C
Docs

AI Refinement

Use AI to improve and polish your proposal content

Updated 2026-03-3017 min read

AI Refinement

After generating a proposal, use AI refinement to improve clarity, compliance, and impact. Refinement helps polish rough content into submission-ready proposals.

Refinement Options

Improve Clarity

Make content clearer and easier to read:

  • Simplify complex sentences
  • Improve flow and transitions
  • Add structure and formatting
  • Eliminate jargon (or add where appropriate)

Strengthen Compliance

Ensure requirements are addressed:

  • Check all requirements referenced
  • Strengthen compliance language
  • Add specific requirement callouts
  • Improve traceability

Increase Impact

Make content more compelling:

  • Strengthen value proposition
  • Emphasize differentiators
  • Add quantified benefits
  • Improve executive summary punch

Adjust Tone

Modify the writing style:

  • More formal/less formal
  • More technical/less technical
  • More assertive/conservative
  • Match client culture

Tip

Use "Match client culture" refinement after researching the agency's communication style.

How to Refine

Refine Entire Section

Refine Selection

Refine specific text:

  1. Highlight the text to refine
  2. Right-click and select Refine Selection
  3. Choose refinement type
  4. Review and accept changes

Refine with Instructions

Provide specific guidance:

Refine this section to:
- Emphasize our cloud migration experience
- Add specific metrics from past projects
- Strengthen the opening paragraph
- Reduce length by 20%

Comparing Versions

Side-by-Side View

Compare original and refined versions:

  • Original on left, refined on right
  • Differences highlighted
  • Click any change to see detail

Change Tracking

Track all refinements:

  • Additions in green
  • Deletions in red
  • Modifications in yellow
  • Accept/reject individual changes

Version History

Access previous versions:

  1. Click History in section menu
  2. See all refinement iterations
  3. Restore any previous version
  4. Compare any two versions

Refinement Best Practices

Iterative Refinement

Don't try to perfect in one pass:

  1. First pass: Clarity and structure
  2. Second pass: Compliance and completeness
  3. Third pass: Impact and tone
  4. Final pass: Polish and proofread

Human Review

Always review AI refinements:

  • Verify factual accuracy
  • Check for introduced errors
  • Ensure company voice is maintained
  • Validate technical correctness

Warning

AI refinement is a tool, not a replacement for human judgment. Always verify refined content.

Preserve Key Messages

Before refining, identify:

  • Must-keep phrases
  • Key differentiators
  • Required compliance language
  • Client-specific terminology

Lock these elements or verify they remain after refinement.

Common Refinement Scenarios

Verbose Content

Problem: Generated content is too wordy

Solution:

  • Use "Improve clarity" refinement
  • Set target length reduction
  • Request "concise professional" tone

Missing Requirements

Problem: Not all requirements addressed

Solution:

  • Use "Strengthen compliance" refinement
  • Provide list of missing requirements
  • Request explicit requirement references

Weak Value Proposition

Problem: Content doesn't sell your solution

Solution:

  • Use "Increase impact" refinement
  • Add specific win themes to emphasize
  • Request quantified benefits

Inconsistent Tone

Problem: Different sections have different voices

Solution:

  • Select multiple sections
  • Apply consistent tone refinement
  • Use "Match" feature with a well-written section

Bulk Refinement

Refine multiple sections at once:

  1. Select sections (Cmd/Ctrl + click)
  2. Click Bulk Refine
  3. Choose common refinements
  4. Review each section

Note

Bulk refinement applies the same settings to all selected sections. Use for consistency.

Refinement Credits

Refinement uses AI credits:

  • Small refinements: Minimal credits
  • Full section refinement: Moderate credits
  • Multi-section bulk: Higher credits

View credit usage in the refinement dialog.

