Conveyor Changelog

When a questionnaire goes back to Salesforce, it should look like the file you started with.

• What: Completed document-format questionnaires now export back to Salesforce as the repopulated original file, not a flat CSV.
• Why it matters: Your customers get answers in the format they sent, so the finished questionnaire looks polished and ready to share. No more reformatting a CSV by hand.
• How it works: When a Word questionnaire is completed and pushed back to Salesforce, Conveyor returns the repopulated original document. Spreadsheet exports are unchanged. PDF questionnaires return a summary document, since PDFs can't be repopulated directly.
• How to get started: Nothing to set up. It applies automatically to Salesforce exports.
• Status: Live now

Your security reviews don't all come from one place. Now your intake doesn't either.

• What: Connect more than one intake source to a single org, so you can triage questionnaires from systems like Salesforce and Jira together.
• Why it matters: Larger teams pull security reviews from different systems across the business. Now one Conveyor org can take them all in, so nothing slips through a second tool.
• How it works: Add a new triage source for your org. Each org can connect one source per type, for example one Salesforce source and one Jira source.
• How to get started: Reach out to your Conveyor team to add an additional intake source.
• Status: Live now

Updating your Trust Center shouldn't mean taking it offline.

• What: Stage layout and visual changes to your Trust Center as a draft, then publish when you're ready.
• Why it matters: Teams used to unpublish their whole Trust Center to make changes, or avoid updates entirely for fear of disrupting customers. Now you can edit in the background and ship with confidence, no downtime.
• How it works: Create a draft of your Trust Center layout or content, preview exactly how it will look with a shareable preview link, and publish when it's ready. Your live Trust Center stays untouched the whole time.
• How to get started: Open the Trust Center designer and start a draft.
• Status: Live now
• Learn more: Drafting changes to your Trust Center

The last step of a questionnaire shouldn't need a person.

• What: Set criteria that let Conveyor automatically return a completed questionnaire to the source system once it meets your bar.
• Why it matters: Analysts spend time on a final review even when an answer set is already complete. Define what "done" looks like and let questionnaires flow back on their own, freeing your team for the work that needs judgment.
• How it works: For questionnaires that come in through the Intake workflow, set criteria like a minimum answered rate or no flagged or unanswered questions. When a questionnaire clears your bar, Conveyor returns it to the source system automatically.
• How to get started: Set your return criteria in your Intake workflow settings.
• Status: Live now
• Learn more: Setting up auto-return for Intake

Finding the right visitor or request shouldn't take endless scrolling.

• What: The Contacts and Requests tabs in your Trust Center audience view now support richer filtering.
• Why it matters: As your Trust Center traffic grows, you can zero in on exactly the people or requests you care about, faster.
• How it works: Filter by criteria like email address, email domain, access status, and access groups to manage requests and visitors in a few clicks.
• How to get started: Open the audience tabs in your Trust Center and apply the new filters.
• Status: Live now

Your Trust Center should speak in your terms, not ours.

• What: Customize the agreement and legal text on your Trust Center, including the NDA confirmation language and the Terms of Service and Privacy Policy section.
• Why it matters: Not every company uses an NDA, and some need visitors to agree to their own confidentiality or legal terms. Generic "NDA" wording created pushback. Now the text matches what your legal team requires, so visitors agree without hesitation.
• How it works: Replace the standard NDA modal text, and swap Conveyor's Terms of Service and Privacy Policy for your own custom language, which supports markdown.
• How to get started: Reach out to your Conveyor team to update your Trust Center legal text.
• Status: Live now

Get finished questionnaires back in the format you need, on demand.

• What: New v2 API endpoints let you export filled questionnaires and grab a download URL.
• Why it matters: Teams building their own workflows can pull completed questionnaires straight into their systems, in the format that fits. Less manual exporting, more automation.
• How it works: POST to request an export, then GET to poll until it's ready and get the download link. Choose CSV (instant), XLSX (repopulates the original spreadsheet), or doc exports. Spreadsheet questionnaires route to XLSX automatically. Also available as MCP tools.
• How to get started: Use the v2 export endpoints with your API key.
• Status: Live now
• Learn More: Questionnaires API

Your RFPs already live next to your deals. Now they move with them.

• What: Salesforce users can submit multiple questionnaire files directly from an Opportunity, grouped into a single project.
• Why it matters: Sales teams kick off security reviews without leaving the CRM, and project status flows back to the Opportunity so everyone sees progress in one place. It's how Adyen runs its workflow today.
• How it works: Files are sent to Conveyor through the new Projects v2 API, ConveyorAI processes them, and project status moves from draft to started to completed right on the Opportunity. It's also available via MCP for teams building their own CRM integrations.
• How to get started: Install Salesforce managed package v1.51 and configure a Connected App.
• Status: Live now
• Learn more: Installing our Salesforce managed package

Your best answers should work harder than once.

• What: A new API endpoint lets you mark questionnaire answers as reusable (or not) for ConveyorAI.
• Why it matters: When an answer is reusable, ConveyorAI pulls from it to draft responses to similar questions on future questionnaires. Your team builds a sharper knowledge base without manual upkeep.
• How it works: Call PATCH /api/v2/questionnaires/questions/{id} to flag an answer as reusable. ConveyorAI then references it the next time a similar question comes up.
• How to get started: Available now for API key holders with questionnaire write access.
• Status: Live now
• Learn more: Questionnaires API reference

Using Conveyor inside ChatGPT or Claude just got a lot easier.

