How Pitch Box Builds a Self-Populating Pursuit Knowledge Base for Experiential Agencies

Three RFPs land in the same week. The BD lead who could reconstruct any case study from memory gave notice a month ago. Whoever picks up the pursuit now is flipping through eighteen-month-old decks, hunting for the one activation story that matches this brand's audience profile, and coming up empty by 9pm the night before the first draft is due.

Brian Morgan watched this pattern from the inside, subcontracting pursuit work for experiential agencies including Czarnowski and Sommers House while delivering programs for clients like Intel and HubSpot. "The RFP grind is a capacity problem, not a talent problem," said Brian Morgan, Founder of Sandbox Group LLC and Pitch Box. "Your best people aren't slow; the evidence is scattered."

Pitch Box exists to fix the architecture, not the person. It replaces the shared drive everyone means to update with a knowledge base that builds itself from the pursuit work an agency is already doing.

Why Most Agency Content Libraries Die in the First 90 Days

Someone stands up a shared drive during a slow week, tags three case studies, and calls it a library. Bid season lands six weeks later. Nobody has time to file a new asset, let alone find one. The library doesn't fail from neglect. It fails because it was designed to be maintained, and maintenance is exactly what disappears when RFPs stack up.

Four conditions kill a manual library before the quarter ends: no owner with protected time to run it, no filing convention that survives contributor turnover, no feedback loop routing a submitted proposal back into the library, and no consequence for skipping the tagging step under deadline pressure. Each condition reinforces the next, which is why the pattern repeats across agencies regardless of headcount or which tool sits on top of it.

"Your best case studies are trapped in old decks and in people's heads," said Brian Morgan, Founder of Sandbox Group LLC and Pitch Box. "That's the real bottleneck." The standard content library is built to be maintained. A self-building pursuit knowledge base is built so the library grows from live pursuit artifacts automatically, with no curation step sitting between a submitted proposal and a retrievable proof point.

What a Mature Pursuit Knowledge Base Contains, and What It Costs to Build by Hand

A mature pursuit knowledge base holds flagship case studies with full outcome data and source attribution, proof points tagged by brand vertical and audience demographic, evaluation-criterion maps tied to specific RFP language, boilerplate validated against past win submissions, and win/loss annotations showing which arguments actually scored. That's the inventory a senior BD lead carries in working memory after five years at one agency. The job of a library is to make that inventory institutional instead of personal.

Built by hand, that inventory has a labor cost most agencies have never measured. For a 100-person experiential agency pursuing roughly 40 RFPs a year, a coordinator spending six hours per pursuit tagging and filing artifacts generates about 240 hours of annual curation labor, somewhere between $18,000 and $24,000 at a $75 to $100 blended rate . That figure excludes the senior review hours required to validate a claim before it's trustworthy enough to retrieve again.

And that 240-hour estimate assumes complete follow-through on filing after every pursuit. In practice, tagging drops off sharply during high-volume bid seasons, which is exactly when the library would matter most. The cost isn't just the labor that was never spent. It's the proof points that were never captured and therefore can't be retrieved on the next pursuit.

What Makes a Knowledge Base Self-Building Instead of Self-Maintaining?

A self-maintaining library depends on someone choosing to file and tag a proof point after a pursuit ends. A self-building one skips that step. Pitch Box extracts proof points and source attribution directly from each submitted proposal, so the library grows as a byproduct of pursuit work, not an added chore.

A self-building pursuit knowledge base is a proof-point library that grows automatically from an agency's live pursuit work rather than from manual tagging. It works by ingesting each submitted proposal and case study, extracting proof points with source attribution, and mapping them to evaluation criteria the moment they're created.

