An RFP arrives Monday morning.
You and ProjEx build a defensible draft proposal and RFP compliance gap analysis within a few hours.

ProjEx Propose reads the client's RFP, extracts every requirement, drafts a compliance-bound outline, and generates a proposal grounded in your team's real track record — with a human checkpoint at every stage. Runs on your machine. Cloud LLMs are opt-in — your data stays local by default.

Five Steps — First Draft Done

From raw documents to a compliance-checked first draft — ready in hours for your team to polish and price.

Proposal generation workflow: Load Documents → Knowledge Base + RFPs → RFP Analysis → Proposal Outline → Cover Letter, Requirements Compliance Report, Draft Proposal → Post-Generation AI-Assisted Editing
1

CHATBOT QUERIES

Two chat interfaces let you ask plain-English questions — one against your company Knowledge Base (past proposals, bios, methodologies), one against the uploaded client RFP.

Why it matters: Quickly and informally assess key questions (client needs, contract type, project scope, location) and identify similar past projects and key personnel in your organization with relevant experience. Great for a quick go / no-go discussion.

2

RFP ANALYSIS

A four-pass extractor reads every page of the RFP (text + tables + figures) and produces a structured Analysis grid: Requirements, Personnel Constraints, Schedule Milestones, Evaluation Criteria, Submission Rules.

Why it matters: Every RFP requirement is identified EARLY in the process and then audited against the generated proposal. No last-minute scrambling to ask "Did we miss anything?"

3

PROPOSAL PLANNING

Choose your proposal structure — Build a manual outline, or let the AI auto-generate one that binds every RFP requirement to a specific section for full coverage.

Why it matters: Fix the shape of the proposal before you write a single word. Early planning decreases overall development time.

4

PROPOSAL GENERATION

Proposal synthesis merges RFP Requirements with your Company Knowledge Base to map your capabilities against client requirements. Once complete, separate AI Agents develop the Executive Summary, Conclusion, Cover Letter, and a Requirements Compliance Report scoring the finished draft against the RFP.

Why it matters: You get a complete package that is consistent and coherent — body, Cover Letter, Exec Summary, Conclusion — and the Requirements Compliance Report flags what's covered and what still needs SME attention.

5

AI-ASSISTED EDITING

AI-Assisted Editing helps refine any section with a targeted instruction to address missed requirements, tighten safety language, add case references, tone-match the client, and more. Integration with the Knowledge Base facilitates framing how your capabilities match client needs.

Why it matters: Within a few hours of receiving the RFP, you're finalizing the message, closing compliance gaps, and shaping the voice that will resonate with your client.

Draft ready to jumpstart your normal sales workflow — just review, validate, polish, and price.

Every RFP, tracked from intake to submission

One dashboard for every opportunity — status, client, site, due date. Drill in for RFP intake, analysis, draft proposal generation, AI-Assisted Editing, and an audit-ready compliance matrix.

ProjEx Propose product view.

What Sets Us Apart

This is not a general-purpose AI wrapper — Every design decision is tuned to marry the client's RFP requirements with your company's proven capabilities and engineered to generate winning proposals.

📖

Full Document Understanding

Commodity AI tools only skim the RFP, retrieving the 30-40 chunks their search filter ranks highest and dropping the rest. On a big industrial RFP with 100+ requirements, you can't afford to use this inefficient approach.

We read every chunk of every RFP, in sequence, using a Map-Reduce architecture that keeps every chunk in scope, and understands what your client actually needs.

~95% extraction recall
⚙️

Purpose-Built LLM Stack

No single AI technique handles every job well. ProjEx runs 12+ distinct AI agents — Chat, RFP Analysis, Outline Creator, Proposal Writer, Compliance Auditor, Cover Letter, AI-Assisted Editing, and more — each with its own model, retrieval method, synthesis parameters, and prompt tuned for the task. Hybrid retrieval for chat. Map-Reduce for full-document analysis. Per-category multi-label extraction for structured lists. Iterative self-critique on high-stakes prose.

12+ AI agents, each purpose-tuned
🏰

IP Protection

Your documents, past-project information, personnel qualifications, proposal templates, and methodologies are stored and indexed in the data location you control. ProjEx Advisors does not receive those working files during ordinary Software use and does not use Customer materials received for authorized onboarding or support to train public foundation models. If you select Local AI, document inference stays on your supported computer. Separate network communications still occur for licensing, component distribution, EULA acceptance records, and support you choose to send. If you select Cloud AI, the information needed for each operation is sent directly to your selected provider through your own account and under that provider’s terms and data policies.

You control what data (if any) sees the cloud
🔍

Verified Accuracy + Audit Trail

Quality gates run where they matter most — self-critique on high-stakes prose (Proposal Sections, Cover Letter, Exec Summary, Conclusion), confidence scoring on every structured extraction, hallucination detection on deliverables. Every extracted requirement is source-attributable back to its RFP location. The Requirements Compliance Report auto-maps every RFP requirement to the proposal section that addresses it, with evidence quotes.

Higher quality, less effort
🛡️

Hardware-Safe + Stable

Before every local query, we check whether your GPU has enough memory to handle it. If not, we split the work into multiple passes, right-size the context window for each, and synthesize the results — no truncation, no mid-analysis crashes. All logged locally for audit.

