1 Journée type analyste financier — pré-market, résultats, calls, rédaction

Mode d'emploi : journée réelle analyste junior (alternant Master). Tu couvres 8-12 sociétés secteur banque/finance (ex: petites banques, fintechs, assurances). Outils : Bloomberg/Refinitiv, Excel modèles, call conférence, mail direction. Alternance : 3 jours banque d'investissement ou buyside + 2 jours école.

Lundi 7 h 30 – 12 h 30 — Pré-market, veille, review résultats weekend

Lundi 13 h 00 – 17 h 30 — Calls management, modèles, notes draft

Mercredi–Jeudi — Calls résultats, publication notes, présentation comité

Mardi–Jeudi (École) — Cas d'étude, théorie, présentations

2 12 scripts opérationnels — questions calls, contestations, accès management, valorisations

Mode d'emploi : scripts réels 2026. Tu parles CFO/investor relations, portfolio managers, directeurs d'entreprises. Variables entre [crochets]. Ton objectif : obtenir info précise, défendre thèse analyste, garder ton, rester pro.

Questions calls résultats (tone courtois, précision laser)

Défense valorisation + contestation gérant

Accès management et rédaction note

Publication notes + comité interne

Veille / stratégie sectorielle

3 5 procédures chiffrées — analyser rapport annuel 2h, comparables, DCF, note 4p, modèle Excel

Mode d'emploi : procédures exactes avec cas réels 2026. Chaque temps = minuté, prêt à appliquer. Tout analyste senior peut checker ton travail = reproducible.

Procédure 1 — Analyser rapport annuel en 2 heures (plan de lecture ordonné)

Cas pratique : ABC Assurance rapport annuel 2025 (100+ pages). Tu as 2h avant rédiger note. Vs. lire entièrement = 6+ heures. Ici = scalpel approche.

1. Lecture 15 min — Management discussion/analysis (MD&A) : Tu saute executive summary (marketing fluff), va droit MD&A = honest management view. Scan : results recap (revenues +5 %, profits +8 %?), major changes (new product? M&A? restructuring?), forward guidance. Prends notes rapides : key numbers + tone (confident? cautious?).
2. Lecture 20 min — Segment profit/loss breakdown : Entreprise ABC = 3 segments (Life, Property, Health). Lequel grow (+12 %)? Lequel decline (-3 %)? Margins = improving? Analyze profitability par segment (ROE, combined ratio). Donne toi intuition : « Life segment = cash cow, Health = growth engine, Property = stable. »
3. Lecture 20 min — Financial statements (P&L, Balance Sheet, Cash) : Focus sur : Revenues trends (top line steady or decline?), expenses ratio (fixed vs variable?), GAAP vs adjusted earnings (what's one-time items?), capital structure (debt/equity, insurance reserves). Compare Y-o-Y : 2025 vs 2024 = revenue trend continues? Margin expansion or compression?
4. Lecture 15 min — Risk factors + Legal : Fastest read but crucial. Key risks : climate change (catastrophe risk for insurer = make or break). Regulatory changes (Solvency II capital requirements = binding?). Litigation (any major lawsuits?). If risk = major, adjust valuation conservatively.
5. Lecture 20 min — Cash flow statement + Capital allocation : Operating cash = good or bad vs net income? CapEx needs? Dividends paid? Buyback plans? Capital position vs regulators (solvency ratio). Important = capital deployment philosophy = tell you dividend sustainability + future growth capacity.
6. Lecture 20 min — Notes to financial statements (accounting detail) : Normally skip = death by detail, but scan for : accounting policy changes (could be earnings massage?), impairments (major write-downs?), contingencies (hidden liabilities?). If see red flag = read deeper, ask investor relations.
7. Synthesis 10 min : Close report, summarize : revenues = +5 % stable, profits +8 % = margin expansion, capital = strong, dividend likely increase 5-10 %. Risks = manageable. Now ready to model.

Procédure 2 — Construire tableau multiples comparables (cas chiffré 5 sociétés)

Cas pratique : ABC Assurance vs peers : BNP, Société Générale, Crédit Agricole, AXA. Veux déterminer valuation ABC relative.

Analyse ABC vs peers : ABC P/E = 10.0x (peers avg 9.9x) = fair, légèrement premium (expected = smaller mid-cap = less liquid, justify +0.1x discount). ABC P/B = 0.95x (peers 0.88x) = slight premium = capital allocation decent (ROE 9.8 % vs peers 9.8 % avg) = justified. Conclusion comparables : ABC valuation = fair, not cheap, not dear.

