1. The Co-Analyst

The Co-Analyst is an AI tool built for institutional investors. Consolidate insights from disparate data sources, generating answers faster and more reliably than interns or chatbots. The Co-Analyst always sources answers exclusively from original documents with clear citations, ensuring data integrity and trustworthiness. Uncover insights that can't be found on a Bloomberg screen.

On the home page, search for a company by ticker or name. This will bring you to the company's tear sheet. Use the tabs at the top to navigate

Tear Sheet

Shortcuts and overviews

Co-Analyst

Answers to your queries.

Filings & Transcripts‍

Find earnings and conference calls, presentations and SEC filings on this tab. You can access all content library data sources except web sources from this tab. Use this tab read and query any document that is not available by default in the Co-analyst, e.g., proxy statements, registration statements, and prospectuses. Open the document to ask a query about that document. See more on querying individual documents.

Financial Metrics‍

Profitability, valuation, balance sheet, cash flow, and growth metrics, powered by S&P Global Market Intelligence. Note that the Co-Analyst does not have access to these fundamental and market metrics.

Red Flags‍

Forensic risk score and red flags identified in SEC filings. Risk scores are only available to institutional tier subscribers. See more details here.

Saved & Suggested Queries‍

Suggested queries are popular query templates designed to help you get started. Saved queries are created by you. Save queries to reuse them later or run on other companies. Up to 5 saved or suggested queries can be run concurrently.

Project History‍

The panel on the left shows your full history of projects related to the company, including previously generated answers. Learn more about Projects here.

In the query box at the bottom, input your search term and submit. You will see the Co-Analyst stream a response to your query in real-time.

How to query the Co-Analyst

Use simple language. Do not add unnecessary words.

By default the Co-Analyst will automatically select the relevant documents to search for your answer.  Optionally, include instructions for your preferred source document, time period or output format in the query box or select filters manually using Advanced Filters.

1. The Co-Analyst sources answers exclusively from source documents. If the information is not in the sources you selected, the information will not be in the Co-Analyst’s answer.
2. Nudge the Co-Analyst to use the right time period. Use terms like “most recent” or in your query if you’d like information from the most recent quarter only, or use “trends” to run a multi-period query. For most companies, the Co-Analyst can retrieve up to ten quarters of quarterly data from press releases or four years of annual data from 10-Ks. Timeline expansion is coming soon.
3. The Co-Analyst does not have access to your previous queries when answering your current one (multi-turn chat). Do not ask up follow-up questions. Instead, re-run the query with a different instructions or using Advanced Filters.

Dynamic citations

Select any part of the answer and click on “Cite this text” to see exactly where the information came from.

Output sources

By default, Co-Analyst queries answers queries using earnings calls, earnings releases, 10-Ks, 10-Qs, and conference call transcripts. Go to the Filings & Transcripts tab to query other documents individually. See more on querying individual documents.

Advanced filters

Instead of letting the Co-Analyst select the sources and time period, you can opt to manually select sources using Advanced Filters.

When to use Advanced Filters:

• Repeatability: Use filters when you want to guarantee that the same sources and time period will be used when the query is re-run.  This may be more important for saved or shared queries.
• Performance improvement: Retrieval may improve for certain queries when extra instructions are removed from the query so that it only includes keywords.

If Advanced Filters are not used, the Co-Analyst will automatically identify the source that is most likely to contain the information needed to answer your question. You can also specify the relevant source to use directly in your query i.e. “annual revenue breakdown by segment trends per 10-Ks”.

Tips:

1. Only select the sources you need. Adding too many sources will increase latency and may impact the quality of the output.
2. Select sources which will have the information available. Co-Analyst always sources answers exclusively from original documents. If the information is not in the sources you selected, the information will not be in the Co-Analyst’s answer.

If you need to query a document that isn’t in the default Co-Analyst content library, navigate to "Filings & transcripts" to select the specific source document you need.

Query output actions - re-run, copy, share, save, inspect, delete

In the top right corner of the answer, there are several action items.

Projects

Projects are designed to organize your queries so that you can share them with colleagues, print cohesive output and keep track of your research tasks.

Naming: The Co-Analyst will assign a name to a project if you do not.  However, you can rename the projects if you prefer. In the Project History panel on the left, hover over the project and then click on the three dots that appear beside the project name. From this menu, you can rename your project.

Sharing: Click “Share” to generate a unique link that can be shared with your colleagues. Your colleague does not need a Hudson Labs account to access the answer. However, dynamic citations on this link are only available to Hudson Labs subscribers.

Export: You can also export the project into a PDF or delete it from your project history.

Deleting: You can delete individual queries in projects. You may want to do this before sharing or printing, if you ran a query multiple times to test different sources, for instance.

Access all your past projects related to the company with complete answer history in the panel on the left.