Advanced Refinement Techniques

Target-Driven Refinement

Refine with specific objectives:

Compliance Enhancement

Strengthen compliance for requirements 3.1-3.5:
- Add explicit requirement references
- Include acceptance criteria for each
- Quantify performance targets
- Add traceability to compliance matrix

Length Optimization

Reduce section length by 25% while maintaining:
- All requirement coverage
- Key differentiators
- Quantified metrics
- Professional tone

Value Proposition Strengthening

Enhance value proposition by:
- Adding quantified ROI calculations
- Emphasizing cost avoidance benefits
- Highlighting mission impact
- Comparing to alternative approaches

Style Matching

Match the RFP's or client's communication style:

Example - Formal Government Style:

  • Longer sentences with subordinate clauses
  • Passive voice acceptable
  • Traditional terminology ("shall", "will", "the contractor")
  • Conservative, risk-averse language

Example - Modern Tech Style:

  • Shorter, direct sentences
  • Active voice preferred
  • Contemporary terminology
  • Innovation and agility emphasized

Competitive Positioning Refinement

Strengthen your position against likely competitors:

Against Incumbent

Refine to emphasize:
- Fresh perspective and innovation
- Proven transition expertise
- Cost efficiency improvements
- Modern technology stack

Against Large Integrators

Refine to emphasize:
- Agility and responsiveness
- Partner-level attention
- Decision-making speed
- Specialized expertise

Against Small Firms

Refine to emphasize:
- Enterprise scale and resources
- Financial stability
- Comprehensive capabilities
- Deep bench strength

Tip

Only refine for competitive positioning if you have intelligence about likely competitors. Generic positioning weakens proposals.

Evaluator-Focused Refinement

Optimize for how evaluators read proposals:

For Technical Evaluators

Refine to:
- Use precise technical terminology
- Include architecture diagrams and technical details
- Reference standards and specifications
- Explain trade-offs and alternatives considered

For Business Evaluators

Refine to:
- Emphasize business benefits and ROI
- Minimize jargon, explain technical concepts
- Focus on risk mitigation
- Highlight management and oversight

For Executive Evaluators

Refine to:
- Lead with strategic value
- Use compelling opening statements
- Include visual summaries
- Focus on mission impact

Refinement Workflows

Three-Pass Refinement Strategy

Pass 1: Structure & Compliance (20% of time)

  • Verify all requirements addressed
  • Check section flow and transitions
  • Validate heading hierarchy
  • Ensure consistency across sections

Pass 2: Content Quality (50% of time)

  • Improve clarity and readability
  • Strengthen value propositions
  • Add quantified benefits
  • Remove redundancy
  • Enhance technical depth

Pass 3: Polish (30% of time)

  • Fine-tune tone and style
  • Perfect grammar and mechanics
  • Optimize for page limits
  • Final proofreading
  • Formatting consistency

Section-Specific Refinement Priorities

Executive Summary

  1. Compelling opening hook
  2. Clear value proposition
  3. Win theme integration
  4. Brevity (2-3 pages max)

Technical Approach

  1. Requirement traceability
  2. Technical accuracy
  3. Rationale for decisions
  4. Clarity for non-technical readers

Management Approach

  1. Specificity to this project
  2. Clear governance structure
  3. Risk mitigation integration
  4. Client collaboration emphasis

Past Performance

  1. Relevance to current RFP
  2. Quantified achievements
  3. Client reference details
  4. Similarity demonstration

Bulk Refinement for Consistency

Apply consistent refinements across multiple sections:

Good Uses for Bulk Refinement:

  • Harmonizing tone across all sections
  • Standardizing terminology (e.g., "client" vs. "customer" vs. "agency")
  • Applying length reductions proportionally
  • Adding consistent win theme emphasis

Avoid Bulk Refinement For:

  • Section-specific technical improvements
  • Compliance enhancements (requirements differ by section)
  • Major restructuring

Refinement Quality Measurement

Before/After Metrics

Track improvement with these metrics:

MetricBeforeAfterTarget
Readability (Flesch-Kincaid)455850-60
Average Sentence Length28 words22 words20-25
Passive Voice %25%12%<15%
Jargon DensityHighMediumLow-Med
Requirement Coverage85%98%>95%
Compliance Score7289>85

Refinement Impact Score

Each refinement shows expected impact:

  • High Impact (🔥): Significant quality improvement expected
  • Medium Impact (⚡): Moderate improvement
  • Low Impact (💡): Minor polish

Focus on high-impact refinements first.