Procédure 3 — Monter DCF simplifié (hypothèses/WACC/valeur terminale, cas chiffré)

Cas pratique : ABC Assurance DCF 2025-2030, terminal growth 2.5 %, WACC 7.2 %.

1. Données inputs : 2025E net income = 320M€, capex needs = ~10M€/an (insurance = low capex), tax rate = 25 %, dividend paid = 160M€ (50 % payout). Free cash to equity = net income - capex - debt changes + dividend = approx 310M€.
2. Growth assumptions 2025-2030 : Conservative = revenues grow 3 %/an (in line GDP), margins stable = earnings grow 3 %/an = 2025: 320M€ → 2030: 375M€.
3. WACC calculation : Cost of equity (using CAPM) = risk-free 2.5 % (10Y French government bond) + beta 0.9 × market risk premium 5 % = 7.0 %. Cost of debt = 2.8 % after-tax (ABC minimal debt = insurance model). Blended WACC = 7.0 % × (equity weight 98 %) + 2.8 % × (debt weight 2 %) = 6.9 % → round 7 %.
4. Terminal value (Year 2030+) : Year 2030 FCFE = 375M€. Terminal growth = 2.5 % (GDP). Terminal value = FCFE(2030) × (1 + growth) / (WACC - growth) = 375M€ × 1.025 / (0.07 - 0.025) = 8.6Md€.
5. DCF valuation : Discount cash flows 2025-2030 + terminal value to present (Year 0). FCFE each year : 2025 = 310M€ / 1.07 = 290M€, 2026 = 319M€ / 1.07^2 = 279M€, ... 2030 = 375M€ / 1.07^5 = 267M€. Sum discounted = ~1.55Md€. Terminal value discounted = 8.6Md€ / 1.07^5 = 6.1Md€. Total enterprise value = 1.55 + 6.1 = 7.65Md€. Equity value (no net debt) = 7.65Md€. Per share (100M shares) = 76.5€ per share.
6. Sanity check : DCF PT = 76.5€ vs current price 68€ = +12 % upside. vs. comparables = ~75€ (10.0x P/E × 7.5€ EPS). DCF vs comparables = aligned = good sign (not diverging = model robust).

Procédure 4 — Rédiger note d'analyse 4 pages (gabarit section par section)

Structure standardisée : Page 1 = Thesis + Recommendation, Page 2-3 = Valuation + Analysis, Page 4 = Risks + Financials.

