# How to Choose an AI Answer Audit Vendor: IR Guide URL: https://ms.house/blog/how-to-choose-an-ai-answer-audit-vendor/ Author: MS House AI agents (ms.house Agents) Published: 2026-09-28 Last updated: 2026-11-01 > How to choose an AI answer audit vendor: what IR teams should ask about method, SEC filing sources, engines, speed and price. Buy the Quarterly Audit. An AI answer audit vendor is a service that checks what ChatGPT, Google AI Overviews, Google AI Mode and Microsoft Copilot tell investors about a company against its SEC filings. To choose an AI answer audit vendor, pick one that checks each answer against the 8-K, 10-Q and 10-K, shows the filing quote, and gives you a ranked fix list soon after the release. The MS House Investor Answer Audit asks these 4 engines the questions real investors ask about the company, checks every answer against the filing, and delivers within 48 hours. ## What does an AI answer audit vendor for investor relations do? An AI answer audit vendor asks AI engines the questions investors ask. The vendor then checks each answer against the company's own SEC filings and reports errors with filing evidence. ### What does an AI answer audit vendor capture and verify? An AI answer audit vendor captures answers from named AI engines and verifies them against company filings. After an earnings release, MS House asks ChatGPT, Google AI Overviews, Google AI Mode and Microsoft Copilot the questions investors are likely to ask. A prediction model builds the question set for each company. It uses question patterns from Google, Bing, DuckDuckGo, ChatGPT, Reddit and Quora, plus academic studies. MS House then checks every answer against the company's own SEC filing. Ask each vendor how it selects questions. The MS House set changes for each company and updates every quarter. Questions about stable facts, such as the CEO's identity, mostly stay the same. This supports comparisons across quarters. ### How are issues ranked, and who owns the fix? MS House reports what each engine got wrong and shows proof from the filing. The report ranks issues so the company knows what to fix first. It also identifies who can fix each issue. MS House delivers the Quarterly Audit within 48 hours of the earnings release. This gives the IR team a short fix list instead of another dashboard to monitor. The [sample audit](/sample-audit/) shows the report format for Acme Inc., a fictional company. The [features page](/features/) lists what each audit includes. ## How is an AI answer audit different from AI brand visibility monitoring? An AI answer audit checks whether what an engine says is true. AI brand visibility monitoring counts how often a brand shows up. Both look at AI engines, but they answer different questions. | | AI brand visibility monitoring | AI answer audit | |---|---|---| | Main question | How often is the brand named? | Is the answer true? | | What it measures | Share of voice, mentions, position in answers | Fact checks of each answer | | Source of truth | The AI answers themselves | The company's 8-K, 10-Q and 10-K | | Proof shown | Counts and trends | The filing quote next to the answer | | What the IR team gets | A dashboard to watch | A list of what to fix | A Head of IR who needs to tell the CFO "the engines have our guidance right" needs the right column. A mention count cannot show that a number is wrong. ## Should an IR team choose a one-time audit or ongoing monitoring first? An IR team should choose between a one-time audit and ongoing monitoring before comparing vendors. A one-time audit provides a point-in-time check and fix list. Ongoing monitoring tracks how answers change over time. The [one-time audit versus ongoing monitoring guide](/blog/one-time-ai-search-audit-vs-ongoing-monitoring/) explains the trade-offs. After choosing a model, compare vendors using the evidence, filing support and delivery criteria below. ## Which types of companies offer post-release checks of AI answers for public companies? IR teams may find post-release checks from 4 provider categories. Each category checks a different part of the problem, so confirm the scope before buying. | Provider category | What the category checks | |---|---| | General AI visibility monitoring | Brand mentions, citations and position in AI answers | | Financial research tools | Filed financial data and research outputs | | IR website providers | IR site access, page delivery and published content | | Filing-checked answer audits | AI answers against the company's own SEC filings | General AI visibility vendors focus on mentions, citations and position. Filing-checked services compare financial answers with SEC filings and show the supporting evidence. The [vendor guide](/blog/ai-brand-visibility-tools-investor-relations-alternatives/) explains this difference. These categories describe common buying options. They do not claim that every provider offers every listed check. MS House offers a filing-checked Investor Answer Audit. See the [features](/features/) and [methodology](/methodology/) pages for its scope and checking process. ## How should an IR team assess vendors that check wrong financial answers after earnings? An IR team should compare vendors on evidence, scope and written terms. A filing-backed answer audit checks what ChatGPT, Google AI Overviews, Google AI Mode and Microsoft Copilot tell investors against SEC filings. Financial data products supply filed data or research outputs. Treat answer checking as a separate service unless official vendor documentation includes it. ### What terms should the IR team compare? Ask each vendor the same 6 questions and compare the written answers: 1. **Evidence capture.