• What: The Conveyor MCP server now supports OAuth 2.1 sign-in, alongside API keys.
• Why it matters: Consumer AI tools like ChatGPT, Claude.ai, and Claude Desktop only accept OAuth-authenticated MCP servers. Now your team can connect Conveyor to the AI tools they already use, no API key juggling.
• How it works: Paste https://mcp.conveyor.com/mcp into your AI client's "Add custom connector" flow, log in, and authorize. On enterprise plans like ChatGPT Business or Claude Enterprise, an admin adds the connector once for the whole org.
• How to get started: Add the connector in your AI client and sign in with your Conveyor account.
• Status: Live now
• Learn more: MCP server setup

When a visitor opens a SOC 2 report, they shouldn't have to guess which products it applies to.

• What: Each document in your Trust Center now displays its associated product lines in the document viewer properties.
• Why it matters: Customers and visitors told us they were unsure which products a document covered, especially when an organization sells multiple products under one Trust Center. This removes that ambiguity.
• How it works: The viewer lists each product line individually, or shows "all product lines" if the document applies across your whole portfolio. No action needed from you, this displays automatically based on how each document is tagged.
• How to get started: Open any document in your Trust Center to see the new field in the properties panel.
• Status: Live now
• Learn more: Trust Center document settings

If you serve buyers in multiple languages, your NDA settings just got easier to manage.

• What: A redesigned NDA and Access settings page that surfaces translation status, signing requirements, and per-language NDA controls at the top level.
• Why it matters: You no longer have to dig into the English NDA settings to see what's live in other languages, or jump between pages to update them. Everything is visible and editable in one place.
• How it works: Open NDA and Access settings to see every language with its NDA status. Edit a language inline, or jump straight to the Trust Center Designer for that language with one click.
• How to get started: Go to Settings, then NDA and Access.
• Status: Live now
• Learn more: NDA Settings

Not every API key needs the ability to open any door.

• What: Scope API keys to specific resources and specific actions on those resources. Update permissions on existing keys at any time.
• Why it matters: As more teams plug AI agents and MCP clients into Conveyor, you need tight control over what those agents can do. Now you can issue a read-only key for reporting, a Knowledge-Base-only key for an editing agent, or a full-access key for trusted automation.
• How it works: When you create or edit an API key, pick the resources and actions it can access. The available scopes mirror the Public API documentation, so every endpoint can be locked down individually.
• How to get started: Go to API key settings to update existing keys or create a new one with custom permissions.
• Status: Live now
• Learn more: API key creation (Settings > Integrations)

Automate keeping your subprocessor list current, in real time.

• What: List, add, and remove subprocessors from your Trust Center profile through the Public API and MCP server.
• Why it matters: When you sign with a new vendor or sunset an old one, your Trust Center should reflect that the same day. Now you can automate it from procurement workflows, internal tools, or an AI agent instead of editing each entry by hand.
• How it works: Use the POST endpoint to add a subprocessor by vendor name. Fuzzy matching catches misspellings and suggests the closest matches so you can confirm the right vendor. Also exposed as MCP tools for Claude Code and similar clients.
• How to get started: Enable the "Trust Center Designer" capability on your API key in API key settings.
• Status: Live now
• Learn more: Subprocessor API reference

You'll lose your head over this one. You can now manage your Customer Trust workflows via the API or from your favorite AI tool via our MCP.

• What: PATCH a questionnaire to update its status, due date, assignee, or notes through the Public API.
• Why it matters: MCP clients and integrations can now push questionnaires through review without anyone clicking through the UI. Less swivel-chair work for analysts, faster turnaround on inbound reviews.
• How it works: Send a PATCH request to the questionnaire endpoint with the fields you want to change. Pair it with the MCP server to drive updates from Claude, ChatGPT, or your own internal tools.
• How to get started: Existing API tokens with questionnaire write access already work. No new permissions to enable.
• Status: Live now
• Learn more: Questionnaire API reference

One little change can break big chunks of your knowledge. What if you could use the Conveyor API or AI to help make cleaning and curating your knowledge easier? Well now you can!

• What: Manage your Conveyor Knowledge Library through the Public API and / or AI via the MCP server. Create Q&A pairs, update verified status, and run bulk curation prompts from your AI client (Claude, OpenAI, Cursor, etc).
• Why it matters: When a fact changes (a renamed product, a deprecated cert, a new sub-processor), it can require a big clean up of your Q&As / knowledge. With this update you no longer need to hunt and edit each Q&A by hand. You can now point your favorite AI tool at your library and let it do the cleanup.
• How it works: Use POST /api/v2/knowledge_base/questions to create Q&A pairs directly, or connect Conveyor's MCP server in Claude Code or the OpenAI playground to run prompts like "unverify every answer that mentions TLS v1.1." The MCP server scopes to whatever your API key allows.
• How to get started: Generate an API key with the Knowledge Base Write capability, then follow the MCP setup guide.
• Status: Live now
• Learn more: Knowledge Base API and MCP setup guide

Your Trust Center is usually the first point of contact with a prospect's security team. If it's confusing, unprofessional, and generally mediocre, customer trust takes a hit. We want Conveyor customer trust centers to wow users, so we made some updates.

• What: A redesigned Trust Center landing experience that improves look and feel, makes it easier for visitors to self-serve answers, and adds proper mobile support.
• Why it matters: Most security reviews start with someone scanning your Trust Center. The new design gets visitors to the answers they need faster, on any device, without leaning on your team.
• How it works: The above-the-fold area surfaces what matters most up top. Visitors land in a layout that guides them toward documents, Q&As, and the Trust Center Agent. Everything is responsive so it works on mobile.
• How to get started: Live across all Trust Centers as of May 11. Preview is available from your admin panel.
• Status: Live now
• Learn more: What's new in Trust Center

Excel questionnaires rarely come in a clean two-column format. Now Conveyor handles the mess.