A/B Comparison

Compare refinement options side-by-side:

Common Refinement Scenarios

Scenario 1: Too Much Technical Jargon

Problem: Technical evaluators may be fine, but business evaluators are lost.

Refinement Strategy:

Simplify for non-technical audience:
- Explain acronyms on first use
- Replace jargon with plain language equivalents
- Add analogies for complex concepts
- Include glossary references

Example Transformation:

Before:

"Our microservices architecture leverages containerization via K8s with service mesh implementation for inter-service communication, deployed on a multi-AZ cloud infrastructure with auto-scaling HPA policies."

After:

"Our modern system architecture breaks the application into independent components (microservices) that can be updated and scaled individually. These components run in containers (isolated packages with everything needed to run), managed by Kubernetes (K8s). This design provides high availability across multiple data centers and automatically scales to handle demand spikes."

Scenario 2: Generic, Template-Like Content

Problem: Content feels like it could apply to any company or any project.

Refinement Strategy:

Add specificity and company differentiation:
- Replace generic statements with specific examples
- Add quantified achievements from past projects
- Include unique methodologies or tools
- Reference specific RFP requirements
- Name key personnel who will be involved

Example Transformation:

Before:

"We have extensive experience in software development and have successfully delivered many projects on time and within budget."

After:

"Our team has delivered 47 web applications for federal agencies since 2018, including three Department of Veterans Affairs systems serving 2.3M users. In the most recent VA project (Contract #VA-123-18-D-0045), we delivered 18 releases over 24 months with 99.7% uptime and zero security incidents, coming in 8% under budget. Project Manager Sarah Chen and Lead Architect Michael Torres from that project will lead this engagement."

Scenario 3: Weak Value Proposition

Problem: Proposal describes what you'll do, but not why the client should care.

Refinement Strategy:

Transform features into benefits:
- For each capability, explain the client benefit
- Quantify impact where possible
- Connect to agency mission
- Address client pain points explicitly
- Use benefit-oriented language

Example Transformation:

Before:

"We will implement an automated testing framework with continuous integration and deployment pipelines."

After:

"Automated testing with continuous integration will reduce your software defect rate by an estimated 60% (based on our similar VA project), cutting the time from bug discovery to fix from 3 weeks to 3 days. This means faster delivery of mission-critical features to warfighters, fewer production incidents disrupting operations, and lower long-term maintenance costs—we estimate $450K in cost avoidance over the 3-year contract period."

Scenario 4: Exceeding Page Limits

Problem: Content is 20% over the RFP page limit.

Refinement Strategy:

Reduce length by 20% while preserving:
- All requirement coverage
- Key differentiators
- Quantified metrics
- Compliance statements

Methods:
- Remove redundancy and repetition
- Combine similar points
- Convert prose to tables or bullets
- Move supporting details to appendices
- Tighten sentence structure

Editing Tactics:

  • Replace "in order to" with "to"
  • Replace "due to the fact that" with "because"
  • Eliminate "very", "really", "quite"
  • Combine short, choppy sentences
  • Remove passive constructions

Scenario 5: Missing Requirements

Problem: Compliance matrix reveals gaps in requirement coverage.

Refinement Strategy:

Add requirement coverage for [list requirement IDs]:
- Create new subsections for uncovered requirements
- Add explicit requirement callouts
- Include acceptance criteria
- Reference requirement IDs
- Update compliance matrix

Structured Addition:

For each missing requirement:

  1. Create a clear subsection with requirement reference
  2. State how you comply
  3. Provide evidence (past performance, methodology, etc.)
  4. Include validation/acceptance approach
  5. Cross-reference in compliance matrix

Scenario 6: Inconsistent Terminology

Problem: Same concept referred to differently across sections.

Refinement Strategy:

Standardize terminology across all sections:
- "the System" → "the Modernized Logistics Platform (MLP)"
- "client" → "the Agency"
- "end users" → "warfighters"
- "we" → "Acme Corporation"

Apply consistently in all sections.

Use bulk refinement to apply terminological consistency across the entire proposal.