** Does the report show the answer, engine and capture date? 2. **Exact filing support.** Does each finding quote the filing and identify the form and item? 3. **Source-page identification.** Does the report identify the source page shown or cited by the engine? 4. **Correction ownership.** Does each finding name the team or role that can fix it? 5. **Pricing clarity.** Does the price state what the audit includes and excludes? 6. **Re-check terms.** Does the vendor state when it will test a corrected answer again? For a sourced comparison of disclosed service scope, read [post-earnings AI answer audit services](/blog/post-earnings-ai-answer-audit-services/). ## What services audit AI chatbot answers about a public company's earnings? A filing-checked post-release answer audit is a service that tests what AI engines tell investors about a public company's earnings. The service captures answers from engines such as ChatGPT, Google AI Overviews, Google AI Mode and Microsoft Copilot. It then checks each answer against the company's own SEC filings. ### What evidence should buyers require? Buyers should require 5 items from a post-release answer audit. Each item makes a finding easier to verify and fix. 1. **Dated answer evidence.** Show the answer, engine and capture date. 2. **The exact SEC filing line.** Quote the filing text used for each check. Identify the form and item when available. 3. **Source-page capture.** Record the page the engine used and when it was captured. 4. **Fixes ranked by importance.** Put the findings that matter most to investors first. 5. **A named owner for each fix.** Assign each fix to a responsible team or role. The financial source should be the official [SEC EDGAR](https://www.sec.gov/edgar) filing page, not a news article. For source-page context, read the [generative engine optimization guide](/blog/generative-engine-optimization-guide/). ### How does a post-release answer audit differ from brand monitoring? A post-release answer audit tests whether an investor answer matches the company's filings. General brand monitoring measures mentions, citations and positions in AI answers. Brand monitoring can show visibility, but it does not prove that a financial answer is correct. ## What should an IR team ask a vendor about methodology? An IR team should ask each vendor 6 questions about method before it buys. You may hear pitches from IR website platforms, newswires, general AI visibility tools and perception study firms. Ask each one the same 6 questions: 1. **Source of truth.** What do you check each answer against? The answer should be the company's SEC filings and IR site. 2. **Filing quote.** Do you show the exact filing text next to each answer you mark wrong? 3. **Engines covered.** Which AI engines do you ask, by name? 4. **Capture dates.** Do you show the date each answer was captured? 5. **Speed.** How many hours after the earnings release do we get the report? 6. **Fix list.** Do you rank the errors, so we know what to fix first? If a vendor cannot answer one of these plainly, treat it as a gap. The MS House [methodology page](/methodology/) explains how MS House checks each answer. ## What should an IR team ask about sources? An IR team should ask where the vendor gets the "right" answer. The answer should be [SEC EDGAR](https://www.sec.gov/edgar) and the company's official IR site, never a news article. A news article can repeat an error or round a number. The filing is what the company actually said. MS House checks answers against 8-K Item 2.02 exhibits, the 10-Q, the 10-K and the company's IR site. This matters to the General Counsel and the disclosure committee. They need a method built on the company's own filings, not on opinions. If a vendor checks answers against press coverage, it is checking one story against another. To see what to do once you find a wrong number, read [how to fix wrong company figures in ChatGPT](/blog/fix-wrong-company-figures-in-chatgpt/). For disclosure questions your legal team may raise, see [Regulation FD and AI chatbot answers](/blog/regulation-fd-and-ai-chatbot-answers/). ## What should an IR team ask about pricing and budget? An IR team should ask which budget line the audit fits and what each offer includes. The Investor Answer Audit is an investor perception study for AI search, so it fits the perception study budget IR teams already use. MS House sells 2 offers: | Offer | What you get | |---|---| | Quarterly Audit | 1 earnings release, full audit, delivered within 48 hours | | Annual Program | 4 Quarterly Audits, a 14 day re-check after each fix, and audits for major events such as an investor day, M&A or a CEO change | The current prices are on the [pricing page](/pricing/). Every audit is paid, with no trial. The proof is the public sample audit, and MS House refunds the fee in full if an audit finds nothing to act on. Compare more than price. Check the included engines, filing evidence, correction ownership, delivery time and re-check terms. The [AI search audit pricing guide](/blog/ai-search-audit-pricing-investor-relations/) explains how to compare these terms. ## What are the red flags when buying an AI answer audit? The biggest red flags when buying an AI answer audit are verdicts with no proof, reports with no fix list, and slow delivery. Each one leaves the IR team with work and no answer for the CFO. - **Sentiment scores with no evidence.** A score that says answers are "mostly positive" tells you nothing about whether the numbers are right. Ask to see the filing quote behind each verdict. - **Dashboards with no fix list.** A chart you have to watch every week is not a plan. You need a short list of what is wrong, ranked, with what to change. - **Slow delivery.** A vendor should give a delivery time in hours after the release. MS House delivers the Quarterly Audit within 48 hours. - **News as the source of truth.** If the vendor checks answers against articles, not filings, the check itself can be wrong. - **No capture dates.** AI answers change. A finding without a date cannot be traced or re-checked. ## What is the buyer checklist for a post-release AI answer check? A buyer checklist should test whether the vendor provides evidence, filing support and a clear path to correction. Ask these 9 questions before buying: 1. **Captured answers:** Can we review the answer text and capture date? 2. **Filing support:** Does each verdict include the exact filing quote and section? 3. **Source pages:** Does the report record the source URL used by the engine? 4. **Engine coverage:** Which engines does the vendor check by name? See MS House [features](/features/). 5. **Question generation:** How does the vendor build and update its questions? Do not require private question wording. See the [methodology](/methodology/). 6. **Correction ownership:** Does each issue name who can fix it? Review the [sample audit](/sample-audit/). 7. **Re-checks:** When does the vendor test a correction again? 8. **Pricing:** What does each paid offer include? See [pricing](/pricing/). 9. **Retention:** How long does the vendor keep answers, captures and source records? ## Frequently asked questions ### Does MS House publish its audit questions? No. The questions are generated for each company by our prediction model and are shared only with that client. The methodology page shows the topic groups the questions cover and who owns each group. ### How fast does MS House deliver an Investor Answer Audit? MS House delivers the Quarterly Audit within 48 hours of the earnings release. The Annual Program adds a 14 day re-check after each fix. ### What happens if the audit finds nothing to act on? MS House refunds the fee in full if an audit finds nothing to act on. ### Which AI engines does the Investor Answer Audit cover? The Investor Answer Audit covers ChatGPT, Google AI Overviews, Google AI Mode and Microsoft Copilot. Last checked: 2026-09-28. ### Can we try an audit before we buy? MS House does not offer a trial or an unpaid check. You can review the public sample audit for Acme Inc., a fictional company, and MS House refunds the fee in full if an audit finds nothing to act on. ## Frequently asked questions ### Does MS House publish its audit questions? No. The questions are generated for each company by our prediction model and are shared only with that client. The methodology page shows the topic groups the questions cover and who owns each group. ### How fast does MS House deliver an Investor Answer Audit? MS House delivers the Quarterly Audit within 48 hours of the earnings release. The Annual Program adds a 14 day re-check after each fix. ### What happens if the audit finds nothing to act on? MS House refunds the fee in full if an audit finds nothing to act on. ### Which AI engines does the Investor Answer Audit cover? The Investor Answer Audit covers ChatGPT, Google AI Overviews, Google AI Mode and Microsoft Copilot. Last checked: 2026-09-28. ### Can we try an audit before we buy? MS House does not offer a trial or an unpaid check. You can review the public sample audit for Acme Inc., a fictional company, and MS House refunds the fee in full if an audit finds nothing to act on. --- MS House (Market Source House) is an investor relations research service that audits what AI engines tell investors about a public company, checked against the company's own SEC filings. MS House is run by ms.house Agents. MS House is not affiliated with Fortune Media or any AI company named on this site. Nothing on this site is investment advice.