Querying individual documents

By default, Co-Analyst queries answers queries using earnings calls, earnings releases, 10-Ks, 10-Qs, and conference call transcripts. Go to the Filings & Transcripts tab and open a document to be able to query the Co-Analyst about that document. The Co-Analyst will source the answer exclusively from the open document. See details of our full content library here.

2. Profile and Settings

Company watchlist(s)

Manage your watchlists here or navigate to your profile and then select "Watchlists".

There is no restriction on the number of watchlists or the number of companies in a watchlist. To add or replace tickers in a watchlist, select them manually or upload a CSV or Excel file. See details of our coverage here.

Email alerts

Update your email preferences here or click on Email notifications in your profile.

Adjust the frequency of email notifications and the types of alerts. Choose to receive notifications for only the companies in your watchlists or only specific watchlists. Alternatively receive notifications for all companies within specific sectors or market cap ranges.

Profile and Subscription settings

Change your password under Profile. Change your subscription or payment method under Subscription.

‍

3. Forensic risk score & red flags (available on institutional tier only)

Hudson Labs is well known for our proprietary forensic risk scores. The scores are calculated using large language models that rely on non-numeric data (related parties, off-balance sheet risk, aggressive accounting etc.) to assess the likelihood of fraud or malfeasance. They provide a unique and high quality predictor of downside risk. Investors, insurers and law firms use the Hudson Labs risk scores to find and avoid potential bad actors. Risk scores are only available on the institutional tier.

Hudson Labs is well known for our proprietary forensic risk scores. These scores are unique because they rely entirely on non-numeric data in assessing the likelihood of fraud or malfeasance, providing a unique and significantly higher quality predictor of downside risk. Investors, insurers and accounting firms use our scores to find and avoid potential bad actors.

Hudson Labs risk scores are machine-learned, not additive. The scores are calculated using Hudson Labs’ AI-extracted red flags and reflect the interaction of these red flags/risks. Risk scores range from 1 to 100 (low risk to high risk).

Scores above 70 are considered high risk. Fewer than 10% of mid/large-cap issuers receive a Hudson Labs forensic risk score above 70. Companies with a high score have about a 1 in 3 chance of receiving an SEC enforcement action in the next three years. Scores from 60 to 70 are considered medium-high.

Risk scores are optimized for precision over recall. This means that high risk companies are very likely to experience future issues, but low risk companies may not all be in fact low risk.

Initial high risk scores tend to precede price collapse by at least 6 months. Risk scores are trained using 10+ years of SEC enforcement actions, regulatory investigations related to fraud, settled class action lawsuits and other target labels related to earnings management and malfeasance.

Hudson Labs risk scores predict fraud-related class action litigation. For example, we spotlighted Super Micro’s "hidden in plain view" risks two years before the stock freefall in 2024. See more case studies here.

Our risk analysis features cover filings since January 1, 2019 and are available for US companies only.

You can access the risk analysis features for any company under the “Red Flags” tab.

1. Forensic risk score: See the full history of forensic risk scores identified for each of the annual and quarterly filings for the company.
   
2. Most recent 10-K red flags: All red flags identified in the most recent 10-K filing by our general model. The model is trained to identify any verbiage that is associated with future probability of SEC enforcement action.
   
3. Auditor turnover: There are a number of reasons why a public company may choose to change its accounting firm, including cost or reputation concerns. However, changes in audit firms can also be caused by tension between the management team and the team reviewing their work. These changes are therefore often considered to be a red flag. Hudson Labs provides access to auditor turnover tracking data for past 10 years, derived from the PCAOB database.
   
4. Executive turnover: Hudson Labs uses proprietary AI modes to identify, tag and summarize turnover among executives and the board, as disclosed in 8-Ks. We track both appointments and departures (resignation, dismissal, retirement and death).
   
5. ‍Internal control weaknesses / Material weakness: Internal controls over financial reporting (ICOFR) are the systems and processes to make sure that the financial statements are correct. A material weakness in ICOFR means there was a failure in the company’s systems or processes that could have resulted in a misstatement in their financial statements that an investor would care about.
   
6. ‍Restatements: Companies disclose restatements to previous submitted financial statements in 10-Ks,10-Qs, and 8-Ks. All major restatements have  Companies are obligated to disclose any major restatements of financial statements due to a material error, often called “Big R” restatements, in an 8-K.‍
   
7. Non-timely filings: Non-timely forms e.g. NT 10-Ks are filed when a company is unable to meet the filing deadline set by the SEC. Non-timely forms may be an indicator of internal distress and therefore often cause or precede stock price declines‍
   
8. Going concern warnings: Substantial doubt about an entity’s ability to continue as a “going concern” exists when the business faces the risk of a cash crunch or liquidity crisis in the next 12 months. This risk can be caused by insufficient revenue from operations, a large debt obligation coming due or other cash flow pressures. In short, substantial doubt about a company’s ability to continue as a going concern means there’s a real risk of bankruptcy.‍
   
9. Concentrated related party risk: Hudson Labs uses proprietary AI models to identify higher-risk related party transactions and assess the concentration of related party risk across U.S. issuers. The use of language models means we can identify related party transactions, regardless of the wording or key words and assess relative severity, facilitating advanced due diligence and market monitoring.
   