• What: Conveyor now maps and answers every response column in a spreadsheet, and automatically pulls in context columns like "Guidance Notes," "Priority," or "Details & Instructions" to inform ConveyorAI's answers.
• Why it matters: Five answer columns? No problem. Your reviewers also see the same context the AI saw, so they're not flying blind on what the question actually meant. Cleaner imports, sharper answers, less manual remapping.
• How it works: On import, Conveyor detects multiple response fields and context columns and maps them automatically. You can remap manually if needed. The AI uses any mapped context columns when drafting answers.
• How to get started: Upload an Excel questionnaire and check the mapping screen. It's on by default.
• Status: Live now
• Learn more: Flexible Response Fields and Question Context docs

A new endpoint built for AI agents to cross-check your KB at scale.

• What: A new PATCH /api/v2/knowledge_base/questions/:id endpoint lets you programmatically mark KB answers as verified or unverified. Status changes show up in the UI exactly as if a human had reviewed them.
• Why it matters: This unlocks AI agent verification workflows. An external agent can cross-check your KB answers against a separate source of truth, then flag or approve at scale. Your team focuses on the flagged items instead of reviewing everything by hand.
• How it works: Send a PATCH request with status=verified to approve or status=unverified to flag for human review. Also available as an MCP tool (knowledge_base_patch_question) for use in Claude Code and similar clients.
• How to get started: Grab your API key from admin settings and try it out.
• Status: Live now
• Learn more: API token docs

Three new endpoints unlock the question and conversation data your analytics team has been asking for.

• What: Three new GET endpoints. /api/v2/questionnaires/questions returns every questionnaire question with ConveyorAI answers, final answers, grading, status, last editor, and cited sources. /api/v2/oneoff_questions returns every question asked through the API, Slack, browser extension, web app, and Trust Center Agent, with answers, confidence levels, sources, and who asked. /api/v2/trust_center/conversations returns full Trust Center Agent conversations with nested messages, visitor emails, and referenced documents.
• Why it matters: Find Knowledge Base gaps by pulling low-confidence one-off questions. Audit Slack answer quality by source. Track Trust Center Agent conversations end to end. Build the dashboards you've been wanting in Looker, Snowflake, or wherever your analytics live.
• How it works: All three endpoints support filtering by date range and other parameters. Each is also available as an MCP tool (analytics_get_questionnaire_questions, analytics_get_oneoff_questions, analytics_get_trust_center_conversations) for customers using Claude Code or other MCP-compatible clients.
• How to get started: Grab your API key from admin settings and hit the endpoints.
• Status: Live now
• Learn more: API token docs

If you don't share Q&As, your Trust Center won't look like you're hiding a "0" anymore.

• What: The Knowledge Base section is automatically hidden on Trust Centers that have no Q&As shared publicly.
• Why it matters: Plenty of customers reserve their Trust Center for document sharing only. The old layout displayed a "0" next to Knowledge Base, which looked off. Now the section is removed entirely and Documents expand to full width.
• How it works: If you have zero Q&As shared to your Trust Center, the KB section disappears and the Documents area takes the full width. Add a Q&A and the section reappears.
• How to get started: Already live. Nothing to configure.
• Status: Live now

One less reason to open a support ticket.

• What: Admins can now customize Trust Center Agent text configurations directly in the Agent settings page, no Active Admin access required.
• Why it matters: Until now, changing your Agent's disclaimer text, file disclaimer, or questionnaire fallback messaging meant going through Active Admin. Now your team can own these settings directly and update them whenever your messaging needs to change.
• How it works: Go to your Trust Center Agent settings. You'll find editable fields for the disclaimer text shown below the chat, the file disclaimer at the top of downloaded files, and the fallback text for unanswered questionnaire questions. Edit, save, done.
• How to get started: Head to Trust Center Agent settings and look for the text customization options.
• Status: Live now
• Learn more: Enabling the Trust Center Agent

Curated Q&As now support the same test environment controls that documents already have.

• What: Curated Q&As now have test environment access settings, just like documents. Knowledge Admins can control whether a Q&A is available in live questionnaires, test questionnaires, or both.
• Why it matters: Previously, any curated Q&A you added was immediately used by ConveyorAI across all questionnaires. There was no way to validate new answers before they reached customers. Now you can add Q&As in test mode, run them through test questionnaires, and promote to live once you're confident.
• How it works: Open any Q&A in Knowledge Base, and you'll find a new "ConveyorAI Environments" section in its Settings. Toggle between Live, Test, or Both. New Q&As default to both environments enabled. Test Q&As are used in One-Off and Questionnaires running in test mode only. Slack and Trust Center Agent interactions are not affected by test-only Q&As.
• How to get started: Navigate to Knowledge Base, then Q&A. Click into any question and look for the "ConveyorAI Environments" section in Settings.
• Status: Live now
• Learn more: Curated Q&A documentation

AI assistants are only as good as the access and context they have. Now yours can tap directly into Conveyor.

• What: The Conveyor MCP server is available in Labs, letting you connect AI tools like Claude, Cursor, and other MCP-compatible clients directly to Conveyor.
• Why it matters: Your AI assistant can now do the actual work inside Conveyor, not just answer questions. Approve pending access requests, understand customer activity in your Trust Center, spot stale Knowledge Base content before it causes problems, track questionnaire completion rates, and see which documents customers keep downloading all from whatever tool you're already using.
  
  • Example queries might include:
    
    • Which customers are most engaged with our Trust Center this month?
    • What pending access requests do we have? Approve all requests from @acme.com and add them to the Enterprise access group.
    • Audit our Knowledge Base: which questions are most viewed, and which ones need attention?
    • Summarize questionnaire completion rates over the last 90 days?
    • Which documents are customers downloading most this month?

