Building an RFP Response Workflow

Navigate to Workflows → New Workflow → Trigger. Choose your intake method:

• Email trigger (Gmail): Connect your Google account and specify a label or filter pattern (e.g., subject contains "RFP" or "Request for Proposal" or is sent to rfp@yourcompany.com). RenderDraw monitors the inbox and fires the workflow when a matching email arrives with a PDF attachment.
• Email trigger (Outlook/Microsoft 365): Connect via OAuth. Configure the same label-or-folder-based filter. Works with shared mailboxes.
• Webhook trigger: If your company receives RFPs through a procurement portal (e-BUILDER, Procore Bid Management, BidSync), configure the webhook endpoint and map the payload fields to the workflow's document input.
• Manual trigger: For testing and one-off RFPs. Upload the document directly to the workflow run interface.

Set a trigger filter to prevent noise from triggering full workflow runs. For email triggers, filter on: attachment present AND (subject contains "RFP" OR "Request for Proposal" OR "Solicitation" OR "Invitation to Bid").

After the trigger, add a Parse Document block. This block handles the raw PDF or DOCX and converts it to structured content the downstream AI blocks can work with.

Configure the parser block with:

• Document source: Map to the trigger's attachment output (trigger.attachments[0])
• Parser mode: "RFP Document" (optimized for procurement language, extracts sections automatically)
• Extract sections: Enable all: Introduction, Requirements, Evaluation Criteria, Submission Instructions, Pricing Requirements
• OCR fallback: Enable for scanned PDFs. Set confidence threshold to 0.85 — below this, flag for manual review before continuing.
• Output format: Structured JSON with requirement objects (each containing: text, section, requirement_type, is_mandatory, evaluation_weight)

The parser block's output is the foundation for every subsequent step. Take time to test it with 3–5 representative RFPs before proceeding.

Add a Classify & Score block to evaluate the RFP before committing full automation resources to it. This block scores the opportunity against your predefined criteria and routes it accordingly.

Configure scoring dimensions:

• Industry match: Map extracted industry keywords against your target sectors. Score 0–10.
• Geographic fit: Extract project location and compare to your service footprint. Score 0–10.
• Contract size: Extract estimated value or project size. Configure your target range. Score 0–10.
• Capability coverage: Compare the requirements list against your product/service catalog. What percentage do you have a credible response for? Score 0–10.
• Deadline feasibility: Extract submission deadline and compute available days. Flag if under 7 days.

Configure routing rules:

• Score ≥ 35: Continue to full automation pipeline
• Score 20–34: Continue but flag for priority human review
• Score < 20: Route to no-bid notification (send email to sales lead, archive the workflow run)

You can adjust thresholds and dimension weights after reviewing your first 20–30 scored RFPs. The initial configuration is a starting point, not a final tuning.

Add an Extract Requirements AI block to take the parsed document structure and produce the definitive requirements list that will drive response generation. This is where your AI context configuration matters most.

AI block settings:

• AI provider: Claude (recommended for long-document comprehension with 200K+ token context window)
• System prompt: Use the "RFP Analyst" preset or customize with your industry's compliance vocabulary
• Input: Full parsed document text + structured sections from the parser block
• Output schema: Requirements array with fields: id, text, section, requirement_type (mandatory/preferred/informational), response_required (boolean), evaluation_weight (1–5), related_requirements (array of ids)

After extraction, add a Write to Workbook action to persist the requirements list. This creates the RFP workbook that tracks response status throughout the workflow.

Learn about RFP Workbooks →

Add a Query Knowledgebase block. This block takes each extracted requirement and finds the most relevant content from your indexed knowledge sources.

Configure the query block:

• Knowledgebase: Select your RFP-specific knowledgebase (or create one — see setup guide)
• Query mode: Per-requirement search (each requirement generates its own semantic query)
• Results per requirement: Top 3 (configurable; more results increase draft quality but slow generation)
• Minimum relevance score: 0.72 (below this, the requirement is flagged as "no content match — human authoring required")
• Source filters: Optionally restrict by content type (technical specs only, past proposals only, compliance docs only) per requirement type

The output of this block is a requirements-to-content mapping: for each requirement, a list of ranked, relevant content chunks from your knowledgebase. This feeds directly into the draft generation block.

For RFPs that include a pricing schedule or bill of materials requirement, add a Fetch Pricing block before draft generation. This block queries your CPQ or pricing system for current prices on the identified product/service scope.

Configure for your CPQ:

• Salesforce CPQ: Connect via OAuth. The block runs a price rule evaluation against the extracted scope. Maps RFP line items to Salesforce Products by keyword and part number. Returns a structured quote with line-level pricing and discount eligibility.
• Logik.io: Pass the extracted configuration requirements to Logik's runtime API. Validates configuration feasibility before pricing — catches specification conflicts before they reach the human reviewer.
• RenderDraw Native: If you use RenderDraw's configurator engine for your product catalog, the pricing block calls the built-in CPQ directly with no external integration required.
• Manual pricing: For scopes that can't be priced automatically, the block creates a pricing placeholder in the workbook and flags the line for manual completion before the draft generation step runs.

See all pricing provider integrations →

Add a Generate Draft AI block. This is the primary content generation step — it takes the requirement list, matched knowledgebase content, and pricing data and assembles the full proposal document.

Configuration options:

• Template: Upload your Word or PDF proposal template. The block maps generated content to template sections by section header or variable tags.
• AI provider: Claude Sonnet or Opus (recommended for long-form coherence). See AI Context for model selection guidance.
• Generation mode: "Synthesize from sources" (uses knowledgebase content as primary material, generates only where gaps exist) vs. "Full generation" (uses sources as context but writes original prose throughout).
• Include compliance matrix: Auto-generates a requirement-by-requirement compliance matrix as an exhibit. Maps each extracted requirement to its response location in the document.
• Confidence annotations: Flag sections where knowledgebase match score was below 0.80 with inline reviewer notes. These are the sections that need the most human attention.
• Pricing schedule: Auto-formats the CPQ output into the proposal's pricing exhibit template.

The output is a fully formatted draft document in your template, saved to the workflow run's document store and linked to the RFP workbook.

Add a Human Gate block — this is mandatory. No proposal ships from an automated workflow without human sign-off. Configure who reviews, how they're notified, and what the deadline is.

• Reviewer assignment: Route to a specific user, a role-based queue, or dynamically assign based on RFP properties (industry, geography, contract size). Configure escalation if no action within N hours.
• Review interface: Reviewers see the full draft with compliance gap highlights, low-confidence section flags, pricing summary, and the original RFP side-by-side. Inline editing, comments, and section-level approval are all available.
• Approval levels: Configure multi-level approval for large contracts (e.g., technical review by SME, then commercial review by sales director, then legal sign-off).
• Deadline tracking: The gate displays the RFP submission deadline and the hours remaining. Sends escalation notifications as the deadline approaches.

After the gate is approved, add a Deliver Proposal action block:

• Conga Composer: Merge the approved content into a Conga template and deliver via Salesforce opportunity record.
• DocuSign: Send for electronic signature before delivery if required by the procurement process.
• Email delivery: Send PDF attachment via the connected Gmail/Outlook account to the RFP contact extracted during parsing.
• Portal submission: For procurement portals with submission APIs, the delivery block handles the upload and returns the confirmation number.

Configure the post-delivery action: update the CRM opportunity, add the submitted proposal to the knowledgebase, and start the win/loss tracking timer.