1. Page 1 — Title + Investment Thesis (top half) : « ABC Assurance — Initiate With Buy Rating | €75 PT (12 % Upside) ». Thesis = 1 paragraph : « ABC well-positioned mid-cap insurer, benefiting from: (1) steady earned premium growth in Life (France market resilient), (2) strong solvency ratio 180% (capital for dividend growth/M&A), (3) valuation attractive relative ROE (9.8% vs peers 9.8%, but smaller liquidity discount unjustified). Consensus €73 PT slightly conservative on growth potential. »
2. Page 1 — Recommendation & Price Target (bottom half) : Table : Rating = « BUY », Price Target = « €75 (12M horizon) », Current Price = « €68 », Upside = « +12 % », Bull Case PT = « €82 (cost of equity decline) », Base Case = « €75 », Bear Case = « €65 (recession) ». Add concise text : « We initiate ABC with Buy rating, €75 PT based on DCF (50 %), Comps (40 %), Sum-of-parts (10 %). Key upside catalyst = capital deployment (dividend +5-10 % p.a., potential buyback). Downside if underwriting deteriorates or rates fall > 100bps. »
3. Page 2 — Business overview + segment breakdown (half page) : « ABC is French-listed insurer with €3.2Bn market cap, €320M 2025E earnings. Three main segments: (1) Life insurance (45% earnings) — steadily profitable, France retirement trend tailwind, margins 4-5%; (2) Property & Casualty (35%) — competitive, combined ratios 95-98%, exposed to inflation; (3) Health (20%) — niche player, small but stable 5% growth. » Add brief table : revenue/earnings by segment, growth rates, margins.
4. Page 2-3 — Valuation (1 full page) : « Our €75 PT reflects blended methodology : (1) DCF analysis (base case 2.5% terminal growth, 7.0% WACC) = €76.5/share; (2) Forward P/E multiples (peer avg 9.9x, ABC quality deserves 10.0-10.2x premium) × 2025E EPS €7.50 = €75-76; (3) Sum-of-parts (value each segment separately, roll-up) = €74/share. Sensitivity analysis : if WACC +100bps = PT €70 (bear), if terminal growth 3% = PT €82 (bull). Stock trades €68 = 9.1x P/E, discount to intrinsic = entry opportunity. »
5. Page 3 — Key investment points + catalysts (half page) : Bullet points : « • Earnings resilient through cycles (diversified portfolio) » « • Dividend history strong (5% yield, growing >5% p.a.) » « • Capital position allows shareholder returns (solvency 180%, regulatory min 100%) » « • Catalysts: Q1/Q2 earnings growth confirmation, capital deployment announcement, potential equity raise ruling out 2026. »
6. Page 4 — Risk factors (1/3 page) : « Downside risks to rating/PT: (1) Catastrophe event (hurricane/flood 1-in-10-year = underwriting loss ~€50-100M, earns down 15-30%), (2) Rate decline (100bps cut = NII headwind for invested assets, ~€30M income impact), (3) Competition (peer pricing pressure erodes margins 50bps), (4) Regulatory (Solvency II capital hikes require dilutive equity raise). »
7. Page 4 — Financial summary table (bottom, 1/3 page) : 3-year historical + 2-year forward : Revenue, Net Income, EPS, Book Value/share, ROE %, Dividend/Share, Dividend Yield, Solvency Ratio. Example : 2024A €312M / €7.20 / €38 BVPS / 9.5% ROE / €1.15 divi / 3.8% yield / 175%. 2025E €320M / €7.50 / €40 / 9.8% / €1.25 / 4.1% / 180%. Quick visual = shows growth trajectory + sustainability.

Procédure 5 — Construire modèle 3 états Excel (structure onglets, formules standard)

Modèle complet : 3 onglets (Income Statement, Balance Sheet, Cash Flow) + 1 Valuation onglet. Tout lié (balance = income, cash = balance, valuation = tous).

1. Onglet 1 « Income » : Columns = Historical (2023A, 2024A) + Forward (2025E-2030E). Rows = Revenues (top line), Operating Expenses (claims/commissions), EBITDA, D&A, Operating Income, Non-op items (investment gains), Taxes, Net Income. Formulas = each row grows at assumption % (e.g., revenues +3%/y growth). Bonus = calculate EBIT margins, tax rate % for validation.
2. Onglet 2 « Balance » : Assets = Cash, Investments (main for insurer), Receivables, Fixed Assets, Goodwill. Liabilities = Insurance reserves, Debt, Payables. Equity = Book value, Accumulated earnings. Key checks = Assets = Liabilities + Equity (always). ROE = Net Income / Avg Equity (validates). Capital ratios (debt/equity, leverage) track regulatory mins.
3. Onglet 3 « Cash Flow » : Operating cash = Net Income + D&A - Change in working capital. Investing cash = -CapEx (use assumption ~3-5% revenues for insurer). Financing = Dividends paid, Debt changes, Equity raise. Free cash to equity = Op CF - Investing CF - Debt service + New borrowing. This feeds DCF valuation onglet.
4. Onglet 4 « Valuation » : Top = assumptions box (growth rates, WACC, tax, capex %). Middle = DCF calculation (project FCFE 5-10 years, discount to PV, add terminal value). Bottom = Per share metrics (shares outstanding, implied share price, upside %). Add one Sensitivity table (PT varies by WACC/growth assumptions grid) = shows robustness.
5. Best practices : (1) Color code = blue inputs (you change), black outputs (formulas), grey = prior year data. (2) Row labels clear (avoid « Misc » = track every item). (3) All formulas linked (change revenue growth assumption = whole model updates auto). (4) Annual updates = copy prior year, adjust for actuals, roll forward forecast. (5) Test model : « If I double net income = does valuation double? » Yes = model is sound logic.

4 Tableau de bord de couverture — valeurs suivies, recos, performance, multiples clés

Mode d'emploi : template mise à jour hebdo (lundi matin). Chef coverage + sales consultent pour client pitches / position sizing. Tu = central source vérité.

