The proposal that wins is not the most impressive-sounding one. It is the one that most clearly demonstrates that the firm understands the client’s problem, has solved it before, and has the team and methodology to solve it again.
These three elements (problem understanding, relevant experience, credible approach) are present in every winning proposal from a capable firm.
They are present inconsistently because the proposal process at most firms depends on who is available to write and how much time they have.
An AI-assisted proposal process makes all three elements present in every proposal regardless of who is writing and how much time they have. The AI context pack is the prerequisite — without it, proposals sound like they could have been written for any firm in the category.
This article describes the specific AI-assisted proposal process for a professional services firm: the Foundation build, the section-by-section workflow, the quality gate, and the proposal library that makes every subsequent proposal better than the last.
The proposal workflow is one of the three highest-return starting workflows for most firms. For the full prioritised list of workflows to deploy first, see the ten operations workflows your company should automate with AI first.
The Foundation build — five documents, eight hours of work
Document 1: Proposal format standards (90 minutes)
The section structure the firm uses for its standard proposal type, with depth and length guidelines for each section.
Standard section structure for a professional services proposal:
1. Cover page and executive summary (1–2 pages): the firm's understanding of the problem,
the proposed approach in plain language, the key differentiator, the team, the fee range.
2. Project understanding and approach (2–4 pages): the firm's analysis of the project
requirements, the proposed methodology, the phase sequence, the key decision points.
3. Relevant experience (1–2 pages): 3–4 project profiles demonstrating experience
with comparable work.
4. Team qualifications (1–2 pages): the proposed project team, their relevant experience,
and their specific roles.
5. Fee and schedule (1 page): the fee summary, the billing structure, the project schedule.
6. Firm qualifications and references (1 page): firm background, size, 2–3 client
references.Length guidelines: total proposal 8 to 12 pages for standard RFPs, adjusted for the client’s stated page limits.
Include in this document: the specific headings the firm uses, the tone guidelines (formal vs conversational varies by client type), and any RFP-specific format requirements the firm frequently encounters.
Document 2: Capabilities vocabulary guide (60 minutes)
The specific terminology that describes the firm’s approach, methodology, and differentiators: the language that characterises the firm’s work at the level of quality that distinguishes it from generic descriptions.
Three sections to include:
Firm differentiators: how the firm describes what makes it different, not “we are committed to quality.” Specific differentiators the firm can demonstrate. For an engineering firm: “our embedded project review process identifies constructability issues at design development rather than construction documents, reducing contractor RFIs by 30 to 40% on comparable projects.”
Methodology vocabulary: the specific terms the firm uses for its approach. For an engineering firm: the specific phases, the specific review processes, the specific tools and their role. For a legal firm: the specific work product categories, the matter management approach, the specific regulatory frameworks the firm navigates.
Client type vocabulary: how the firm describes its work differently for different client types: government clients vs private developers vs non-profits. The vocabulary calibration that signals sector expertise.
Document 3: Project portfolio library (3 to 4 hours — the highest-leverage investment)
Structured descriptions of 15 to 30 past projects in a format AI can use to match relevant experience to RFP requirements. Each entry follows a consistent structure:
PROJECT: [Project name and location]
CLIENT TYPE: [Government / Private developer / Non-profit / Corporate]
PROJECT TYPE: [Category — e.g., healthcare facility, commercial office, transportation]
SCOPE: [What the firm did — specific services, specific phases]
SCALE: [Size — square footage, budget, duration]
CHALLENGE: [The specific problem or constraint the firm navigated]
APPROACH: [What the firm did differently — specific to this project]
OUTCOME: [Measurable result: delivered on schedule, reduced cost, improved performance]
RELEVANT FOR: [RFP types where this project is most useful as a reference]Building the library:
Assign the 15 to 30 most relevant past projects to team members who worked on them. Each team member completes the structured entry form in 15 to 20 minutes. Review for consistency and quality. Upload to the Proposals Project.