The seeding mechanism is an engineering decision, not a marketing claim. Pitch Box scrapes an agency's existing case study pages and ingests past RFP responses at initial setup, extracting proof points with full source attribution. Every pursuit afterward adds to the base automatically. Grounded generation is the methodology behind that ingestion, and it runs in a fixed sequence:

1. Ingest the RFP and extract requirements, evaluation criteria, and the buying committee.
2. Retrieve matching proof from the agency's own case-study library and self-building knowledge base.
3. Draft each section with every claim traceable to that evidence and zero invented facts.
4. Run the Bid Qualifier go/no-go check before senior hours go into the pursuit.
5. Hand the sharpened draft to the human team, who owns the final submission.

"An AI that invents your proof points will lose you the pitch," said Brian Morgan, Founder of Sandbox Group LLC and Pitch Box. "Grounding isn't a feature; it's the whole game." Every generic answer library examined during Pitch Box's design required a person to decide a proof point was worth filing, apply a taxonomy, and validate the claim before it became retrievable. Pitch Box inverted that sequence: curation happens as a byproduct of pursuit activity, not a prerequisite to it.

How Each Pursuit Makes the Next One Cheaper and Sharper

Start with one pursuit. If locating and verifying proof points previously took roughly 20 senior hours per RFP at a blended rate of $150 to $200 an hour, and retrieval against a self-building library cuts that to about 6 hours, the recovered time is 14 hours, or $2,100 to $2,800 per RFP . These are illustrative figures, not guaranteed outcomes.

Scaled to a 100-person agency running 40 pursuits a year, 14 recovered hours per pursuit adds up to roughly 560 hours annually, or $84,000 to $112,000 at the blended senior rate. That number grows over time rather than holding flat: each submitted proposal adds proof points that reduce retrieval time on the next pursuit, so the per-pursuit cost keeps falling across the life of the library.

This is the framing a managing partner or CFO recognizes as structurally different from a flat software subscription. It's not just labor saved once; it's an asset that gets more valuable with use.

Proof, Traceability, and the Turnover Test

A library's real value shows up under two conditions: a senior BD lead gives notice mid-pursuit, or a loaded bid season arrives with more qualified RFPs than the team can pursue at once. In both cases, value isn't how much is stored. It's retrieval speed and claim reliability under time pressure.

Pitch Box extracts requirements and evaluation criteria from an incoming RFP (roughly 26 sections, parsed in about 60 seconds) and matches those criteria to specific case studies and proof points already in the knowledge base. Every retrieval carries its source: the original document or case study page the proof point came from. That source traceback is why 100% of drafted claims stay traceable to evidence and why the engine produces zero hallucinated facts. This is the whole architecture: the human sharpens and owns the final pitch, and the engine removes the blank-page grind and the scramble to find the right proof point.

The platform runs alongside tools an agency already has: Google Drive, Box, Slack, HubSpot, Salesforce, Asana, and Canva, coordinated through a built-in Workflow system for user and task management. Every tier includes unlimited seats, so the whole pursuit team works in one place without a per-seat cost conversation. Once a program is won, the Consistency Engine locks in a North-Star for that account: the goals, KPIs, voice, themes, factual ground, and scope on which it was won, and delivered work gets measured against it from there.

Pursuit capability that lives in a person's inbox is a retention risk. Pursuit capability encoded in a knowledge base with traceable source attribution is an institutional asset that survives the departure.

What to Build First

Start with three to five flagship case studies that carry full outcome data and source attribution. These are the highest-reuse proof points in any experiential agency's pursuit library, and seeding them first gives the ingestion engine a validated base before the first live RFP run.

Before expanding to the full archive of past proposals, run retrieval against one live or recent RFP. Compare the proof points surfaced to what a senior BD lead would have picked manually, note the gaps, and adjust the source material before scaling. Skipping this step is the most common implementation failure: a large library with low retrieval precision erodes trust before the tool earns it.

"You don't need a faster blank page," said Brian Morgan, Founder of Sandbox Group LLC and Pitch Box. "You need your own wins, retrievable at pursuit speed." For a managing partner watching pursuit spend buried in salaries with no line item to point to, or a stretched BD lead rebuilding a case study from a stale deck at 9pm, the fix isn't a faster writer. It's a library that was never designed to be maintained in the first place, because it was designed to build itself.