Improving AI Output Quality

Get better proposals by optimizing your inputs, Proposal Brain, and generation settings.

Not getting the quality you expect from CraftRFP? Here's how to improve your AI-generated proposals.

Quality Factors

AI output quality depends on:

  1. RFP clarity - How well the AI understands the requirements
  2. Proposal Brain - Examples the AI can learn from
  3. Settings - How you configure generation
  4. Context - Additional information you provide

Improving RFP Input

Clean Up Your RFP

Before uploading:

  • Remove headers, footers, and page numbers if possible
  • Ensure text is selectable (not scanned images)
  • Include all relevant sections

Paste Key Sections

If the full RFP is messy, paste just the critical parts:

  • Scope of work
  • Requirements list
  • Evaluation criteria
  • Submission instructions

Optimizing Your Proposal Brain

Upload Winning Proposals

The AI learns best from:

  • Proposals you've won - These teach what works
  • Similar industries - Match to your current target
  • Recent work - More relevant than older proposals

Quantity Matters

Past ProposalsAI Quality
0Good (uses best practices)
1-5Better (learns your basics)
5-15Great (understands your style)
15+Excellent (highly personalized)

Tag Strategically

Help the AI find relevant examples:

  • Tag by industry
  • Tag by contract size
  • Tag by service type

Generation Settings

Tone Selection

ToneBest For
ProfessionalGovernment, enterprise clients
FriendlySmall businesses, startups
TechnicalIT, engineering RFPs

Length Setting

  • Concise - When RFP has page limits
  • Standard - Most situations
  • Comprehensive - Complex technical proposals

Adding Context

Use the notes field to provide:

  • Your company's unique strengths
  • Relevant past project examples
  • Specific win themes to emphasize
  • Client relationship history

After Generation

Review the Quality Score

  • Below 70% - Consider regenerating with better inputs
  • 70-85% - Good base, edit to improve
  • 85%+ - Strong draft, minor tweaks needed

Iterate with Section Regeneration

Instead of regenerating everything:

  1. Keep strong sections
  2. Regenerate weak sections with specific guidance
  3. This preserves good content while improving weak areas

Still not getting good results? Contact support - we can review your setup.

Was this article helpful?

Let us know if this article answered your question or if you need more help.