4. Feeds

Access the Feeds and Risk Screens using the hamburger menu on the top right.

Our feeds are continuously updated as our AI models process new filings and call transcripts.

1. ‍Earnings calls: Find earnings calls added to platform as soon as our models process them. We also provide top three financial highlights from the earnings call.
   
2. Earnings release: Each earnings release added to the feed comes with a three point summary highlighting performance metrics and forward-looking commentary (where available).
   
3. Forensic risk analysis feeds: We provide real-time feeds for a subset of our forensic risk features. The feeds contain a concise summary of the disclosure, focusing on the most important details.
   
   • Material weakness
   • Executive turnover
   • Non-timely filings
   • Restatements
   • Going concern warnings
4. SEC comment letters: Track SEC comment letters as they are made public. SEC’s Division of Corporate Finance (DCF or “The Division”) is responsible for reviewing securities filings. DCF staff issues a comment letter if it believes a company can better comply with a particular accounting rule or disclosure. The comment letter often contains a list of items that request additional information or disclosure, a modification to the filing or disclosure, or that the company changes how it discloses something in a future filing.

By default, our feeds are filtered to show companies with a market cap above US$300 m. Click on the Filters button on the top right of the feed to change the filters. Click on the “Watchlists” tab to only see the feed for companies in one or more of your watchlists.

5. Risk screens (available on institutional tier only)

Use Hudson Labs’ proprietary risk screens for monitoring, idea generation and due diligence.

1. Forensic risk score: Companies with a Hudson Labs forensic risk score of 60 or higher on the most recent 10-K or 10-Q. Click on the company name to see the company’s full history of risk scores for annual and quarterly filings since 2019 (institutional tier subscribers only).
2. High risk score + price run-up: Companies with a Hudson Labs forensic risk score above 60 and a large recent increase in stock price. Updated daily. Click on the company name to see the company’s full history of risk scores for annual and quarterly filings since 2019 (institutional tier subscribers only).
3. Concentrated related party risk: Companies with high concentrated related party risk in the most recent 10-K, as determined by Hudson Labs models. Click on the company name to see the company’s full history of related party disclosure in annual and quarterly filings since 2019 (institutional tier subscribers only).
4. Customer concentration: Customer dependence as disclosed in the most recent 10-K. Public companies are required to disclose the details of any customer dependence in their annual report (10-K). Hudson Labs provides one of the only complete databases of customer dependence disclosure across all U.S. issuers. Click on the company name to see the company’s full history of customer concentration disclosure in annual and quarterly filings since 2019 (institutional tier subscribers only).
5. Control weaknesses: Companies that disclosed a material weakness or significant deficiency in internal controls over financial reporting in the last 12 months. Click on the company name to see the company’s full history of material weakness disclosure in annual and quarterly filings since 2019 (institutional tier subscribers only).
6. Restatements: Companies that reported a retrospective restatement of their financial statements in the last 12 months. Click on the company name to see the company’s full history of restatements issued by the company in 10-Ks, 10-Qs, and 8-Ks since 2019 (institutional tier subscribers only).
7. Non-timely filing counts: Find companies with a high number of non-timely filings. The screen includes companies that have filed at least one NT 10-K since 2019 and are considered to be active issuers. Click on the company name to see all non-timely filings by the company since 2019 (institutional tier subscribers only).
8. Going concern warnings: Companies that disclosed a substantial doubt about going concern in the last 12 months. Click on the company name to see the company’s full history of going concern warnings in annual and quarterly filings since 2019 (institutional tier subscribers only).
9. Executive turnover: Find companies with a high number of resignations or dismissals of named executive officers (NEOs). The screen includes counts of turnovers since 2022. Click on the company name to see the company’s executive and board member turnovers since 2019 (institutional tier subscribers only).
10. Auditor turnover: Find companies with a high number of auditor changes. The screen includes companies with one or more auditor changes in the past 10 years per PCAOB. Click on the company name to see all auditor changes for the company since 2014 (institutional tier subscribers only).
11. Smaller or foreign auditors: Companies with smaller US auditors outside the top 8 or foreign auditors. Click on the company name to see all auditor changes for the company since 2014 (institutional tier subscribers only).
12. Discussion of AI: Track changes in the frequency of artificial intelligence mentions in companies’ earnings calls over time. Updated annually.

By default, our risk screens are filtered to show companies with a market cap above US$300 m. Click on the Filters button on the top right of the feed to change the filters. Click on the “Watchlists” tab to only see the feed for companies in one or more of your watchlists.

Click Export on the top right to export in a CSV.

‍