Yearly running cost Assuming 240 questionnaires a year
By hand $90K / year 2,000 hours of work
Subscription $10.8K+ / year
This workflow $864 – 1.7K / year Machine usage only; setup, hosting and review are extra. Company sources Where should the answers come from? Use approved product documents, security policies and past answers. Get cited drafts and a list of questions your team needs to resolve.
Upload approved documents Upload your sources and export answers for review $864 – 2,880 setup Connect my document library Sync approved sources and connect your review queue $1,440 – 4,800 setup
Which system should it connect to?SharePoint document library Existing review queue Another system (specify below) Your team approves the answers. Nothing is submitted to your customer automatically.
Files and connections Guide prices assume your provider adapts a tested workflow using AI-assisted development, and you supply approved sources and test answers. Complex permissions or custom connections need a separate quote.
Uses approved uploaded sources and returns a workbook of cited drafts for review. No connection to your systems.
The estimate covers English XLSX or CSV questionnaires of up to 100 questions, using approved text or PDF sources. Scanned questionnaires, tender portals and complex document layouts need a separate scope.
Answer accuracy How thoroughly should answers be checked? Choose targets for supported facts and complete answers. Missing evidence should be flagged, not guessed.
Source-checked drafts Cited answers with missing or conflicting information flagged. ≥98% source-supported facts · ≥90% complete answers Included Higher verification targets More demanding checks on factual support and answer completeness. ≥99.5% source-supported facts · ≥95% complete answers Setup +$432 – 1,440 My own requirements Describe your acceptance targets in the notes below. Quote separately
How accuracy is measured These are acceptance targets to test on your questionnaires, not measured results. Complete answers are counted only where your approved sources contain the information needed.
Targets for your selected checks What is checked Target Facts supported by sourcesShare of verifiable factual claims in the drafts that a valid, authorised source supports. ≥98% Citations that hold upShare of factual claims with a correct location, version and supporting evidence. Claims without a citation count against it. ≥98% Answerable questions fully coveredShare of answerable questions whose draft covers every required sub-question without a material error, as marked by your experts. ≥90% Accepted with light editsShare of all questions the expert accepts with light edits and no new facts. Unknown and conflicting questions stay in the denominator. ≥70%
Test on questionnaires not used to build the workflow. Check each factual claim against its source. Unknown answers, expired documents and conflicting information go to your team; your team reviews every answer during the pilot.
Speed How soon do you need the draft answers? Choose a target for a questionnaire of up to 100 questions. Your team's review time is separate.
Within 15 minutes Included Within 10 minutes Setup +$144 – 480 Within 5 minutes Setup +$288 – 960
Timing details Measured after your sources are indexed, including queueing, drafting, checks and workbook export. The target applies to at least 95% of test questionnaires, with 1 in progress at a time. Your provider confirms what is achievable on your hardware.
Running-cost target How much do you want to spend per questionnaire? Select how much your AI workflow can spend drafting answers to up to 100 questions. This may affect how many checks and revisions it can run.
Up to $17.00 Included Up to $10.00 Setup +$144 – 480 Up to $5.00 Setup +$288 – 960
Cost details Reference machine cost: $6.60 – 16.20 per questionnaire, at 20 questionnaires a month after indexing. Includes model calls and allocated hosting, not expert review or maintenance.
Your selection is a target, not a guaranteed charge. Lower costs and faster answers can require more setup work; the provider tests the combination against your accuracy requirements.
Your environment Where do you want it to run? Choose where your documents are processed and whether an external AI model can read them.
Run it onMy cloud account My own computer or server AI model accessUse an external AI model Keep all processing inside my environment
Data and access details Documents, indexes and review records stay in your account. Whether any data leaves is decided by the model access option, not by where it runs.
Authorised questions and retrieved passages go to the agreed model provider. The guide running cost assumes this option.
Answers and citations must follow the same document permissions. Fully offline processing cannot sync an external library; use uploaded documents or agree an internal connection separately.
Interface How do you want to use it? Use your existing AI agent, a dedicated agent or a web page. You can choose more than one.
My existing AI agent Use MCP to send a questionnaire and retrieve the draft answers. Setup +$144 – 480 A dedicated AI agent Chat with an agent to upload questions and review its answers. Setup +$288 – 960 A web page Upload files, check sources and download the completed workbook. Included
Estimated setup price $864 – 2,880
This guide price is a starting point for your conversation with a provider. Agree the final scope and price directly with your provider.
Your requirements
Sources Upload approved documents
Answer checks Source-checked drafts
Speed target 15 minutes / questionnaire
Cost target $17.00 / questionnaire
Environment My cloud account
Interface A web page Per questionnaire of up to 100 questions.
Prepare my brief → Generate a draft prompt to share with your connected RenX agent.
Enable JavaScript to configure a brief, or prepare your task in RenX .