# How it works | ReqFit Inside the AI pipeline that reads your RFP and proposal together and tells you what is missing. Built on Google Cloud Vertex AI for bid and proposal teams. URL: https://reqfit.com/how-it-works --- How it works # How ReqFit works Inside the AI pipeline that reads your RFP and proposal together and tells you what is missing. Built on Google Cloud Vertex AI for bid and proposal teams. The journey ## Three steps. One report. 01 ### Upload Drop in the RFP and your proposal, or a questionnaire and your responses. Word, PowerPoint, PDF, or a mix of all three. No setup. No prompt engineering. It is drag and drop simple. 02 ### AI analyses The 5 stage AI pipeline reads your documents, extracts the requirements, maps your responses, and scores how well each requirement is covered. 03 ### Get your report View online or download a structured PDF or Word report you can act on, share with your team, or attach to an internal sign off. One click. Three steps. Read on for what is happening underneath. See it in action ## See it in action Here is ReqFit running for real. Below are three screenshots from the product, plus a short walkthrough showing the full flow from upload to finished report. ### Upload your documents Drop in the RFP and your proposal. Word, PowerPoint or PDF formats. ![ReqFit upload screen with an RFP and a proposal file selected. Full size view.](/images/screenshots/screenshot-01-upload.png) ### Watch the AI pipeline run The pipeline runs end to end without your input. ![ReqFit processing screen showing the AI pipeline running. Full size view.](/images/screenshots/screenshot-02-processing.png) ### Download your AI checked report View online, download as Word, or download as PDF. Your choice. ![ReqFit completed review with options to view the report online, download as PDF, or download as Word. Full size view.](/images/screenshots/screenshot-03-report-viewer.png) Click any screenshot to view it at full size. ### Watch a full review, end to end That is the whole product. No prompt engineering, no setup, no manual step. Read on for what the AI is actually doing under the hood. The pipeline ## Inside the AI pipeline ReqFit runs on Google Cloud Vertex AI. It is not a chatbot you have to coach. It is a multi-stage pipeline purpose-built for one job: comparing a proposal against the requirements in an RFP and producing a structured review you can use. Here is what happens between upload and report. 01 ### Document ingestion All documents are parsed into structured text. Section structure, tables, headings, and embedded text from PDFs are preserved. Word, PowerPoint and PDF formats are supported. 02 ### Requirement extraction The AI reads the RFP or questionnaire and identifies the requirements, evaluation criteria, and mandatory questions it finds. Each one is given an ID and tagged with its source location and priority. 03 ### Coverage analysis Your proposal is cross-referenced against the extracted requirements. The AI maps responses to specific requirements, identifying where coverage is direct, where it is partial, and where it is missing. 04 ### Scoring and gap identification Each requirement receives a coverage rating: Strong, Adequate, Weak, or Missing. Gaps are tagged by priority and the pipeline produces an overall coverage score plus a section by section breakdown. 05 ### Report generation The full analysis is rendered into a structured report: a summary with the headline coverage score, submission readiness, priority actions, compliance confirmations, and a per-category breakdown. The whole pipeline runs end to end without your input. ## Also works for due diligence questionnaires A due diligence questionnaire is a structured set of questions, which is the same shape of problem the pipeline is built for. Upload the questionnaire and your completed responses, plus any attachments your answers refer to. ReqFit treats each question the way it treats a requirement, working through them one by one and showing you which your responses cover and which still need attention before you return the document. It checks how well your responses cover what is asked. It does not judge whether your answers will satisfy the assessor. Cloud platform ## Built on Google Cloud Vertex AI ReqFit is built on Google Cloud Platform's Vertex AI. We use a multi-model approach. Different models are routed to different stages of the pipeline based on the task. Document parsing, requirement extraction, coverage reasoning, and report generation each have a dedicated model configuration tuned for the job. This matters for three reasons. ### Capability A general purpose AI tool cannot reliably produce a structured proposal review. It can give you opinions, suggest rewording, and answer questions about your bid. It will not work through the requirements in a forty page RFP, map them to your responses, and score coverage in a structured way. The ReqFit pipeline is engineered for that specific task. ### Reliability