Runs on a single consumer-grade GPU
👤

You're in Control, Always

AI agents dramatically speed up the process — but you stay in control at every stage. Human checkpoints at RFP analysis, outline approval, section edits, and compliance tracking. Override any AI output. Nothing ships without your explicit approval. The AI drafts fast; you direct and decide what's right.

You approve every stage

Frequently asked questions

Grouped by topic. Click a question to expand the answer. Don't see what you're looking for? Email sales@projexadvisors.com.

Security & Data

Does my data leave my computer?

Your document files, local search index, analyses, and proposal files remain in the data folder you select and are not automatically uploaded to ProjEx Advisors. Indexing runs locally. If your tier and hardware support Local AI and you select it, document inference occurs on your computer. If you select Cloud AI, the text and, where applicable, images needed for the requested operation are sent directly to the cloud AI provider through your own account. A large analysis may send substantial portions of an RFP through multiple requests. Other network activity includes activation and periodic license checks, secure transmission of EULA acceptance records, delivery of approved software and model components, and support messages or diagnostic logs you choose to send.

Does ProjEx see my RFPs or proposals?

Not during ordinary Software use. ProjEx does not operate a product-content server that automatically receives or stores your working RFP files, Knowledge Base, local search index, or generated proposal files. During customer-led onboarding, calibration, training, or support, ProjEx personnel may view materials you choose to display or provide for that purpose. If you send an optional diagnostic log, the Software applies automated, best-effort redaction on your computer and lets you preview the payload before submission. Review the preview carefully because automated redaction cannot guarantee removal of every sensitive detail.

Does ProjEx train AI models on my materials?

ProjEx does not use Customer materials it receives for authorized onboarding or support to train public foundation models. During ordinary use, ProjEx does not receive your locally stored working documents. If you enable Cloud AI, information sent directly to your provider is governed by your provider account and that provider’s current terms, data policies, and settings. ProjEx does not control a third-party provider’s retention or independently selected data-sharing or training settings.

Can I submit what ProjEx Propose writes without review?

ProjEx Propose is an AI-assisted productivity tool that creates draft analyses and proposal content. It does not replace qualified subject-matter review or certify that an RFP response is accurate or compliant. Users remain responsible for thoroughly reviewing the applicable RFP and amendments; reviewing, editing, and approving generated content; verifying facts, figures, personnel information, qualifications, certifications, and compliance statements; and making all final pricing, strategy, legal, technical, and submission decisions. Generated content should not be submitted without qualified human review.

Setup & Requirements

What GPU and VRAM do I need for local LLM execution?

A Windows 11 (x64) computer equipped with an NVIDIA graphics card that supports sm_120 with at least 32 GB VRAM. The app was developed and tested on an RTX 5090; the RTX PRO 4500 / 5000 / 6000 Blackwell cards are also supported. Also recommended: 16 GB RAM (32 GB better) and ~50 GB free disk.

What if I don’t have an approved GPU?

For cloud-only usage (OpenAI backend), any modern Windows machine will run the app fine — no GPU required. Just supply an OpenAI API key on usage Tier 2 or higher (Tier 3 recommended), so rate limits and token throughput are adequate during proposal generation. You also need Microsoft Word for Windows: every deliverable is a Word document, and the app uses Word to make PDFs.

How long does install take?

The installer itself takes about 3 minutes to load, plus an additional 10-15 minutes to install the local LLM (if enabled). At installation, you will be given a license key to enter which will enable either Cloud-Only or Cloud + Local capabilities.

How many users can run the app at the same time?

One. ProjEx Propose is a single-user product — one license per machine, one active instance per database. At this time, the underlying data store is not built for shared, concurrent access; two people editing the same opportunity, knowledge base, or proposal at once is not supported.

Workflow

What file types can I ingest?

You can ingest the following file types into your Company KB or the RFP module: PDF, DOCX, DOC, PPTX, PPT, XLSX, TXT, MD, and common image formats (PNG, JPG, TIFF). Images and embedded diagrams get processed by a vision model.

What is the finished format of deliverables (Analysis, Proposal, etc.)?

Every deliverable — Analysis grid, Draft Proposal, Cover Letter, Executive Summary, Conclusion, and Requirements Compliance Report — exports to Microsoft Word (.docx). Each paragraph is assigned a custom Word Style (headings, body text, bullets, numbered lists, at three depth levels). Edit the style template before generation (font, size, color, spacing) to accommodate your corporate standards.

Do I need to write LLM prompts?

No. Prompts are managed by the vendor and ship with the app in your vertical's Sales Pack. For AI-Assisted Editing, you have the option to use pre-configured prompts (more concise, add case study, limit section to 3 paragraphs of 4 sentences each, etc) or to write your own custom prompts to get the desired results.

How do I submit vendor support requests?

The app has a built-in Contact Vendor form where you describe the issue and, optionally, attach a diagnostic log. Before anything is sent, the app applies automated, best-effort redaction on your computer (API keys, license keys, document text and other sensitive details are replaced with markers) and shows you a preview of exactly what will go. Review the preview: automated redaction cannot guarantee removal of every sensitive detail. Prefer email? Reach out to support@projexadvisors.com.

See it work on your RFPs

Contact us for more information.

We handle the information you send as described in our Privacy Policy.

Prefer email? Reach us at sales@projexadvisors.com.