Best AI for Portfolio Managers

Portfolio management has an attention allocation problem that rarely gets discussed openly. The work that actually drives investment returns, the deep analysis of a business, the assessment of competitive dynamics, the careful evaluation of a management team's capital allocation track record, requires uninterrupted focus. But a portfolio manager's day is also full of things that demand time without requiring that same depth: synthesizing the twelve sell-side notes that came in this week on a position, drafting the quarterly letter for LPs, writing up the investment thesis for the investment committee, updating position memos as situations evolve.

AI tools are good at the second category. They're not good at the first. The PM who uses AI to write faster and research faster gets more of their time back for the work that actually differentiates their returns. The PM who expects AI to generate investment insight will be disappointed.

This guide covers three tools that fit real portfolio management workflows. The focus is on where AI produces genuine time savings for high-quality output, with honest notes on the limitations that matter.

The data and confidentiality context

Portfolio managers working at registered investment advisers, hedge funds, or asset managers face a specific constraint with AI tools. Positions, portfolio composition, non-public research, and any information that could constitute MNPI are not appropriate for consumer AI tools.

The practical rule for most portfolio managers: use AI tools for writing, public-source research, and synthesis of your own analysis, without pasting specific position data, non-public company information, or material non-public information into consumer tools. For institutional deployment with access to proprietary data, you need enterprise tools with appropriate data agreements.

With that constraint named: here's what helps.

1. Claude (claude.ai)

Claude is the right tool for portfolio managers who want a capable AI assistant for writing, synthesis, and working through analytical problems.

For investment memo drafting, Claude's value is concrete. Investment memos have a consistent structure: company background, investment thesis, key variant views versus consensus, valuation framework, risks, and catalysts. A PM who has done the analysis has the substance in their head; the time goes to writing the document. Give Claude a structured set of notes: the business description, the main points of the investment case, the key risks, the valuation approach and current multiples, and it produces a solid first draft. The PM edits for accuracy, sharpens the language, and adds the nuance that reflects the actual judgment. But drafting from notes rather than from a blank page cuts the time meaningfully.

For quarterly client letters, the same logic applies. The structure is consistent: what markets did, what the portfolio did, what changed in key positions, what the portfolio looks like now, and the outlook. Claude handles this structure well and produces professional, clear language from the bullet-point notes a PM would write for themselves anyway. Most quarterly letters that take two to three hours to write can be done in under an hour with AI-drafted first sections.

For research synthesis, Claude handles long documents well. Paste the relevant excerpts from multiple research reports, earnings transcripts, or industry analysis documents, ask Claude to identify the key themes, where sources agree and disagree, and what the key questions remain. It's a faster way to get through a stack of research than reading everything sequentially. The synthesis requires the PM's judgment to interpret, but getting oriented on what the research says is faster.

For thinking through analytical problems, Claude works well as a conversational tool. Describing a situation ("I have a position in X, the recent earnings call suggested Y, but the sell-side consensus is still modeling Z, help me think through what the discrepancy implies") and working through the implications is a useful thinking exercise. Claude doesn't know your position or the company, but it's a reasonable thinking partner for structuring the analysis.

The standard data caveat: don't put specific position sizes, non-public company information, or anything that could constitute MNPI into the standard Claude.ai plan. Use it for public-information synthesis and writing based on your own analysis.

Best for: Investment memo drafting, client letter writing, research synthesis from public sources, and analytical problem structuring. Pricing: Free tier available; Claude Pro at $20/month.

2. Perplexity

Perplexity covers the external research work for portfolio managers who want fast, cited answers on current company and market developments.

The specific use cases are: company news and press release summaries, recent analyst commentary and rating changes from public sources, sector trend research, management team background research, and quick fact-checking on specific claims or metrics.

For staying current on positions without reading every piece of coverage individually, Perplexity gives you a faster entry point. Ask "what has happened with [company name] in the last two weeks" and get a cited summary of the key developments. That's not a replacement for reading the actual press releases and filings, but it's a faster way to decide what requires your attention and what you can handle with a summary.

For research on new position ideas from public sources, Perplexity helps with initial orientation: the company's business model, the competitive landscape, recent analyst perspectives, and the key questions worth investigating further. This research is based on public information and appropriate for consumer tools.

At $20/month for Perplexity Pro, it's worth having as the external research layer in a PM's toolkit.

Best for: Current company news, public analyst commentary, sector research, and initial position research based on public sources. Pricing: Free tier available; Perplexity Pro at $20/month.

3. Glean

Glean is for portfolio managers at larger firms where the internal research and prior analysis is the knowledge asset that's hardest to find.

The problem is familiar: an analyst wrote a detailed note on a company three years ago, and the PM working on a similar situation doesn't know it exists. The investment committee reviewed a closely related opportunity last year and reached a conclusion about the industry's dynamics, but the memo isn't findable quickly. Prior portfolio manager reports on a sector contain institutional context that's relevant to a current decision but isn't in anyone's head.

Glean connects to enterprise document storage, email, and knowledge management tools. It indexes everything with permissions intact and makes it searchable in plain language. A PM can search "European consumer discretionary debt levels" and find the prior research and memos that address that specific question in seconds.

For funds where prior work represents significant intellectual capital, the ability to find it quickly has real value. Analysts building investment cases on companies that the firm has looked at before can find the prior work and build on it rather than starting fresh.

The access control layer is important for fund management. Position data, non-public information, and confidential research need appropriate access controls. Glean respects existing permissions so the knowledge retrieval capability doesn't create a confidentiality problem.

This is enterprise-only with custom pricing. For a small fund with a three-person team, Glean is overkill. For a larger asset manager with years of accumulated research that analysts can't find efficiently, it's worth a serious look.

Best for: Larger funds and asset managers where finding prior research, investment memos, and institutional knowledge is a recurring bottleneck. Pricing: Enterprise only; custom pricing.

How portfolio managers build an AI workflow

The tools on this list cover different parts of a PM's written and research work. Most portfolio managers who use AI effectively end up with a two-tool setup: Claude for writing and synthesis, Perplexity for external research. That's $40/month combined and covers most of the writing and research support a PM needs without requiring any institutional procurement.

The time savings compound over a year. Portfolio managers who build consistent AI workflows for writing and research report getting meaningful time back for the work that actually requires their investment judgment. The client letter that used to take all Thursday morning takes an hour. The investment memo first draft that took a full day now takes a focused afternoon. That recovered time goes somewhere, ideally into the analysis and thinking that drives actual investment performance.

Frequently asked questions

Can AI tools help with portfolio risk analysis and factor attribution?

For writing about risk and attribution, yes. Claude handles the narrative sections of risk reporting well, given the quantitative inputs. For the actual factor attribution calculations and portfolio risk analytics, purpose-built portfolio analysis tools handle that; AI tools write about the results rather than generating them.

What about AI tools that are specifically built for investment research?

Purpose-built investment research AI tools are emerging, with features like direct integration with SEC filings databases, earnings call transcript analysis, and company-specific knowledge graphs. Most are still maturing. The tools on this list are the ones that work reliably now for the writing and research tasks most PMs spend time on. As investment-specific AI tools mature, they'll likely be worth adding to or replacing parts of this stack.

How should I handle compliance review of AI-assisted client communications?

Check with your compliance officer. The SEC has issued guidance on AI use in investment adviser client communications, and many firms have added AI-specific review requirements to their compliance policies. At minimum, any AI-assisted client letter should go through the same compliance review process as any other client communication. Don't skip review steps just because AI drafted the text.