Built on managed, production-grade Google Cloud infrastructure, so the service behaves predictably and is there when a deadline is. Each review runs against the same extracted requirements baseline, so a re-run measures your new draft against a fixed standard. No prompt engineering required from you. ### Security Vertex AI processes your documents inside Google Cloud. Source documents are discarded once your report is generated; the extracted requirements list is retained against your account and cleared when you delete your account, and reports stay in your dashboard for 90 days. We do not log document content, and we never train our AI on your documents. We are Cyber Essentials certified. Full detail on our [security page](/security). Platform Google Cloud Vertex AI Multi-model architecture. Different stages of the pipeline route to different Vertex AI models, each tuned for the task. Source documents are discarded once your report is generated, and we do not log document content. Purpose built ## Not a chatbot. Not a writing tool. A purpose built AI reviewer. ReqFit is not a general purpose AI assistant. It does not write your proposal, suggest replacement copy, or answer questions about your bid. It does one thing. It tells you, with evidence, where your proposal does and does not match the brief. The final check on accuracy and completeness before you submit stays with you. ### Structured output Every report follows the same format: the same sections, the same scoring scale, the same evidence trail. You can read the second one as fast as the first because you already know where to look. ### Domain reasoning The pipeline understands what an RFP or questionnaire is. It knows the difference between an evaluation criterion and a context paragraph, between a hard requirement and a nice to have, between a question that needs a specific answer and one that needs a narrative. ### Grounded in your documents ReqFit works from the source text of your RFP and proposal, not from generic assumptions. Each review measures your draft against the same extracted requirements baseline, so a re-run is judged against a fixed standard. The report ## The report Every ReqFit report follows the same structure, so you and your team always know where to look. 1. 01 Cover and about this report 2. 02 At a glance 3. 03 Submission readiness 4. 04 Priority actions 5. 05 Compliance confirmations 6. 06 Detailed breakdown 7. 07 Next steps ![Preview of a ReqFit review report showing the cover, coverage scorecard, and requirement by requirement analysis pages.](/images/pipeline/report-preview.png) The report is viewable online and can be downloaded in both Word (.docx) and PDF format. Edit it, share it, or attach it to an internal sign off. It is yours. Privacy ## Your data, your control Source documents are discarded once your report is generated. The extracted requirements list is retained against your account so re-runs measure against the same baseline, and is deleted when you delete your account. Reports stay in your dashboard for 90 days. We do not log document content, and we never train our AI on your documents. Full detail on data handling, the Vertex AI processing region, AI regulation alignment, and your rights is in [How we use AI](/how-we-use-ai). ### What that means in practice - **Tiered retention.** Source documents discarded after processing. Requirements list retained until account deletion; reports retained 90 days. See our [security page](/security) for the full data lifecycle. - **Sole sub-processor.** Document processing happens within Google Cloud Vertex AI. No other third party touches your document content. - **Encrypted in transit.** All uploads and processing use TLS 1.3. - **No human review.** ReqFit staff do not access customer documents at any time. The pipeline runs without manual intervention. - **No model training.** Your documents are never used to train, fine-tune, or improve any AI model. - **UK GDPR aligned.** Our standard [Data Processing Agreement](/dpa) is live; a bilateral signed copy is available on request to [security@reqfit.com](mailto:security@reqfit.com). [View our security model](/security) ### Compliance at a glance Sub-processor Google Cloud Vertex AI Sole sub-processor for document processing. Encryption TLS 1.3 in transit AES-256 within Google Cloud infrastructure. Retention Tiered Source documents purged from runtime memory after report generation. Requirements list retained until account deletion. Reports retained 90 days. Regulatory UK GDPR aligned ICO registration ZC111039. Our standard [Data Processing Agreement](/dpa) is live; a bilateral signed copy is available on request to [security@reqfit.com](mailto:security@reqfit.com). For procurement and security review questions, contact [security@reqfit.com](mailto:security@reqfit.com). ## Get your one free report No credit card. No commitment. One free report on every account. [Get started free](https://review.reqfit.com/register)