Interprétation : Portfolio = 3 BUY (BNP, ACA, AXA), 2 HOLD (GLE, SAF) = bullish tilt (60%). Avg upside +9 % = reasonable, not stretched. Average P/E 11.6x vs sector avg 10.5x = small premium (justified = analysts overweight quality names). Yield 3.7 % attractive for income investors. Catalysts next 6m = mixed (earnings, capital, restructuring risks) = keep monitoring.

5 Grille de lecture express — 20 points rapport annuel, 15 min lecture

Mode d'emploi : checklist rapide avant modéliser. Si 15+ points = green, rédige note confiante. Si < 12 = warning, investigate deeper. Outil anti-surprises.

Score interpretation : 18-20 checks = Excellent (high confidence buy/coverage), 15-17 = Good (normal analysis), 12-14 = Fair (some concerns, need deeper dive), < 12 = Poor (red flags, consider avoiding or wait). Use as first screen before committing 2+ hours model/note.

6 8 prompts IA — synthèse résultats, veille sectorielle, scénarios, notes draft

Mode d'emploi : copie-colle chaque prompt dans Claude/ChatGPT, adapte variables en CAPS. Résultats = 80-90 % de note draft (tu affines final 20 %). Gain temps = ~1-2h/note.

Tu es equity analyst junior. Voici transcript appel résultats [NOM_SOCIETE] publiée [DATE].

Génère résumé exécutif (1/2 page, bullet points):
1. Highlights (revenues, earnings vs expectations)
2. Management guidance 2025-2026
3. Margin trends (positif/négatif drivers)
4. Capital allocation (dividends, buyback, CapEx)
5. Tone/confidence (was mgmt bullish or cautious?)
6. Key Q&A insights (unexpected questions/answers)
7. Financial forecast revisions (if any)

Format: bullet points concis, include specific numbers from call, no fluff.

[PASTE TRANSCRIPT HERE]

Tu es analyste secteur banque/finance. Mon coverage = [LIST 5 STOCKS].

Génère veille hebdo (1 page):
1. Market moves last week (sector index vs broad market)
2. Deals/M&A announced (consolidation, acquirers, pricing)
3. Regulatory news (ECB policy, Solvency II, competition)
4. Macro backdrop (rates, growth forecasts, inflation)
5. Earnings calendar (who reports this week, key expectations)
6. Impact on my coverage (which stocks affected positively/negatively?)
7. Thesis changes (do I need to adjust recos?)

Format: clear, actionable, avoid generic commentary.

Analyste. Stock [TICKER] [NOM], current price €[PRICE].

Build 3 scenarios for 2026 earnings:
— Base case: [YOUR BASE CASE ASSUMPTIONS — revenue growth, margin, EPS]
— Bull case: [UPSIDE DRIVERS — e.g., capital deployment, market share gains]
— Bear case: [DOWNSIDE DRIVERS — e.g., recession, margin compression]

Pour chaque scenario, calculate:
1. 2026E Earnings per share (EPS)
2. Forward P/E multiple (justify why higher/lower than base)
3. 12-month price target
4. Probability (e.g., Base 60%, Bull 20%, Bear 20%)

Weighted DCF PT = (Base PT × 60%) + (Bull PT × 20%) + (Bear PT × 20%)

Output: 3-scenario table + commentary on key assumptions/sensitivities.

Analyste covering [TICKER]. Peers = [LIST 3-4 PEERS].

Construis tableau comparables:
— Ticker, Price, Market Cap
— Revenue growth 2025E, EBITDA margin, Net margin
— P/E 2025E, PB ratio, EV/EBITDA
— ROE, Dividend yield
— Outliers (qui expensive? qui cheap?)

Analyse pour [MON STOCK]:
1. Valuation vs peers (is it cheap, fair, dear?)
2. Quality vs valuation (why different multiple — ROE, growth, risk?)
3. Fair value PT (based on peer average multiples + adjustments)

Conclusion: is relative valuation support buy/hold/sell?

Tu es analyste. Rédige note [TICKER] [NOM] initiation (BUY/HOLD/SELL):

Structure 3 pages:
— Page 1: Thesis + PT + Recommendation summary
— Page 2: Business overview, valuation methodology (DCF/Comps), key drivers
— Page 3: Risks, Financial summary table (3-yr history + 2-yr forward)

Tone: professional, data-driven, concis. Include:
— Specific numbers (don't generalize)
— PT rationale (why €X, not €Y?)
— Key catalysts next 12 months
— Risk weighting (probability)

Base case: [INSERT YOUR ASSUMPTIONS]
Bull/bear scenarios: [INSERT IF RELEVANT]

Output: full draft note, ready for light edit + publish.