• How it works: Install the Conveyor MCP server and connect it to your preferred AI client. Once connected, your AI tools can query Conveyor and/or your Knowledge Library directly, pulling accurate, up-to-date answers into whatever workflow you're already in.
• How to get started: Follow the setup guide in the MCP Server docs to get connected.
• Status: In Labs (open Beta).
• Coming soon: We're adding Oauth and submitting to Claude and others for official listing.
• Learn more: MCP Server Setup Guide

Scaling your Customer Trust program means allowing more teams, like sales, the process security asks. Now you can manage exactly what team can do what inside of Conveor.

• What: Custom roles and permissions for Conveyor, giving Superadmins granular control over what each user can access and modify.
• Why it matters: As more teams touch Conveyor, from security analysts to sales reps, admins need to set guardrails. Custom roles make it safe to roll out Conveyor broadly without worrying about accidental changes to Knowledge, settings, or workflows.
• How it works: Superadmins go to Settings > Members & Access > Roles, then create custom roles with specific permissions. Assign roles to team members to control what they see and do across the platform. All users also get a refreshed Launchpad and cleaner Settings layout.
• How to get started: Navigate to Settings > Members & Access > Roles and click "Add Role."
• Status: Live now
• Learn more: Inviting and Managing Your Team

Most security review workflows start in a ticketing system. A case comes in through Salesforce, a ticket lands in Jira, a request pops up in Slack. Conveyor's intake agent already pulls questionnaires from those systems automatically. But once the questionnaire was done, you still had to download it and upload it back yourself. That gap in the workflow is gone.

• What: Completed questionnaires can now be returned directly to the ticketing system they originated from, including Salesforce, Jira, Slack, and Zendesk.
• Why it matters: This closes the last manual step in the intake-to-completion workflow. No more downloading from Conveyor and re-uploading to your ticket. The questionnaire goes back to where it started, with one click.
• How it works: When the intake agent is set up, the action button on a completed questionnaire changes to "Send to Salesforce" (or Jira, Slack, etc.). Click it, and the completed file is attached directly to the originating ticket. You can still download the original format or CSV from the overflow menu if you need it.
• How to get started: No setup required. The new export button appears automatically for customers using the Conveyor intake agent.
• Status: Live now
• Learn more: Exporting a Questionnaire

Spreadsheet questionnaires come in all shapes. Now Conveyor handles more of them automatically, and makes it easier to fix the ones it doesn't.

• What: Conveyor now auto-detects multi-column, multiple choice formats (Yes/No/N/A checkboxes) when you upload a spreadsheet questionnaire. We've also redesigned the manual mapping experience into a guided, step-by-step flow.
• Why it matters: Multi-column checkbox formats are one of the most common structures in enterprise security reviews, and until now they required manual reformatting before upload. That friction is gone. And when auto-mapping needs a correction, the new guided sidebar walks you through it in a few clicks instead of leaving you to figure it out on your own.
• How it works: Upload a spreadsheet questionnaire as usual. Conveyor reads the column headers and maps multi-column answer choices automatically. If a tab needs adjustment, click "Change Pattern Selection" to open the new step-by-step sidebar: select the question cell, choose the answer type, pick answer cells, and apply with Smart Select. One pattern applies across the entire sheet.
• How to get started: Upload any spreadsheet questionnaire in Conveyor. Multi-column detection works out of the box, no setup needed.
• Status: Live now
• Learn more: Importing from Spreadsheets

Conditional question support for spreadsheet questionnaires

How it benefits you

Questionnaires often include follow-up questions that only apply based on a prior answer. ConveyorAI now detects parent/child relationships in Excel questionnaires at import, answers parent questions first, and automatically cascades "N/A" to child questions when a section doesn't apply. Change a parent answer and all affected children regenerate automatically. Each child question includes a context card showing the dependency so you can jump to the parent in one click.

Get started

This is live now. Import an Excel questionnaire with conditional logic and ConveyorAI will handle the rest.

Your best security answers are shared  in Slack. Don't let them stay there 💬

• What: Turn Slack conversations into Trust Center Q&A pairs, just by reacting to a message with an emoji. 🧠
• Why it matters: New information surfaces fast in Slack. Now you can capture it in your Knowledge Library without leaving the conversation, keeping your Trust Center Agent accurate as things evolve.
• How it works: React to any Slack message with your chosen emoji. Conveyor’s AI parses the thread, drafts clean Q&A pairs for you, and flags any conflicts with existing answers in your library. Review, approve, and push to Conveyor — all without switching tools.
• How to get started: Set up the Slack integration in your Conveyor admin settings and choose your trigger emoji.
• Status: Live now
• Learn more: Adding Q&As via Slack

Not everyone leaves feedback in the moment — so now we follow up 📬

• What: Approved visitors now get a follow-up email 24 hours after their Trust Center visit — asking them to rate the experience, confirm if they completed their review, and leave comments.
• Why it matters: Not every visitor responds to an in-app prompt. This extends your feedback loop beyond the active session so you hear from more of the people who matter. You can also send yourself a test email to preview exactly what visitors will see.
• How it works: Enable post-visit email surveys in your Trust Center settings. Approved visitors automatically receive the email 24 hours after their visit.
• How to get started: Enable post-visit email surveys in your Trust Center admin settings.
• Status: Live now

The Trust Center Agent just got a speed upgrade 🏎️ — especially for document-heavy queries.

• What: Targeted speed optimizations for document retrieval queries.
• Why it matters: If your team manages a large document library (65+ docs), you may have noticed slower response times on queries like “give me all policies.” This was the most consistent feedback from Trust Center Agent users, so we tackled it first.
• How it works: We restructured how document data is passed to the agent, stripping out unnecessary overhead so it can retrieve what you need faster.
• Results: Faster on document retrieval queries, improvement in overall average response time. This is the first of several planned speed optimizations.
• Status: Live now
• Learn more: Trust Center Agent docs

Now you can collect feedback directly from visitors while they're still in your Trust Center 📋

• What: Trust Center visitors can now rate their experience and leave comments from within the Trust Center itself. No follow-up required. An optional email survey (sent 24 hours post-visit) is coming soon.
• Why it matters: You get a direct signal on whether visitors are finding what they need, surfacing friction in the review experience before it costs you a deal. Real feedback, not guesswork.
• How it works: Once enabled, visitors see a Feedback option in the Trust Center header. They can submit a 1–5 rating, optional comments, and an optional email for follow-up. Both authenticated and unauthenticated visitors can respond.
• How to get started: Head to Trust Center → Designer → Visitor Feedback and toggle on feedback collection. Each survey type can be enabled independently.
• Status: In-app survey live now. Post-visit email survey coming soon.