Scenario 7: Weak Opening Statements

Problem: Sections start with boring, generic statements.

Refinement Strategy:

Create compelling opening hooks:
- Start with the client's key challenge
- Lead with a surprising statistic or insight
- Begin with a strong value statement
- Reference a relevant past success

For each major section, craft a strong opening sentence.

Example Transformations:

Before:

"This section describes our management approach for the project."

After:

"Effective management transforms complex IT modernizations from high-risk ventures into predictable successes—our structured approach has delivered 23 consecutive on-time, on-budget deployments."

Before:

"We have significant past performance in similar projects."

After:

"When the Department of Veterans Affairs needed to modernize their legacy benefits system serving 20M veterans, they chose Acme Corporation—and we delivered in 18 months with zero downtime during cutover."

Refinement Ethics & Best Practices

What AI Refinement Should Do

Improve clarity - Make complex ideas easier to understand ✅ Enhance structure - Organize content more logically ✅ Strengthen language - Replace weak verbs, improve flow ✅ Ensure consistency - Standardize terminology and style ✅ Tighten prose - Remove wordiness, improve conciseness ✅ Verify compliance - Check requirement coverage

What AI Refinement Should NOT Do

Invent facts - Don't add capabilities you don't have ❌ Overstate - Don't exaggerate achievements ❌ Plagiarize - Don't copy competitor or reference materials ❌ Remove required content - Don't delete mandated elements ❌ Change technical accuracy - Don't alter correct technical details ❌ Make commitments - Don't add promises without approval

Warning

Always validate that refinements maintain factual accuracy. AI can inadvertently introduce errors or overstatements. Human review is mandatory.

The 80/20 Rule

  • 80% of value comes from refining high-impact sections (Executive Summary, Technical Approach, Past Performance)
  • 20% of value comes from refining supporting sections

Focus your refinement effort where it matters most:

  1. Executive Summary (highest impact - read by all evaluators)
  2. Top-weighted evaluation factor sections (often Technical Approach)
  3. Past Performance (proof you can deliver)
  4. Other required sections
  5. Optional sections (if time permits)

When to Stop Refining

Diminishing returns set in after 3-4 refinement passes:

Stop When:

  • Quality score >85 and stable
  • No more obvious improvements identified
  • Requirements 100% covered
  • Within page limits
  • Tone and style consistent
  • Technical accuracy validated

Don't Fall Into:

  • Endless tweaking of minor word choices
  • Over-engineering perfect sentences
  • Refinement paralysis (afraid to submit)

Tip

Perfect is the enemy of good (and timely). Aim for 85-90% quality, not 100%. Use saved time to strengthen other sections or get peer review.

Collaborative Refinement

Team Review Workflows

Distributed Review:

  1. Assign sections to subject matter experts
  2. Each expert reviews and refines their sections
  3. Technical lead reviews all technical sections for consistency
  4. Proposal manager reviews all for compliance and flow
  5. Executive reviewer reads executive summary and spot-checks

Round-Robin Review:

  1. Each team member reviews a different section
  2. Rotate sections among reviewers
  3. Fresh eyes catch issues the writer missed
  4. Combine insights from multiple perspectives

Comment-Driven Refinement

Use comments to guide refinement:

Approval Gates

Set up approval stages:

  1. Author Self-Review: Author refines to their satisfaction
  2. Peer Review: Colleague reviews and suggests refinements
  3. SME Review: Subject matter expert validates technical accuracy
  4. Compliance Review: Proposal manager checks requirements coverage
  5. Executive Review: Senior leader approves final version

Use proposal status to track progress through gates.

Refinement Keyboard Shortcuts

Speed up refinement with shortcuts:

ActionShortcutDescription
Refine SelectionRefine highlighted text
Refine SectionRefine entire section
Accept ChangesAccept refinement
Reject ChangesReject refinement
Compare VersionsShow side-by-side diff
Undo RefinementRevert last refinement
Redo RefinementReapply last refinement

Frequently Asked Questions

Next Steps

Was this page helpful?

AI Refinement | Cothon Docs | Cothon