Proposals: Over a Week to About a Day and a Half. Win Rate: From Around 20% to Over 30%.
A 45-person management consulting firm in Washington, DC with 15 years of engagement history.
A 45-person management consulting firm in Washington, DC with 15 years of engagement history.
What They Were Facing
A 45-person management consulting firm in Washington, DC had been in business for 15 years and had nothing to show for it institutionally. That's not a commentary on their work quality, which was strong. It's a statement about their knowledge management, which was nonexistent. Deliverables from past engagements lived on individual consultants' laptops, in disconnected SharePoint folders, and in email archives that nobody searched. The symptoms were predictable. Consultants spent roughly a quarter to a third of their time recreating frameworks, templates, and analyses that someone in the firm had already built for a previous client. A consultant working on a market entry strategy for a healthcare client had no way of knowing that a colleague had done nearly identical work 18 months earlier. So they'd start from scratch, spending dozens of hours producing something a predecessor had already figured out. The proposal process was where it hurt most. Each RFP response required over a week of effort: pulling credentials, writing case summaries, drafting technical approaches, assembling team bios, and creating cost estimates. Much of this content existed in some form from previous proposals, but finding it required asking around, searching inboxes, and hoping someone remembered. The firm was responding to fewer opportunities than they should have been because each proposal consumed so much capacity. Their win rate reflected the inefficiency. At around 20%, they were below the industry average for firms their size. The proposals weren't bad, they were just generic. Without easy access to their best past work, consultants fell back on boilerplate rather than tailoring approaches with the specific experience and results they'd actually delivered.
Consultants spending roughly a quarter to a third of their time recreating frameworks and analyses already built for previous clients
Past deliverables scattered across laptops, disconnected SharePoint folders, and email archives
Each RFP response required over a week of effort pulling credentials, case summaries, and cost estimates
Win rate around 20%, below industry average, due to generic proposals lacking specific past experience
Firm responding to fewer opportunities than capacity allowed because proposals consumed too much time
How We Solved It
The first step was an archive excavation. We worked with the firm's partners and a designated project team to collect, organize, and index 15 years of deliverables, proposals, case studies, templates, and internal research. This was roughly 8,500 documents totaling more than 40,000 pages. Much of it was in inconsistent formats, unnamed files in nested folders. We built a processing pipeline that extracted key metadata from each document: client industry, engagement type, date, team members, methodologies used, and outcomes delivered. The knowledge assistant we built sits on top of this indexed archive. Consultants can search in natural language. A query like "market sizing frameworks for B2B SaaS" returns the actual deliverables where the firm used those frameworks, not just document titles. The system surfaces relevant sections within documents, so a consultant can find the specific pricing model from a 2021 engagement without reading the entire 80-page report. For proposals, we built a drafting system that works from the firm's own history. When an RFP comes in, the system analyzes its requirements, identifies the most relevant past proposals and engagement summaries, and generates a first draft that pulls from the firm's actual credentials and experience. The draft includes suggested team compositions based on relevant experience, technical approach sections adapted from similar past wins, and placeholder sections that need original thinking. It's a majority-complete draft that a senior consultant can refine and customize, not a finished product. The system also generates "knowledge nudges." When a consultant starts working on a new engagement, the system proactively surfaces related past work based on the client profile, industry, and engagement scope. This was designed to address the problem of people not knowing what to search for because they didn't know what the firm had done before they joined.
Collected and indexed 8,500 documents totaling 40,000+ pages spanning 15 years of engagement history
Processing pipeline extracting client industry, engagement type, team members, methodologies, and outcomes
Natural language search returning relevant sections within documents, not just document titles
Proposal drafting system generating majority-complete first drafts from the firm's actual credentials and past wins
Knowledge nudge system proactively surfacing related past work when new engagements begin
Measurable Outcomes
Quantifiable improvements delivered within the project timeline
Reduced from over a week of effort to about a day and a half per proposal
Improved from around 20% to over 30% on proposal submissions
Consultant utilization increased roughly 15% from recovered rework time
Documents indexed spanning 15 years of engagements
Firm now responds to roughly 40% more RFPs per quarter with the same team
New hire ramp time reduced by an estimated 30%
The win rate improvement from around 20% to over 30% is the headline metric, and the firm's managing partner attributes it directly to proposal quality. With easy access to the firm's full body of work, proposals now include specific case examples, relevant methodologies the firm has actually used, and team compositions that align with the prospect's industry. They stopped sounding generic and started sounding like a firm with 15 years of relevant experience, because that's what they are. The utilization improvement translated to over a million dollars annually in recovered billable capacity, calculated at the firm's blended billing rate. Some of that capacity went to serving existing clients better (deeper analyses, more thorough deliverables). The rest went to pursuing new business, which is where the increased proposal volume came from. One partner put it simply: "We went from a firm that accidentally forgot everything it learned to one that builds on every engagement. It changed how we compete."
Implementation Timeline
A structured approach from discovery to deployment
Collected 8,500 documents and built metadata extraction pipeline
Weeks 1-3Collected 8,500 documents and built metadata extraction pipeline
Natural language search system with section-level retrieval
Weeks 4-6Natural language search system with section-level retrieval
RFP analysis and first-draft generation from firm history
Weeks 7-8RFP analysis and first-draft generation from firm history
Proactive surfacing integrated with project management tools
Week 9Proactive surfacing integrated with project management tools
Feedback collection and system refinement
Weeks 10-11Feedback collection and system refinement
Organization-wide access for all consultants
Week 12Organization-wide access for all consultants
Frequently Asked Questions
How did you handle client confidentiality across 15 years of deliverables?
Client confidentiality was a non-negotiable constraint. During the indexing phase, we worked with the firm's legal team to establish rules for what could be searchable and by whom. Client names and identifying details are stripped from the knowledge base index. The system returns sanitized references ("a Fortune 500 pharmaceutical company" rather than the actual client name). Consultants with specific engagement history can access the full documents through existing permission controls, but the search layer shows anonymized content by default.
What if two consultants are working on similar engagements and don't know it?
The knowledge nudge system addresses this directly. When new engagements are logged in the project management system, the knowledge assistant identifies overlap with other active engagements and notifies both project leads. This has already prevented duplicate work on two occasions in the first quarter. It also facilitates knowledge sharing between teams working in adjacent spaces.
How do you keep the knowledge base current as new engagements complete?
New deliverables are added to the knowledge base at engagement close as part of the firm's updated project closeout process. The processing pipeline extracts metadata automatically, and the project lead reviews and tags the submission. We built this into the existing workflow so it takes about 15 minutes per engagement rather than requiring a separate archival effort. The system also flags when an older document in a topic area might be outdated based on newer work in the same space.
Can the proposal drafting system match different RFP formats and requirements?
The system parses each RFP's specific requirements and structures the draft response accordingly. It's not a template system that fills in blanks. It generates section-by-section content matched to the RFP's evaluation criteria and page limits. That said, the output is always a draft. The firm's partners review and reshape every proposal before submission. The system handles the assembly and first-pass writing; humans handle the strategy and final polish.
Services Used in This Project
Ready for Similar Results?
Schedule a discovery call to discuss your specific challenges and learn how we can deliver measurable outcomes for your organization.
Schedule Discovery Call