Sierra-Cedar and Cassidy: Applying AI to Proposal Development

Sierra-Cedar is an Oracle and AWS consulting and managed services partner with over 30 years of client success, helping higher education institutions, public agencies, healthcare organizations, and commercial enterprises implement, modernize, host, and support mission-critical enterprise applications. In April 2025, the firm began a structured effort to evaluate AI tools for targeted use cases. Cassidy was one of the tools selected in that first round, and it was applied to a single, well-defined problem: RFP response and proposal development.

The Problem

Sierra-Cedar's proposal team responds to a steady volume of RFPs, many with short response windows. The constraint was capacity, not capability, and a second constraint was access to institutional knowledge at proposal speed. Opportunities with two-week turnarounds were sometimes declined because the team could not produce a response at the quality level it wanted in the time available, especially when the best source material lived in prior proposals, technical narratives, and the experience of specific consultants.

The team reviewed several tools built specifically for RFP response and proposal management. Based on that review, those tools appeared to be organized primarily around document assembly, template structure, and response tracking. Those were useful capabilities, but they did not address the core constraint. What Sierra-Cedar needed was a way to surface its own institutional knowledge quickly and adapt it, in the firm's own voice and grounded in its prior delivery experience.

What the Team Built through Cassidy

1. A knowledge base that brings together prior proposals and RFP responses, technical service descriptions, SOW language, and other reusable delivery content.
2. A set of AI agents configured for specific proposal types, higher education and public sector responses among them, each drawing from that Sierra-Cedar content base rather than from generic source material.
3. A workflow in which the agent generates an initial draft of the proposal or RFP response using previously approved content, and the proposal team then reviews, edits, verifies, and finalizes the submission.

The team retained review and sign-off on each response. The tool assembled starting material and surfaced relevant institutional knowledge. Consultants wrote the final document.

What Changed

1. First drafts are produced faster, which shortens the time between receiving an RFP and beginning substantive review.
2. Short-deadline opportunities that would previously have been declined are now being pursued.
3. More of the team's time is spent on strategy, positioning, review, and tailoring the response rather than locating, interpreting, and reformatting prior content.
4. Institutional knowledge is easier to scale across proposal work because relevant prior language and delivery examples are available to the team earlier in the process.

"With the same team, our throughput is meaningfully higher. We can respond to RFPs that we previously could not take on because they were due in two weeks," said Michele Ayau, Proposal Team Lead at Sierra-Cedar. "Rather than putting in boilerplate to deliver responses faster, we have more time to improve the quality of the response."

Data Handling

Sierra-Cedar works with higher education institutions, public agencies, and government organizations, and its internal content carries client-specific and commercially sensitive material. Review by the firm's legal and security teams preceded adoption. One important factor was Cassidy's representation that customer content stored in the knowledge base is not used to train external LLM models.

"The knowledge base is what drew us in. We reviewed Cassidy's data handling representations with our legal and security teams, including the representation that our data would not be used to train LLM models, confirmed where our data would be stored, and moved forward on that basis," said Ken Holmes, VP of Technology Solutions at Sierra-Cedar.

Where It Stands

Sierra-Cedar's use of AI has continued to evolve since April 2025. Cassidy's home in that evolution is proposal development, where it has become a core part of how the team operates, not a tool they're still evaluating.

The pattern behind that outcome is worth noting. The proposal team identified a specific, recurring constraint, selected a tool to address it, and kept a person accountable for the output. It was a narrow starting point, and it worked.