Nova
Log inGet started
AI training · Finance and accounting

AI training for buy-side portfolio managers and analysts

You already use terminals, brokers and assistants. What changes at expert level is the way you work: a model in which every number points back to its source, an earnings review that starts on its own, a grid that screens forty companies. The investment decision and the conviction stay with you.

Get startedTry the interactive demo

Free. Just your email. You try the real product yourself.

In brief

AI for buy-side portfolio managers and analysts, in brief

AI helps buy-side portfolio managers and analysts with tasks such as “Work through an earnings release,” “Query five years of a company’s reports,” and “Build a model where every number has its source.” Start with a single task, in Claude: allow 20 minutes the first time, and always review the result before you use it.

  • 9concrete use cases
  • 6recommended tools, 3 of them free to start
  • 15 to 60 minper use case, the first time

The tasks where AI helps most

  • Building and maintaining models
  • Reading earnings and calls
  • Monitoring the watchlist
  • Writing memos and preparing the investment committee
  • Preparing meetings with management teams and clients
  • Checking valuations before a decision
Use cases

9 concrete AI use cases for buy-side portfolio managers and analysts

For each task: how it goes today, then with the tool. And the time it takes the first time.

Claude · Attached files, in the conversation

Work through an earnings release

TodayThe release lands at 7am, the market opens at 9am, and you already need to say what surprised.

With ClaudeA variance table with the page of every number, and three questions to put to the thesis.

  • 20 min
  • every week

Gemini Notebook · One notebook per company

Query five years of a company’s reports

TodayFive fiscal years, three hundred pages each, and the answer hides in an appendix note.

With Gemini NotebookA history of capital allocation, where every number points to its page.

  • 20 min
  • as needed

Claude · Claude in Excel

Build a model where every number has its source

TodayThe model for a new stock is built by hand, tab by tab, and nobody remembers where a cell came from.

With ClaudeA model in your template, with linked formulas and a source for every historical number.

  • 45 min
  • as needed

Also

Hebbia · Matrix

Screen 40 companies in a grid

A forty-row grid where every cell points to its source document.

  • 60 min
  • as needed

Claude · Claude in Word

Write the investment committee memo

A memo in the house template, with contradictions in your notes flagged.

  • 30 min
  • every month

AlphaSense · Monitoring

Put a list of stocks on watch

A short list of dated, sourced facts, sorted by thesis.

  • 25 min
  • every day

Claude · Project

Prepare a meeting with management or a client

A one-page brief with the history, the questions and the topics to avoid.

  • 20 min
  • every week

Perplexity

Look for the counter-thesis, with sources

A list of opposing arguments, with their source, their date and the fact to watch.

  • 15 min
  • as needed

Daloopa · Excel add-in

Update a model’s numbers from their source

An updated history, every cell linked to its document.

  • 30 min
  • every month
One use case, start to finish

Claude for buy-side portfolio managers and analysts: work through an earnings release, step by step

  • 20 min
  • every week

What to check

A release can mix reported and adjusted figures. Open the page before using a number.

Work through an earnings releaseClaude · Attached files, in the conversation
  1. Download the release and the presentation from the company’s website.
  2. In Claude, click the + button in the input field, then “Add files or photos”, and drop in both files.
  3. Paste the prompt below, with your estimates in square brackets.
  4. Recalculate each variance from the release before you touch your thesis.

The prompt to copy

You are assisting a portfolio manager who follows [company], held in the portfolio at [weight]. From the attached release and presentation, build a table: metric, reported figure, my estimate, variance in value and in percentage. My estimates: [revenue, operating margin, free cash flow]. Then list the non-recurring items, the scope and currency effects, and what management says about the targets for the year, with the page. End with three questions these numbers raise for my thesis: [two-sentence summary of the thesis]. Give no position advice.

My path

What you’ll be able to do, one step after another

In Nova, “My path” starts from your level: one next step at a time.

  1. Talk to AI like a colleague

    The first moves, with free tools.

  2. Write faster, without mistakes

    In your jobWrite the investment committee memo

  3. Find an answer you can trust

    In your jobPut a list of stocks on watch

  4. Create an image or a presentation

    In your jobWork through an earnings release

  5. Never retype anything again

    In your jobQuery five years of a company’s reports

For a team: everyone moves forward in their own job. You see the team’s progress and who finished a training you ask for, never the rest of one person’s activity.

The tools

AI tools for buy-side portfolio managers and analysts, and what they cost

Claude

Free to start

The panel in Excel to build and check a model, the earnings and committee review agents, and connectors to your licensed data.

Recommended planFree to try, with no internal documents. Panel in Excel: Pro, $20 per month ($17 per month if billed yearly). For a company: Team, $25 per user per month ($20 if billed yearly).

Gemini Notebook

Free to start

A company’s annual reports, presentations and calls in one notebook: every answer points back to the page of the document.

Recommended planFree: 100 notebooks, 50 sources per notebook, 50 questions per day.

Perplexity

Free to start

A sector update or a counter-thesis with the source and date of every number, before you open a position.

Recommended planFree: 3 “Pro” searches per day and 1 “Research” report per month. Then: Pro, $20 per month (€22 on the French App Store).

Daloopa

Fundamentals and KPIs drawn from public documents, every number linked to its source, with an add-in that updates Excel.

Recommended planEnterprise subscription, see the vendor’s website.

AlphaSense

Search across millions of documents, monitoring boards with regular alerts and scheduled agents.

Recommended planEnterprise subscription, see the vendor’s website. A free trial is offered.

Hebbia

A grid where each row is a company or a document and each column a question, with the source documents attached to every cell.

Recommended planEnterprise subscription, see the vendor’s website.

Prices taken from the vendors’ official pages, dated on each tool page. These subscriptions are not included in Nova. Brands are named to identify the tools. No affiliation.

Ground rules

The rules to follow with AI in this job

Always do this

  • Inside information: no non-public information on a listed company in a tool your compliance team has not approved, or in a personal account. A meeting under embargo stays out of the assistant.
  • Every number used in a model or memo can be traced back to its source. An assistant can confuse a reported figure with an adjusted one.
  • Your vendors’ data (prices, consensus, estimates, fundamentals) is licensed: check in the contract what may go through an assistant.
  • The decision, the position size and the report to the client commit you and your management company: AI prepares, you decide and you sign.

Never hand this to AI

  • Deciding on a position, a size or a sale on the sole answer of an assistant.
  • A consensus, price or valuation figure given by an assistant without a source you have opened.
  • What a management team told you in a private meeting, in a tool compliance has not approved.
Questions

Frequently asked questions about AI for buy-side portfolio managers and analysts

Which AI tools should buy-side portfolio managers and analysts use?

Nova recommends 6 to start with: Claude, Gemini Notebook, Perplexity, Daloopa, AlphaSense, and Hebbia. 3 of them can be used for free to get started.

How does Claude help buy-side portfolio managers and analysts?

The panel in Excel to build and check a model, the earnings and committee review agents, and connectors to your licensed data. The plan to choose: Free to try, with no internal documents. Panel in Excel: Pro, $20 per month ($17 per month if billed yearly). For a company: Team, $25 per user per month ($20 if billed yearly).

Where should buy-side portfolio managers and analysts start with AI?

With “Work through an earnings release,” in Claude (Attached files, in the conversation). Allow 20 minutes the first time. A variance table with the page of every number, and three questions to put to the thesis.

Is this an accredited course that leads to a certificate?

No. Nova is software, not an accredited training provider: it issues no diploma or certification. It costs $22 per month for one person, and $149 per month for a team of up to ten people.

Go further

AI training: related jobs