This library is the most important single document in the AI-assisted proposal Foundation. The quality and completeness of the library directly determines the quality of the relevant experience sections AI produces.
Document 4: Client communication standards (45 minutes)
How the firm communicates in proposals for different client types and procurement contexts.
Dimensions to specify:
Government and public agency RFPs vs private client proposals vs negotiated engagements: tone shifts from formal to more conversational, technical depth shifts from maximum to calibrated.
Technical client reviewers (an engineer reviewing an engineering proposal) vs non-technical client reviewers (an owner’s representative reviewing the same proposal): the language accessibility calibration.
Competitive sealed bid vs negotiated selection vs invited shortlist: the competitive positioning intensity calibration.
Document 5: Competitive positioning guide (45 minutes)
How the firm positions when competing against the most common competing firms or approaches.
Structure: one paragraph per competitive situation the firm regularly encounters.
Example: “When competing against [Firm type A], our differentiation is [X]. The comparison that favours us is [Y]. The comparison that favours them is [Z]: address this by [positioning approach].”
This document enables AI to produce executive summary language and approach sections that are positioned for the competitive context, not just generically describing the firm’s capabilities.
The section-by-section workflow
The section-by-section workflow follows this sequence: project portfolio matching first, executive summary last.
Most firms write the executive summary first. This is backward. The executive summary written before the technical sections are complete cannot accurately synthesise the firm’s specific approach for this engagement.
Step 1: Project portfolio matching (10 minutes, before any drafting begins)
Input to the Proposals Project: the RFP requirements summary (the scope of services, the client type, the project type, the stated evaluation criteria).
Ask AI to identify the three to five most relevant project portfolio entries and explain why each is relevant to this RFP.
Output: a ranked list of relevant past projects with the specific relevance to this RFP’s requirements identified. The business development lead confirms the selection or overrides based on client relationship context.
This step eliminates the most time-consuming pre-writing task: the research through past project files to identify what is relevant.
Step 2: Relevant experience section (20 to 30 minutes)
Input: the selected project portfolio entries plus the RFP requirements summary. Specify the number of project profiles to include (typically 3 to 4) and the page limit.
AI output: the relevant experience section with 3 to 4 project profiles formatted in the firm’s standard project profile format, with the relevance to this RFP’s specific requirements drawn out in each profile’s lead sentence.
Human review: the business development lead confirms the relevance framing and adds any project-specific detail that is not in the portfolio library entry.
Step 3: Team qualifications section (15 to 20 minutes)
Input: the proposed project team (names, roles, relevant experience highlights) and the RFP requirements for team qualifications.
AI output: the team qualifications section with each team member’s relevant experience framed against this RFP’s requirements, using the firm’s standard team bio format.
Human review: project manager confirms the role descriptions and adds any specific qualifications relevant to this project that are not in the standard bios.
Step 4: Technical approach section (45 to 60 minutes)
Input: the RFP scope of services, the proposed methodology (provided as bullet points or rough notes by the lead professional), the relevant experience section (already drafted), and the competitive positioning context.
AI output: the technical approach section in the firm’s standard format: the methodology description, the phase sequence, the key decision points, and the approach differentiators framed against the RFP requirements.
Human review: this section receives the most rigorous review of any proposal section.
The lead professional who is proposing the methodology reviews every paragraph for technical accuracy, appropriate depth, and competitive positioning. This is where the firm’s professional expertise most directly enters the document.
Step 5: Fee and schedule section (15 minutes)
Input: the fee amount and structure (determined by the firm’s estimating process, not by AI), the project schedule, the billing structure.
AI output: the fee and schedule section formatted in the firm’s standard format with the relevant assumptions and exclusions included.
Human review: partner or managing director confirms the fee amount and the assumptions.
Step 6: Executive summary (20 to 30 minutes — after all technical sections are complete)
Input: all completed technical sections, the RFP evaluation criteria, the competitive positioning context.
AI output: the executive summary that synthesises the firm’s approach, relevant experience, team, and key differentiators into a 1 to 2 page summary calibrated to the evaluation criteria.
