Skip to Main Content
Concourse & Jira Cloud Portal
Status PM Refinement
Created by Katie Cousino
Created on Aug 12, 2025

AI-Guided Solution Plan Builder for First Four Layers - Solution Components

Simply Put

Enable Solution Architects to generate high-quality, first-draft content for the early stages of the solution plan using AI — reducing manual effort, improving consistency, and accelerating pursuit timelines.




Problem / Pain Point

Teams spend excessive time and effort piecing together the Vision → Capabilities → Architecture → Dev-object chain for each new solution. Inputs arrive in many formats (RFPs, spreadsheets, emails), which makes it hard to keep the plan complete, consistent, and easy to iterate.




🧩 Goal Overview

Enable Solution Architects to generate high-quality, first-draft content for the early stages of the solution plan using AI — reducing manual effort, improving consistency, and accelerating pursuit timelines.


Key Requirements

1. Modular GPT or AI Agent Design or Embedded AI feature

  • Each layer should be generated by a dedicated module with clear logic.

  • Outputs should be editable and exportable (e.g., into slides, Jira, Concourse).

2. Prompt Templates by Layer

  • Defined prompt structures per layer (e.g., “Given this RFP, generate a vision statement with industry relevance”).

  • Reusable and customizable for each offering and practice. Ability to create at a generic level for most practices.

3. Structured Outputs

  • Content should be returned in structured formats (e.g., bullets, tables, slide-ready format), including excel as currently works for Oracle

4. User Controls

  • Allow for sizing of the opportunity

  • Support for tailoring based on client maturity, region, or solution scope.


📥 Key Inputs

Input Type

Description

RFP or Proposal Request

Source document used to extract themes, scope, and goals.

Client Info

Industry, geography, public filings (e.g., 10-K), size, maturity.

Pursuit Metadata

Opportunity type, offering, team roles, size, client tier.

Prebuilt Templates

Vision statement formats, architecture layouts, benefits frameworks.

Past Pursuits / Proposals

Structured examples tagged by offering, industry, and plan layer.


📤 Expected Outputs (By Layer)

Layer

AI-Generated Outputs

1. Vision

Executive summary-style narrative tying client goals to PwC solution vision; aligned with sector trends and strategy insights.

2. Capabilities → Benefits

A table or bullets mapping business capabilities to quantified benefits and KPIs; tailored to industry and offering.

3. Architecture

Logical or technical diagrams (or scaffold text) outlining future state architecture, integration points, and data flows.

4. Dev Objects / Components

A list of solution components (e.g., RICEFW objects, user stories, config elements) based on project type and scope.



Metrics:

  1. General use of the tool = X use per tool

  2. Copy or download of suggestion provided

Goal (What do we want to learn?)

Signal (What shows it’s happening?)

Metric (How will we measure it?)

Are teams assembling proposal layers faster?

Elapsed time from first layer created → final layer saved (event timestamps in Concourse)

Average “proposal-layer build time” (mins)Target: −30 % vs. pre-launch baseline

Are users pulling content from the repository instead of writing from scratch?

“Insert from repository” or AI-assist events vs. manual text entry per layer

% of layers populated via repository/AITarget: ≥ 70 %

Is template reuse driving consistency?

Distinct template IDs reused across proposals

Template-reuse rate (unique template uses ÷ total layers)Target: ≥ 70 %

Is overall proposal throughput increasing?

Completed strategic solution plans per week

Plans completed per weekTarget: +25 % over baseline

Are users satisfied with AI-assisted layer creation?

In-app thumbs-up / thumbs-down after each assist

Positive feedback rateTarget: ≥ 85 %