One AI Engine for Security Reviews
and RFPs

Learn how Conveyor fully automates your entire Customer Trust workflow from hosting a Trust Center to generating instant answers for security questionnaires and RFPs.

There are three metrics Conveyor tracks: answer rate, precision rate, and perfect answer rate so you can monitor how the AI performs over time.

‍

Answer rate is the percentage of questions ConveyorAI returned an answer for (versus flagging as unanswered). Precision rate measures how often answers were used without edits. Perfect answer rate is how often an answer was accepted exactly as-is.

We do not allow training of third-party LLMs using Customer Content. Third party LLM providers are contractually prohibited from using Customer Content for their own purposes.

We only use anonymized and aggregated data when it comes to improving our ConveyorAI algorithms.

Yes, and it's one of the most underused settings in the platform. A custom prompt lets you give ConveyorAI standing instructions — things like your preferred tone, response format, what to do when a question falls outside your scope, or how to handle sensitive topics. Customers who find the custom prompt report it as a significant improvement to initial answer quality.

One important thing to know: the custom prompt runs after source retrieval, not before. That means it can shape how an answer is written, but it can't influence which knowledge sources get pulled. For source-level control, focus on your Knowledge Library structure — product line tagging, document freshness, and curated Q&A pairs have the biggest impact there.

Not directly today, but this is one of the most-requested improvements we're actively working on. Right now, ConveyorAI draws from all eligible knowledge sources — curated Q&A pairs, documents, and past answers — based on relevance. Some customers prefer to weight curated Q&As heavily; others rely more on recent past answers. We've heard this feedback clearly from customers and it's shaping what we build next.

In the meantime, the best way to guide answer quality is to make sure your highest-priority content is well-curated, deduplicated, and tagged to the right product line. The cleaner the signal, the better the output.

This is a known limitation. When your team edits an answer in a questionnaire, that correction gets saved as a past answer — but the benefit only carries forward if a future question is similar enough to pull that specific result. The work your team puts into polishing each response doesn't automatically retrain the AI or raise the quality ceiling across the board.

The most reliable fix is to promote your best edits into curated Q&A pairs in the Knowledge Library. Curated pairs get higher weight and apply more broadly than one-off past answers. Think of it as converting good work into reusable signal rather than letting it sit in a single questionnaire.

The fastest fix is editing the answer directly in the questionnaire. Conveyor saves that correction as a past answer and reuses it for future questions. For patterns of inaccuracy, check whether you have conflicting or outdated sources. The most common culprit is uploading multiple versions of the same document or content that hasn't been refreshed.

Conveyor can handle any format of security questionnaire like excel, word docs, PDFs, and online portals.

For file formats, simply drag and drop into Conveyor or drop in one of your integrated tools like Slack, Salesforce, and others for Conveyor to pick up and process.

For portal security questionnaires, Conveyor has a browser extension that will read the questions, answer and auto-populate the answer back in -- all with one click autocomplete. Conveyor can support any portal with the browser extension; the level of automation it supports changes based on the extension. Learn more here.

Alternatively, you can also paste the URL in Conveyor and Conveyor AI will access the portal questionnaire, bring in the questions and answer them in the Conveyor app.

‍