• Learn more: Documentation

‍

Your Trust Center Agent just got a whole lot smarter about what’s on your own page 🧠

• What: The Trust Center Agent now pulls answers from a broader set of your Trust Center content — not just Q&As and uploaded documents.
• Why it matters: Visitors were asking your agent questions about info that was right there on the page (like subprocessors or your company philosophy) and getting blank stares. That’s fixed.
• How it works: The agent now draws from What We Offer, Our Philosophy, Coming Soon, Announcements, Subprocessors, Quick Summary, and Badges/Certifications. This only affects the Trust Center Agent — no other answer sources are changed.
• How to get started: Nothing to configure. Your agent automatically picks up the expanded content.
• Status: Live now
• Learn more: Trust Center Agent docs

‍Conveyor’s Trust Center Agent is your own AI agent that can now automatically complete security questionnaires for your customers.

When a prospect or customer visits your Trust Center, they can upload their security questionnaire to the Agent directly in the chat window.

Your agent will then:

• Analyze the questions
• Generate answers from your Knowledge Library and documents
• Fill out the entire questionnaire instantly.

It does not require additional credits and is part of existing paid plans. Account admins can toggle their agent on and turn on the questionnaire feature in the Trust Center settings.

‍Benefit: Allows your customers and leads to upload security questionnaires themselves and get instant answers from your agent.

Fully supported portals now offer deep scan, which discovers conditional questions during import and ensures they are available before answering begins. After a deep scan import, users have all the answers they need for exporting back to the portal, rather than just the answers to questions that were visible on first glance. See the demo here.

‍Benefit: Eliminates back-and-forth rescanning for portals and ensures questionnaires are truly complete before export.

Aravo and HyperProof are now officially listed as fully supported portals.

‍Benefit: Conveyor will automatically detect and import all questions from Aravo and HyperProof portal questionnaires and automatically push reviewed answers back to the portal. Check out Conveyor’s latest list of supported portals here.

A new confirmation email can now be sent after a user signs an NDA. Existing admins can enable this setting and preview emails in Designer.

‍Benefit: End users get easy access to their signed NDAs, while admins have full control over their experience.

The default Q&A prompt no longer forces word for word reuse of context, resulting in more concise answers. Users who prefer verbatim wording can still enable it via custom prompts.

‍Benefit: Shorter, clearer answers by default, with flexibility for users who prefer exact wording.

The ConveyorAI Trust Center Agent is a fully embedded, AI-powered assistant that helps your customers and prospects instantly find the security answers and documents they need. Your Trust Center now includes an AI agent that can:

• Instantly answer due diligence questions (with citations)
• Fetch and bundle documents on demand
• Answer questionnaires uploaded by customers (coming in January)
• Understands/answers in your customers' languages (coming soon)
• All with built-in security by design

Conveyor now highlights any answers that already exist in uploaded questionnaires and users can choose whether to overwrite those answers with AI or keep imported answers.

‍Benefit: Prevents accidental blocking of AI answer generation and saves users from having to reimport overwrite-enabled spreadsheets.

Expanded the HubSpot integration to help automate customer access to Trust Centers:

• Admins can now bypass NDAs for recognized visitors
• Domain-based auto-approval for Trust Center access requests
• Exact email address matching within HubSpot

‍Benefit: Reduces manual access approvals for known domains.

Trust Center Featured Documents UX Improvement
The Trust Center admin view for featured documents has been redesigned from a carousel to a list/grid layout with page separators, allowing admins to drag and drop documents across multiple pages.

‍Benefit: Easier reordering of featured documents, especially for Trust Centers with many items.‍

‍

SSO / SCIM Deactivation Bugfix
Fixed an issue where some SCIM-managed users were not correctly deactivated in Conveyor after being marked inactive in Okta (and other IdPs). Conveyor also added more logging to inbound SCIM requests to improve debugging and provide more detail when customers report issues.

‍Benefit: More reliable user lifecycle management and faster resolution of SSO/SCIM-related issues.

Introduced Core Questions in the Knowledge Base Explorer to identify high impact, commonly asked customer questions that are asked in a variety of ways and standardized them to provide clear, consistent responses.

‍Benefit: Takes commonly asked questions that are asked in different ways and answers them concisely and consistently.

‍

Conveyor’s AI now runs analysis on your Knowledge Base. The new Knowledge Base Explorer lets users:

• Run benchmark analyses
  
• Identify gaps in their KB
  
• Create or delegate Curated Q&As
• Rerun benchmarks to track progress

‍Benefit: Users get actionable insights into knowledge coverage and performance over time from Conveyor AI.

Browser Extension Vertical Resizing
The browser extension can now be resized vertically from either the top or bottom edge.

‍Benefit: Prevents the extension from covering key portal UI elements (especially OneTrust).‍

‍

Update to “Minimal” Verbosity Setting
When users choose “minimal” style, Conveyor answers questions more concisely:

• “Yes” answers are now a single word.
• “No” answers may include one short clarifying sentence.

‍Benefit: Cleaner, more consistent responses requiring less editing.

‍

Link Directly to PDFs for Knowledge Base

Instead of having to upload PDFs to Knowledge Base, users can now provide a URL to their document.

