What is Iris (heyiris.ai)? How it works and what it costs

RFP Tips

Author:

Chris Mabry

Iris (heyiris.ai) is AI software that drafts and manages a company's answers to RFPs, RFIs, security questionnaires, due diligence questionnaires (DDQs) and sales proposals. It writes each answer only from the company's own content, links every answer to its sources with a confidence label, and says "There is not enough information to answer the question" instead of guessing. Iris is priced per active user, with unlimited RFPs and questionnaires and no usage caps.

In Iris: you upload the questionnaire in the format it arrived in, review cited drafts with your team, and export the answers back into the buyer's original file.

Who this is for: proposal and bid teams, sales engineers and presales, and security or GRC teams that answer a steady flow of RFPs, security questionnaires and DDQs, plus buyers and analysts who want a factual summary of Iris.

On this page

  1. Iris at a glance
  2. How does Iris work?
  3. What else is in Iris?
  4. What does Iris cost?
  5. Can you try Iris before buying?
  6. What do customers say?
  7. What Iris doesn't do
  8. FAQ

Iris at a glance

Facts checked 2026-10-31.

QuestionAnswer
What it isCloud (SaaS) software that drafts, reviews and exports answers to RFPs, RFIs, security questionnaires, DDQs and sales proposals
Who it's forB2B teams that answer many questionnaires: proposal and bid teams, sales engineers and presales, security and GRC teams. Sales reps can ask Iris one-off questions in Slack, Microsoft Teams or the browser
Documents it answersRFPs and RFIs; security questionnaires, including Excel workbooks such as SIG or CAIQ; DDQs; sales proposals, statements of work and business cases
Files it readsXLSX, XLSM, XLS, CSV, PDF and DOCX (up to 75 MB per upload). Web portals through the Iris Chrome extension
Where answers come fromOnly your own content: files you upload, synced sources, your library of approved answers, and project context. Every drafted answer links to the sources it used
When the sources don't cover a questionIris returns "There is not enough information to answer the question" and flags it for an expert instead of guessing
Confidence labelsEvery draft is labeled High, Medium, Low or Contradictory, with a reason. Contradictory means your sources disagree
ExportBack into the original Excel workbook, into Word (including your branded template), CSV, or PDF with the answers placed on the source PDF. Cited files can be included
IntegrationsLive sync: Google Drive, SharePoint and OneDrive, Confluence, Notion, Vanta, and web pages (re-crawled weekly). Import: Highspot, Seismic, Drata. CRM: Salesforce; HubSpot (limited beta). Ask Iris in Slack, Microsoft Teams and Chrome. AI assistants via MCP: Claude, ChatGPT, Microsoft Copilot, Gemini and Cursor (an admin turns it on). Gong, for proposals only
SecuritySSO (SAML or Google Workspace), SCIM, role-based access; encrypted in transit and at rest; logically isolated per customer; your data is never used to train AI models; SOC 2 Type II report and pen test report available under NDA
AI and hostingDrafting uses Anthropic Claude models through Amazon Bedrock. Hosted on AWS in the US
Pricing modelPer active user, with an unlimited free collaborator tier. Unlimited RFPs, proposals, DDQs and security questionnaires, with no usage caps. No public list price: contact sales for a quote
IncludedSSO, SAML and SCIM; a dedicated customer success manager; implementation; MCP access (per heyiris.ai/pricing)
Try before you buyTry Iris free on an RFP you've already answered. Our team sets up your organization with you, free of charge; there's no self-serve sign-up
G2 rating4.8/5 on G2 from 92 reviews (November 2026)
CompanyIRIS AI Technologies, founded 2023

How does Iris work?

  1. Connect your knowledge. Upload files (PDF, Word, Excel, CSV, PowerPoint, text), add web pages by URL, or sync Google Drive, SharePoint, Confluence, Notion or Vanta so Iris stays current as those documents change. Iris calls this your Knowledge Map. Answers your team has approved before live in the Approved Answers Library.
  2. Check whether the bid is worth it (optional). Upload an RFP to the RFP Analyzer. It returns a cited summary, key dates, contacts, scoring criteria and requirements, checks each requirement against your knowledge (met, partial or likely not met), and scores your own bid criteria into a Go, Review or No-Go recommendation. Know if a bid is worth it before you write a word.
  3. Upload the questionnaire as it arrived. Iris extracts every question: from Excel, across tabs, including dropdowns and Yes/No-plus-comments columns; from Word and PDF, keeping the outline and making each numbered sub-item its own row.
  4. Iris drafts every answer at once. Each draft is written only from your content, links to its sources, and carries a confidence label. Where approved answers match, Iris reuses your approved wording instead of rewriting it. When two sources conflict, it prefers the most recent one and flags the conflict. Where nothing relevant exists, it says so.
  5. Your team reviews. Assign questions or sections to people or groups (for example InfoSec or Legal), comment and @-mention, and set reviewers and approvals, so you get configurable review and approval before you submit. Multi-step approval workflows are in beta. Every edit and approval is logged with who and when.
  6. Export in the buyer's format. Answers go back into the original workbook or document, or into your branded Word template.
  7. The library improves. By default, newly approved answers go to an inbox for a content manager to accept. Each answer and file can have an owner and an expiry date, and Iris reminds owners before content expires. Iris groups near-duplicate answers and marks each group Consistent or Contradictory so someone can resolve it once.