Je dois évaluer [NOM_CFO] (CFO [COMPANY]). J'ai call transcript + MD&A writeup.

Analyze management tone/messaging:
1. Confidence level (bullish, cautious, evasive?)
2. Language patterns (concrete numbers or vague? forward-looking or backward?)
3. Key messages (what do they emphasize?)
4. Red flags (evasion of tough questions? blaming externals?)
5. Capital allocation clarity (is strategy clear?)
6. Risk acknowledgment (do they own problems or deflect?)

Output: 1/2 page assessment. Conclusion: do you trust this management team? Why/why not?

Equity analyst. [NOM_STOCK] earnings call [DATE].

Generate 6 must-ask questions:
1. Specific (e.g., « NIM guidance if ECB stays 4.5% vs 5% » not « how's NIM? »)
2. Based on recent data (prior quarter, competitor moves, macro)
3. Probing assumptions in guidance (force clarity on hypotheses)
4. Q&A format (question → expected answer → follow-up if dodge)

Add brief notes on each:
— Why you're asking (what it tells you for model)
— What « good answer » sounds like
— What « dodge » tells you

Output: Q&A document ready to use on call.

Analyst. Holding [RATING] rating on [TICKER] (PT €[PT]).

New data arrived:
[INSERT NEW INFO — e.g., earnings miss, mgmt guidance cut, merger rumor]

Do I change rating? Walk through:
1. What changed vs. old thesis? (magnitude of surprise?)
2. Impact on 2025-2026 earnings forecast (revise model?)
3. Impact on valuation (PT up/down? how much?)
4. Risk/reward at current price (still attractive?)
5. Catalyst timing (is this transient or structural?)

Conclusion: UPGRADE/DOWNGRADE/MAINTAIN? Explain conviction + timeline.

Output: 1/2 page analysis + clear recommendation.

7 Check-list avant publication note + processus validation comité d'investissement

Mode d'emploi : check-list à utiliser avant submit note final. Comité = approuve ou renvoie (« data wrong », « PT not justified »). Processus = mardi 14h (deadline vendredi 10h publication).

Check-list rédaction/données (avant soumis chef analyste)

• Données factules :
  ☐ Ticker correct + stock exchange
  ☐ Market cap, shares outstanding (vérif Bloomberg)
  ☐ Consensus estimates checked (vs I/B/E/S ou FactSet)
  ☐ Latest price & date captured
  ☐ Financial data (3yr history + 2yr forecast) = sources cited
• Valuation méthodologie :
  ☐ DCF assumptions explicit (growth, WACC, terminal, payout %)
  ☐ Comparables data sourced (peer set justified)
  ☐ PT = blend DCF + Comps + soum-of-parts (weights explained)
  ☐ Sensitivity analysis = table shows PT varies by key assumptions
  ☐ PT is consensus or diverge (if diverge = explain why)
• Contenu investment case :
  ☐ Thesis = 1 clear paragraph (no jargon)
  ☐ Catalysts = 3-5 specific (not vague « earnings » but « Q2 earnings likely beat xx % »)
  ☐ Risks = quantified where possible (e.g., « recession = earnings -15 %, PT → €X »)
  ☐ Competitors identified + why we like ours better
  ☐ Dividend sustainability checked (payout ratio, cash generation)
• Compliance + legal :
  ☐ No forward-looking hype language (« revolutionary » → « incremental »)
  ☐ Price target includes all risks, not just upside case
  ☐ Conflicts of interest disclosed (if any)
  ☐ PT time horizon stated (« 12-month »)
  ☐ Rating distribution stated (firm rating distrib = how bullish is the house?)
• Qualité writing :
  ☐ No typos, consistent formatting
  ☐ Figures = legible (chart/tables clear)
  ☐ Length appropriate (4-5 pages ideal, <3 pages = too brief, >6 = too long)
  ☐ Sources cited properly (url/date footnotes)
  ☐ Tone professional (not too casual, not too stuffy)

Processus comité validation (14h mercredi)

Format : Tu présentes 10 min, comité Q&A 10 min, vote 5 min. Présent = toi + chef analyste + 2-3 senior analysts + compliance.

