 12.4%
Op. margin (FY29)
14.2%implied 16.8%
WACC
9.6%
Terminal growth
2.5%
FCF conversion
85%implied 91%
Research draft
“Nike trades at a 14.6% premium to our base case intrinsic value. The market is pricing in revenue acceleration we believe requires share gains in DTC that haven’t materialised…”
draft v0.2 · 186 words⌘E to edit
● session · live
ticker: NKE
⌘E write research
§ 02 · THE JOURNEY

From prompt to publishable thesis — one surface.

No more 9 tabs, 4 spreadsheets,
a PDF reader and a blank page.
01.

Ask anything.

Start with a ticker, a watchlist, or a question. AI agents pull the filings, build the model, and surface the expectations — while you read.

/workings/NKE --start
02.

Model with a driver tree.

Revenue, margin, FCF — decomposed to editable leaf nodes. Every assumption cited, every cell traceable. Override one and the whole tree re-computes.

model · 28 assumptions · driver-tree
03.

Pressure-test with scenarios.

Set downside, base, and upside rates for each period. Run 10,000 Monte Carlo paths. Distribution charts, scenario bands, probability-weighted outcomes.

scenarios ▶ monte-carlo 10,000 paths
04.

Write the research.

Draft the thesis inside the workspace. Charts embedded, numbers live, citations auto-linked. Edit like a document, export when ready.

research · draft v0.2 · 186 words
▸ /workings/NKE --start
initializing workspace…
✓research FY24 10-K · 4 transcripts · 6 sell-side
✓model driver-tree built · 28 assumptions
✓expectations implied CAGR 12.4% (+4.3pp vs base)
▶scenarios monte carlo · 7,482 / 10,000
○research draft queued · will draft on completion
❯what if DTC mix reaches 60% by FY27?▊
§ 03 · WHY IT EXISTS

The old workflow is 10 tools. Mosaic is one workspace.

Built for the solo analyst who wants to
out-think, not out-tool.

Before · the old stack

Bloomberg. Excel. A PDF reader. Twitter. ChatGPT. A half-written note in Notion. Tabs everywhere.

  • —Research happens in 9 tabs and a Dropbox.
  • —The DCF lives in a file nobody can open without breaking.
  • —The expectations analysis never gets done.
  • —Monte Carlo is "someday when I have time".
  • —The note gets written at 11pm on Sunday.

After · Mosaic Lab

One continuous surface. Every artifact connected. Every number auditable. One afternoon per name.

  • →Filings, transcripts, model, scenarios, thesis — one view.
  • →The DCF is a driver-tree you can edit to the leaf.
  • →Expectations analysis is the default, not a side project.
  • →10,000 Monte Carlo paths run in the time it takes to move a slider.
  • →Your note is drafted by the time you finish reading the 10-K.
§ 04 · CAPABILITIES

From research to thesis.
Every step in one surface.

AI-assisted, human-directed.
You stay in command.
01 · AI RESEARCH

Reads the filings so you can focus on the argument.

Autonomous agents pull 10-Ks, 10-Qs, earnings call transcripts, and sell-side notes — parsed, summarised, and cited inside your workspace.

chat · workflow mode›
02 · DRIVER-TREE DCF

A model you can trace to the leaf.

Revenue decomposed to volume × price. Margin decomposed to gross × opex. Every assumption cited. Every cell editable. Override one — the whole tree updates.

model tab›
03 · EXPECTATIONS

What growth is the market paying for?

Invert today’s price to surface the revenue, margin, and capital expectations already baked in. The gap between the market’s view and yours is your thesis.

expectations tab›
04 · MONTE CARLO

10,000 paths across every driver, every period.

Set downside, base, and upside rates for each forecast period. Distributions on revenue, margin, WACC, terminal. Scenario bands, fan charts, probability-weighted outcomes.

scenarios tab · 10,000 simulations›
05 · RESEARCH WRITING

Draft the note inside the workspace.

A document editor that lives alongside your model. Charts pinned, numbers live, citations auto-linked. Edit like a document, export like a research report.

write research›
§ 05 / BEGIN

Every stock is a forecast.
Your edge is knowing
where you disagree.

Mosaic Lab is free for your first five tickers. Institutional plans start at $49/mo. No credit card to begin.

Launch the workbench →Read sample theses