Benefit: Reduces the steps required to add a PDF to a Knowledge Base.

Conveyor’s AI Librarian now highlights conflicting pieces of information in your Knowledge Base, including both documents and external sources like websites, Confluence, Notion and Google Drive. Users will now see a clear indication of the discrepancy in the comments tab and at the top of the user interface.

‍Benefit: Improves transparency and speeds up conflict resolution during answer review.

The extension now supports Panorays and Archer Legacy portals.‍

Benefit: Conveyor now provides full end-to-end automation support for Panorays and partial support for Archer Legacy.

Admins can now control export permissions for configurable users. By default, exports remain disabled. A redesigned Add Questionnaire modal also launched with a simplified flow for faster uploads of questionnaires for AI to answer.

Benefit: More granular role management

The AI Librarian now uses an improved model and refined prompts to produce shorter, more accurate lists of conflicts only (no duplicates).

‍Benefit: Easier prioritization and improved accuracy in KB conflict detection.

Fully automated answering for Whistic portal is now live. Users can drop a portal URL into Conveyor, and it automatically handles importing all questions, generating answers, and putting reviewed answers back into the portal for you. ‍

Benefit: Conveyor's AI will read and answer questionnaires automatically and place answers back into the Whistic portal for you.

Edit Product Lines on Questionnaires

Users can now edit product lines directly on questionnaires without re-uploading or contacting support.‍

Benefit: Greater flexibility and reduced dependency on support for corrections.

‍
Bulk Answer Regeneration

Users can regenerate AI answers for multiple questions in a questionnaire at once.‍

Benefit: Speeds up large revisions and ensures consistency across answers.

Full import/export automation for GraphiteConnect portal questionnaires is now supported.‍

Benefit: Expands portal coverage and eliminates manual work for security questionnaires in GraphiteConnect portal.

AI agent can now respond to natural language commands in triage queues (e.g., “why was this rejected?” or “process this file”).

Benefit: More autonomous, flexible, and helpful in reducing triage dead ends.

Share Document Folders

Users can share links to entire document folders, with visibility determined by individual file permissions.
Benefit: Easier distribution of grouped compliance documents while maintaining secure access.

‍

Salesforce Permissions Check Modal

A new modal checks whether the connected Salesforce OAuth user has the correct permissions for selected features.
Benefit: Prevents misconfigurations and saves time debugging.

‍

One-Off Questions API Update

The One-Off Questions API now supports optional email addresses for better analytics. Requests without an email default to the API key owner.
Benefit: Improves visibility into who is asking questions and adds API usage insights.

‍Archive Product Lines

Admins can archive product lines without removing associated records. Archived lines are visually distinct and excluded from dropdowns.
Benefit: Simplifies product line management while preserving historical associations.

‍

Bulk Delete Questionnaires

Multiple questionnaires can now be selected and deleted at once from the “All” tab.
Benefit: Saves time during cleanup or general housekeeping.

Users can now import Notion pages into their knowledge base (beta tested with Remote and Notion).
Benefit: Expands knowledge base sources and reduces manual content maintenance.

Images and Rich Text in Answers

A rich text editor with support for images and formatting
Benefit: Improves clarity and professionalism of AI-generated answers, especially in exports.

‍

Performance Improvements

Exact Wording in ConveyorAI Answers

‍ConveyorAI now uses exact wording from sources by default (previously only available via custom prompts).
Benefit: Reduces review time and ensures responses align closely with source material.

‍

Faster questionnaire loading

Massive performance improvements: large questionnaires now load in ~2 seconds (down from over a minute).
Benefit: Smoother, faster workflows and improved customer experience, even on very large RFPs.

‍

Smarter Past Answer Imports

Placeholder answers (e.g., “yes/no/na” or “[vendor to elaborate]”) are now ignored during imports.
Benefit: Cleaner imports and more opportunities for ConveyorAI to provide real answers.

‍

Adjust Past Answer Imports

Users can now edit AI-imported past answers directly from import history (spreadsheet imports only).
Benefit: Greater control and flexibility when refining imported questionnaire data.

‍

Domain Tooltip in Add Questionnaire

A tooltip now explains why providing a domain is required when creating a questionnaire.
Benefit: Reduces confusion and encourages full platform use without wasted credits.

Before, you could only customize prompts at the org level.

Now, you can now fully customize how ConveyorAI answers on a per questionnaire basis!

Perfect for fine-tuning answers when different customers require slightly different tones or levels of detail.

‍

Here's how it works:

• Set a custom verbosity and free response prompt for any individual security questionnaire or RFP and the custom settings will apply ONLY for that project.
  
  • When editing, the form is pre-populated with your organization's default custom style (if one is set), making adjustments quick and easy.

• Once applied, custom styles are displayed in the questionnaire's properties section for easy reference.
• Available everywhere: This feature works in both the Conveyor web app and the browser extension.

‍

More customization has arrived to your Conveyor Trust Center!

‍

You can now update the thumbnail that appears when sharing a Trust Center. This is configurable in the brand settings (same location as logo and cover image)

Highlight your competitive differentiation in security through a video embedded into your Trust Center.

It was previously only possible to make video resources available to authorized Trust Center visitors. Now you make them fully public to any visitor!

Usability Improvements:

ConveyorAI

• Users can now select individual Confluence pages to sync, when adding to their knowledge base.
• Improved pronoun usages in ConveyorAI Answers - e.g. users should see fewer instances of addressing users in an answer as "we/us" or the vendor as "you".
• Past answers can now be filtered by those that have been upgraded to a Q&A pair

Collaboration

• The user page in Account UI now has a search box
• 6 month and 1 year options for 'Default Approval Configuration' now available
• Improvement to Comments:
  
  • Comments are now editable by the user who created the comment
  • Comments are also ordered in ascending chronological order, with the latest comment on the bottom.