1. Présentation 10 min : Tu pitches note rapide : (1) Thesis (30s), (2) Valuation PT justification (2 min — DCF logic + comparables), (3) Catalysts/Risk (1 min), (4) Ask for approval (1 min). Use slide deck (6-8 slides max). Tone = confident, not defensive.
2. Q&A comité 10 min : Membres ask : « Why not higher PT? » « This competitor cheaper, explain. » « What if recession? » « Is dividend safe? » Prepare answers = know your model cold. If asked question you don't know = « Let me verify, I'll come back » (better than guessing).
3. Objection handling : If member says « PT too high », ask : « What assumption you'd change? » (often they'll adjust own assumptions = discussion). If says « Thesis weak », ask to specify (vague objections = not serious). Defend via data, concede if legitimate concern (« You're right, I'll lower capex assumption, PT becomes €X »).
4. Vote : Comité votes : Approve / Approve with conditions / Reject. Typical = Approve (90 %) or Approve + « revise PT down 5 % » / « add 1 more risk discussion ». Reject = rare (means fundamentals/thesis broken or PT way off).
5. Post-vote : If Approve = schedule publication (next available slot = usually Friday 10am). If conditions = tu fix & resubmit lundi morning (quick review, usually rubber-stamped then). If Reject = deep discussion (rare = only if major errors or market timing concern).

8 Tableau « Combien ça coûte » — Bloomberg/Refinitiv, alternatives gratuites, salaires 2026

Mode d'emploi : benchmark réel 2026. Utile pour décisions budget + négociation outils. Tous prix HT France.

Outils données financières (subscription + coûts annuels)

Recommandation 2026 : Junior analyste = firm couvre Bloomberg + Refinitiv (non-negotiable). Solo freelancer / startup = Refinitiv Eikon solo (~20k€) = meilleur rapport qualité/prix. DIY budget = Zonebourse free (French reports) + Edgar (US filings free) + FactSet free lite = workable but 30+ hours/month manual extra. Bloomberg seul = prestige mais overpriced if just starting.

Alternative données gratuites (DIY minimalist toolkit)

DIY workaround : Analyst without Bloomberg = (1) Download annual reports SEC/company website, (2) Build comparables Excel manually (copy/paste data from Yahoo Finance + Zonebourse), (3) Use World Bank/trading view for macro, (4) Check consensus on Zonebourse/SeekingAlpha. Time = +4-6 hours/note vs Bloomberg user (= not viable for 5+ stocks coverage). Verdict : Premium tools = not luxury, = necessity for junior analysts to be productive.

Salaires analyst équité (France 2026) + bonus

Bonus analyst junior (2026) : Breakdown typique = +200 € si 5+ notes publiées (productivity), +300 € si 80%+ PT accuracy (forecasting skill), +250 € si zero compliance issues (integrity), +400 € si call performance (dealers give props) = possible +1150 € / trim = +4600 € / an (30 % bonus). Motivant. Top analyst (confirmé) = bonus = 40-60 % salary (much higher stakes).

Checklist décisions carrière (junior analyste)

• ✓ Coverage focus : Généraliste (multiple sectors) vs specialist (1-2 sectors deep)? → Specialist = longer learning curve but higher value long-term
• ✓ Stock universe size : 5-8 stocks (depth) vs 20-30 (breadth)? → Junior = 8-12 ideal (depth shows expertise, breadth = unfeasible quality)
• ✓ Publication frequency : Note every company quarterly vs selective? → Quarterly = disciplined, selective = higher quality (aim both)
• ✓ Client management : Institutional investors (bigger trades, higher stakes) vs retail (smaller, less demanding)? → Junior starts retail, graduate institutional
• ✓ Tools access : Firm provides Bloomberg? If not = career limiting (negotiate, or job-hop to place with tools)
• ✓ Mentor pairing : Does firm assign senior to review/teach? → Critical for skill development (50 % your success = mentor quality)

Mise à jour septembre 2026. Tous tarifs nets HT France. Taux de change/software pricing fluctue. Salaires = France métro, brut. Bonus = highly variable (individual + firm performance). Analyst role = structurally unchanged (humans beat algorithms still), but AI tools (Claude, ChatGPT) = becoming standard part of workflow (note drafting, research summaries). Future analyst = hybrid (AI-augmented) not replaced.