Human review: the senior professional reviews the executive summary for positioning accuracy and relationship calibration. The executive summary is the most relationship-sensitive section and receives the most attention in the review.
The full proposal: 3 to 6 hours with AI vs 10 to 20 hours without.
The time savings are concentrated in steps 1 to 3 and 5, where AI handles the structural work that previously required the most research and formatting time.
Steps 4 and 6 still require significant human input, but the human time is now concentrated on the judgment and positioning work rather than the production work.
The proposal library as the compound improvement mechanism
Post-submission debrief (15 minutes per proposal)
After every submission, the business development lead records:
- Which sections received positive feedback from the client (if known)
- Which sections required the most client clarification requests during the project
- Whether the proposal won or was unsuccessful
- If unsuccessful: the known reason (price, competitor strengths, qualification gap)
This debrief is not a post-mortem. It is a quality signal. The pattern across 10 to 20 proposals identifies which sections consistently score well, which need improvement, and which competitive positioning approaches are effective.
Quarterly library update (60 to 90 minutes per quarter)
The AI system owner or business development director reviews the debrief records from the quarter and updates:
- The project portfolio library: adds new completed projects, updates the relevance notes based on which past projects were most effective as references
- The competitive positioning guide: updates the positioning paragraphs based on which approaches proved effective and which did not
- The technical approach vocabulary: refines the methodology description based on client feedback patterns
After four quarterly updates: the proposal Foundation reflects sixteen months of learning about what works in this firm’s specific competitive environment. The proposals at month sixteen are measurably better than the proposals at month two — because the Foundation has been refined through the proposal library feedback loop.
Common questions on AI-assisted proposals
”What about page-limited RFPs where every word counts?”
The AI-assisted process is more effective on page-limited RFPs, not less. The AI draft produces consistent density across all sections, making it easier to identify which sections can be trimmed and which require the full page allowance.
The senior professional’s review time on page-limited RFPs shifts from “cutting what was already written” to “calibrating the depth of each section against the page budget and evaluation criteria weight.” This is a better use of their time.
”What about RFPs where the format is completely prescribed by the client?”
Update the proposal format standards document to reflect the prescribed format before running the session. The AI draft follows whichever format the Foundation specifies.
For one-off prescribed formats (unusual section structure, specific response form fields): input the format requirements directly in the session prompt.
The portfolio matching step (Step 1) and the technical approach drafting step (Step 4) are the most valuable components regardless of format prescription. They do not depend on the format standard.
”What if the firm’s best differentiators are confidential or sensitive?”
The project portfolio library and the capabilities vocabulary guide stay within the shared Proposals Project and are not shared externally.
The AI workspace terms of service (Claude Teams or ChatGPT Teams) provide the data handling governance for information loaded into the workspace.
For genuinely confidential differentiators (proprietary methodologies, specific client relationships, competitive intelligence): describe the differentiator at the level of specificity appropriate for the proposal context without naming the confidential source.
The AI uses the description, not the underlying confidential information.
Want the proposal Foundation built, the project portfolio library populated, and the first AI-assisted proposal produced in the next two weeks?
The AI-assisted proposal process reduces proposal drafting time by 60 to 70% while improving consistency and win rates.
After twelve months of the AI-assisted process, the firm’s proposals are not just faster — they are consistently better than the pre-AI baseline, and measurably better than the weakest proposals the firm was submitting before the process was in place.
Path one: build the project portfolio library this week. Assign 10 past projects to the team members who worked on them. Have each complete the structured entry form in 15 to 20 minutes. Compile the entries into a single document and upload it to your Proposals Project. Run the portfolio matching step (Step 1) on the next RFP you receive. Compare the quality and speed of the experience section to your current process.
Path two: bring in a partner. Phos AI Labs builds the Proposals Project Foundation, populates the project portfolio library, and runs the first AI-assisted proposal alongside the business development lead before the end of month two. Thirty minutes, no deck. Start here.
Related articles