For the following portals, you can now auto-import questions, get answers, and fully export back to the portal -- all using only the questionnaire URL.

• UpGuard
• Venminder
• Bitsight
• LogicGate

This integration works exactly the same as our Jira, Zendesk, Salesforce, etc integrations for Conveyor's AI Agent, Sue.

Sue comes to the places where our users are already triaging and managing their questionnaire workflow and with this integration, Sue watches shared inboxes for questionnaires, automatically ingesting, assessing, and completing on security questionnaires.

Upload docs, flag dealbreakers, set win themes, assign teammates, and track every moving piece — all in one place.

ConveyorAI even weaves your win themes into generated responses to help you stand out and submit with confidence.

Now you can keep track of security questionnaires and RFPs separately.

We’ve made big performance improvements for questionnaire answering and one-off questions (in web app, Slack, browser extension)  — so you should see snappier responses and smoother experiences across the board.

You should experience:

1. Higher answer rate
2. Improved accuracy
3. Faster runtime

Directly upload questionnaire requests in Slack for processing by Conveyor's AI Agent, Sue. Sue will triage the request and communicate with requestors in thread.

Import/export questionnaire questions using Conveyor's browser extension with 1 click: now supported for Prevalent questionnaires.

See everything Sue (AI Agent) been triaging behind the scenes: review individual requests, see exactly how Sue classified them, and view any questionnaires created.

You can also approve escalations, handle errors, and stay fully in control — all from one place.

‍

One click to import all questions with the browser extension and one-click to export & fill all answers back into the portal.

"Fully automated one-click" experience is supported for: ServiceNow, OneTrust, BitSight, ProcessUnity, Coupa, LogicGate, Venminder, UpGuard.

Preview Word Docs and PDFs: Just like PDFs, DOC and DOCX questionnaires can now be previewed directly from within Conveyor’s UI for faster reference.

Faster Questionnaire Answering Throughput: Back-end improvements allow ConveyorAI to process more questionnaire jobs in parallel, speeding up full questionnaire completion without affecting single-question performance.

Configurable Access Groups for Auto-Approvals: Auto-approved users can now be automatically added to a specific access group for better control

No More Caching One-off Searches: Repeating a one-off search will now generate a new answer each time, preventing outdated or stale results when source data changes.

Your Conveyor trust center will now auto-translate to French, Spanish, Japanese, and more—automatically adapting to where your customer is located. All you have to do is toggle translation on for the languages you need.

You can now import answers from DOCX/PDF files and provide contextual guidance for complex questions—making AI responses faster and smarter.

Assign questions to users or teams, manage watchers separately, and use new filters and bulk actions to streamline collaboration across Conveyor.

Usability improvement: easier management for admins and internal users to quickly see key document info at a glance.

Questionnaires imported from spreadsheets, PDFs, or DOCX files now retain their original sections, making large questionnaires easier to navigate and work with.

• Sections are automatically detected on import
• You can perform bulk actions by section
• RFP templates now nest questions under sections

Coming soon: section editing, detect sections in portal-based questionnaires.

Custom prompt answers now use the same cited sources as the original AI responses, reducing hallucinations and grounding answers in your knowledge base. This applies to both org-wide and per-question prompts.

Sue, your AI Agent, can now pull in data from multiple sources during triage! If you're using tools like Jira Service Desk or Zendesk, just drop a Salesforce Opportunity link in the ticket—Sue will automatically pull relevant deal info and combine it with the ticket context to make better decisions.

You can now view the original questionnaire PDF directly in the “Overview” tab when working on a questionnaire—making it easier to review AI-generated answers with full context, especially when the file was uploaded by someone else.

Access requests from users for NDA-gated content in your trust center will automatically be approved if there is a matching record in Hubspot.

Don't want to set the AI agent loose without knowing how it will perform? No sweat. You can now test your AI Agent’s behavior in a sleek, conversational interface that reflects Sue’s growing agentic abilities.

What’s new:

• Pre-filled test fields so you’re not starting from scratch every time
• Live rule tracking — watch your guardrails get checked off as Sue processes them
• Natural, descriptive responses — no more robotic outputs
• Chat back and upload files to iterate with Sue in real time

Whether you're validating your guardrails or just want to see how your AI Agent thinks, the new tester makes it faster, easier, and way more delightful.

Conveyor's AInow better understands “rollup questions”—those short follow-ups that only make sense in the context of a previous question.

For example:

User: Hi ConveyorAI! Answer this: “Web Hosting” or "Continuous"
ConveyorAI: 🤔 That doesn't make sense… unless we know the parent question is:
“What service hosting models are provided as part of this service? Select from the list below.”

Now, Conveyor's AI classifies and connects these rollup questions to their parent context, so the AI gives more accurate answers, even when the question alone isn’t clear.

What this means for you:

• Better understanding of short or vague follow-ups
• Improved accuracy on complex questionnaires
• Already seeing a +10% boost in precision and a +9% boost in recall for rollup questions

You can now control Sue using natural language directly in ticket comments! Just mention “Sue” or “Conveyor” in your comment and she’ll know you’re talking to her.

Here’s what you can do:

Start/Stop Listening:

‍
Tired of Sue chiming in mid-discussion? Ask her to stop listening—and bring her back in when you're ready. Here are some example commands:

• “Sue, stop listening.”
• “This is out of scope, Conveyor.”
• “Sue, triage this one please.”
• “Start listening again, Sue.”

Override Custom Rules:

‍
Need to bypass ARR, NDA, or question count rules? You can now ask Sue to override them directly. Here are some example commands:

• “Sue, override the number of questions and finish the ticket.”
• “The NDA is approved. Conveyor, go ahead.”
• “Sue, just create the questionnaire.”

Bulk select questions to update or delete multiple questions at once.