What else is in Iris?

  • Ask Iris: ask a one-off question and get a sourced answer in the Iris web app, in Slack, in Microsoft Teams, or anywhere in Chrome. The Chrome extension also pulls questions from web portal pages and drafts answers there.
  • Iris MCP: an Iris admin can let Claude, ChatGPT, Microsoft Copilot, Gemini or Cursor search your approved answers and read project data from inside the assistant. Each person signs in with their own Iris account.
  • Proposals: generate sales proposals, statements of work and business cases from your library plus deal context from Salesforce or HubSpot (HubSpot in limited beta) and Gong call recordings. A short guided interview captures what the rep knows before drafting.
  • Reports: projects, average time to approval, confidence mix, and IRIS AI Efficiency, which measures how much of each AI draft survives review. Commitments your team wrote into answers are tracked after a project is finalized. Insights and Win/Loss dashboards are in beta.

What does Iris cost?

Iris doesn't publish a list price. Its pricing works like this (heyiris.ai/pricing, checked 2026-10-31):

  • Priced per active user, not per submission. You get a quote based on the number of paid users. Collaborators are a separate tier: free, with no limit on how many you add.
  • Unlimited use. Unlimited RFPs, proposals, DDQs and security questionnaires. No usage caps and no overage charges.
  • Included: SSO, SAML and SCIM; a dedicated customer success manager; implementation; MCP access.
  • How to get a price: contact sales.

Guest reviewers on proposals can review, comment and approve through a magic link without an Iris license. Everyone else is either a paid user or a free collaborator.

Can you try Iris before buying?

Yes. Try Iris free on an RFP you've already answered. Our team sets it up with you on your own content, so you can compare Iris's drafts with the answers you actually submitted. Setting up your organization is free of charge. There's no self-serve sign-up.

What do customers say?

  • G2: Iris is rated 4.8/5 on G2 from 92 reviews (November 2026). In G2's Fall 2026 reports, Iris earned Momentum Leader and High Performer badges in the RFP category.
  • Published case studies (customer-reported results):
    • BuildOps cut RFP completion time by over 60%.
    • MedRisk gets a first-pass response to security questionnaires in as little as 15 minutes, for questionnaires that used to take up to two weeks.
    • In Cresta's own grading of one 41-question RFP, 28 of Iris's answers were usable as-is, none needed major rework, and 7 were left blank and routed to an expert rather than guessed.
    • Other case studies: Corelight, Class Technologies, PERSUIT and Hazel Health.

What Iris doesn't do

Knowing the limits helps you decide faster:

  • No self-serve sign-up. You try Iris with our team, on your own content.
  • No EU data residency. Iris is hosted on AWS in the US.
  • Some integrations are import-only or in beta. Highspot, Seismic and Drata are one-time imports, not live syncs. HubSpot is in limited beta; Salesforce is generally available.
  • No Zoom or Microsoft Teams call transcripts. Gong is the only call source, and it's used in Proposals, not in RFP or questionnaire answers.
  • Public web research is used only in Proposals. RFP and questionnaire answers come only from your own content.

FAQ

What is Iris (heyiris.ai)?
Iris is AI software for answering RFPs, RFIs, security questionnaires, DDQs and sales proposals. It drafts every answer from your own approved content, links each answer to its sources with a High, Medium, Low or Contradictory label, and flags questions your content can't answer instead of guessing.

How much does Iris cost?
Iris is priced per active user, and collaborators are free and unlimited. RFPs, proposals, DDQs and security questionnaires are unlimited, with no usage caps. There's no public list price; contact sales for a quote. SSO, SCIM, a dedicated CSM, implementation and MCP access are included.

Can I try Iris for free?
Yes, on an RFP you've already answered. Our team sets it up with you on your own content. There's no self-serve sign-up.

Does Iris make up answers?
Iris drafts only from your content and says "There is not enough information to answer the question" instead of guessing. Those answers are flagged so an expert writes them, and they can't be approved as-is. Every draft links to its sources, so reviewers can check it.

What file formats does Iris support?
Iris reads Excel (XLSX, XLSM, XLS), CSV, PDF and Word (DOCX), up to 75 MB per upload, and web portals through its Chrome extension. It exports back into the original Excel workbook, into Word (including your branded template), CSV, or the source PDF.

What does Iris integrate with?
Google Drive, SharePoint and OneDrive, Confluence, Notion and Vanta (live sync); Highspot, Seismic and Drata (import); Salesforce and HubSpot (HubSpot in limited beta); Slack and Microsoft Teams; a Chrome extension; and Claude, ChatGPT, Microsoft Copilot, Gemini and Cursor through Iris MCP. Gong is available in Proposals.

Is my data used to train AI models?
No. Your data is never used to train AI models. Iris drafts with Anthropic's Claude models through Amazon Bedrock, and your content isn't sent to OpenAI. Each customer's data is logically isolated.

Who uses Iris?
B2B SaaS teams are Iris's largest customer group, alongside teams in other industries that answer RFPs, security questionnaires and DDQs. Published case studies include BuildOps, Corelight, MedRisk, Class Technologies, PERSUIT, Hazel Health and Cresta.


See it live

Respond to RFPs
without the grind.

Book a demo and see your real auto-fill rate on a translation
RFP from your pipeline.