Select your questions, then:

1. Delete multiple questions at once
2. Apply custom prompts and AI adjustments to a group of questions at the same time

Announcing Conveyor's newest Jira Service Management (JSM) integration! If you use JSM as your case or ticketing system, you can now integrate seamlessly with Conveyor's AI agent to triage, manage cases, collaborate with your team, and upload questionnaires to Conveyor. Installation is quick and easy, via a new Jira Forge App (currently a private link, marketplace listing TBD).

Customers who use Zendesk’s ticketing system to manage requests from sales and other customer-facing teams can now send Conveyor's AI agent straight to their Zendesk queue to do triaging, communications, and questionnaire importing into Conveyor.

Sue, Conveyor's AI Agent for Customer Trust, not only helps teams share their SOC 2 and answer security questionnaires in seconds, she now takes on every manual step of the customer security review.

‍

Sue can make judgment calls and act independently to:

• Manage and complete seller requests while communicating across systems (wrong questionnaire file uploaded to a ticket? Sue will collaborate to get the right one, process it, and keep everyone in the loop)
• Automate questionnaire completion, rejection, or escalation based on your policies.

• Process and answer any questionnaire format—yes, even portals—with no manual tagging required.

• Auto-notify and communicate with teams to get requests reviewed and completed
• Curate a knowledge graph without a giant knowledge base
• Personalize a Trust Center based on customer insights from conversation intelligence tools

Sign up for early access here.

‍

No more tagging cells for the AI to read for Excel questionnaires.

Simply upload and the AI will automatically read any format of excel questionnaire and place answers into the sheet in front of your eyes (instead of just at export). The same experience for Word and PDFs is now available for Excel.

‍

Use Microsoft Teams to get notifications right where you work. After connecting to Teams, you will be able to:

1. Approve requests from customers to access your trust center
2. Tag SMEs in the Conveyor app so they are notified in Teams for easy collaboration
3. Get notified in Teams on questionnaire status changes and updates

‍

You can now work on foreign language questionnaires in English!

Set your review language to English and you can make edits or answer unanswered questions in English, and ConveyorAI will automatically translate and update the answers in the original language before exporting.

‍

Improvements to UX make it easier to see who’s assigned to each question, if there are comments, and where the mapped answers are in the original file.

When ConveyorAI generates an unanswered or flagged question, it will think through whether there is a highly relevant party to ask for help.

If yes, it will present these suggestions in a "Suggestions" slide out panel on the Questionnaire detail page. Users can reject or accept suggestions in order to assign the question to a collaboration group.

‍

Conveyor now understands when a question is asking about internal policy-level scope (like a company policy for Multi-Factor-Authentication used by employees) versus an external product-level scope (like delivering Single-Sign-On as a capability for customers). This means ConveyorAI will answer questions with higher accuracy and the user spends less time on review.

A suite of new options for in-the-moment answer AI Adjustments, including changing the verbosity level, adjusting the tone of an answer, or adding a custom prompt ("If the question mentions privacy, add this link to our privacy policy!")

Plus, apply your preferred settings universally so the AI knows exactly how you like your answers, every time. The ability to set global custom prompt and verbosity settings can now be found on the Organization Preferences page.

All you have to do is add in OneTrust URL and the AI will import questions, answer them, and export each answer back into the right place in the portal for you. (More portals coming soon!)

‍

Refer to the docs site for more on how to use the URL import and how to use the auto-export back into the portal.

‍

Trust Center Trust Center

List View for Subprocessors in Trust Center

Toggle between the "Tile View" and the "List View" when you showcase your subprocessors on your trust center.

‍

Add Custom "Privacy Policy" and "Terms of Service" Links for Trust Center.

Now you can link to your own privacy policy on the "Request Access" modal instead of using Conveyor's.

‍

Conveyor will auto-parse SIG & HECVAT questionnaires Questionnaire Automation

Upload SIG/SIG Lite and HECVAT/HECVAT Lite questionnaires as "Documents" and Conveyor will auto-parse and source questions/answers from these standard formats (same way it works for the CAIQ!)

Need help with a questionnaire? Tag a subject matter expert (SME) and they will get a link to everything they're tagged in - all in one spot. SMEs have one place to go for all the questionnaires that need their help, and they can get there easily from slack so you get answers fast.

‍

ConveyorAI will automatically reflect on “best practices” for security questionnaire answers so you get better answers with fewer edits needed.

‍

Stop maintaining thousands of Q&A pairs and link Conveyor to your Confluence instance to pull the most updated data to use in questionnaire  and RFP answers.

‍

Questionnaire Automation Questionnaire Automation

ConveyorAI Adjusts Answer Verbosity and Tone for RFPs vs. Security Questionnaires‍

ConveyorAI will default to an answer style that is more verbose and detailed when you pick “RFP” as the file type.

‍

‍AI Librarian Proactively Flags Duplicates in Knowledge Base‍

ConveyorAI will proactively notify you when it finds outdated or conflicting answers across past answers or Q&A pairs.

‍

‍Browser Extension Update: Auto-Import All Questions with 1-Click

Simply plug in your URL when you create a new questionnaire in app and ConveyorAI will scan and pull in questions from OneTrust portals for easy question answering, editing, and review.

‍

‍New Analytics Dashboards General

Improved analytics dashboards means you can see:

• Which team members are self-serving answers to questions internally (NEW)
• Precision and accuracy rates of ConveyorAI in your instance (NEW)
• Questionnaires created and completed
• Avg number of days it takes to complete questionnaires
• Avg number of questions
• Security influenced revenue
• Frequently used FAQs and documents
• Trust center customer activity

‍

Have a customer question? Just type “@Conveyor” and ask anything. Plus, safety thresholds determine if questions are escalated to an approver before being returned to your sales reps.

‍