Instant project recall
Find technology stacks, operating systems, requirements, implementation approaches, and scope from past work.
Example prompt
“Which projects used Bluetooth Low Energy?”
← All case studies Making project experience
Case study / Private AI & Sales Intelligence
Making project experience
searchable for business development.
How a mid-sized software development company turned its institutional knowledge into a searchable intelligence platform that accelerates proposals and helps its business development teams win more business.
85%
Time saved
Up to, user-reported
2×
Perceived accuracy
Self-reported improvement
2×
Proposal win rate
Early sample, 10 opportunities
Early feedback and preliminary results. The small opportunity sample does not establish causation.
01 / The challenge
Preparing a proposal
took over a week.
A mid-sized software development company had successfully delivered hundreds of applications across numerous industries, technologies, and platforms. Yet accessing that collective experience had become a significant operational challenge.
Critical project knowledge was fragmented across statements of work (SOWs), technical documentation, project records, and the memories of developers, technical leads, and project managers.
For business development teams, preparing a proposal or responding to a Request for Proposal (RFP) took over a week. The resulting materials often lacked the precision and context needed to secure every client.
Despite vast expertise, the company had no reliable way to articulate it to inquiring customers.
What made research so difficult?
Some employees had left the organization. Others could not recall specific implementation details. Technical leads and project managers lacked time to participate in every proposal. Questions about Bluetooth Low Energy, a particular technology stack, or similar client work could take days to resolve.
02 / Our approach
Start with the people
preparing the bids.
Solaeris interviewed bid producers, account managers, technical leads, and other team members to understand how project knowledge was stored, shared, and retrieved.
Those conversations shaped a centralized platform that helps authorized employees discover and connect existing information, regardless of their technical background.
03 / The solution
Search the company’s experience
through a private AI platform.
Solaeris designed and developed a private AI knowledge platform that transforms thousands of internal documents into an intelligent, searchable repository.
Employees can ask questions in natural language and receive contextually relevant answers grounded in the company’s actual project records.
Retrieval augmented generation (RAG) pipeline
Queryable AI search engine
AI-extracted metadata
Statements of workTechnical documentationProject recordsTeam knowledge
Features
Cross-project search
Connect experience across industries, technologies, and project records, even when the same concept is described differently.
Example prompt
“Find projects in this industry approved by this department head.”
AI-powered RFP matching
Upload an RFP and identify the three most relevant past projects, with supporting documentation to start the proposal.
Example prompt
“Which three projects are the closest match to this RFP?”
Source-grounded answers
Follow citations and direct links to SOWs and scoping records to independently verify the answer.
Example prompt
“Show me the project records behind this answer.”
See it in action / Source-grounded answers
From an answer to its evidence.
Follow a citation into the original project record and verify the detail behind the recommendation.
Verify the source: the answer’s citation opens the HarborCare statement of work and highlights the supporting passage about secure patient messaging.
04 / How it works
A private workflow
for searching project records.
Confidential customer information and proprietary technologies stay within the private processing workflow. Public AI services such as ChatGPT, Claude, and Gemini were excluded.
Private processing environment
A / Prepare the knowledge
- Internal documentsSOWs, scope & project records
- On-premises AIExtract & structure information
- Indexed knowledgeVectors, keywords & metadata
B / Answer a question
- User promptA natural-language question or RFP
- Hybrid searchRetrieve relevant indexed evidence
- Private AI modelGenerate an answer from the evidence
- Cited responseAnswer + links to source documents
How documents become searchable
An on-premises AI model extracted and structured information from thousands of documents. The extracted data was normalized, enriched with searchable metadata, and converted into vector representations for semantic retrieval.
Why combine semantic and keyword search?
Semantic retrieval finds conceptual relationships. Deterministic keyword and metadata matching locate precise technical details, such as named technologies or a department head. Hybrid search combines these approaches to retrieve relevant evidence.
How users can verify an answer
The platform passes retrieved evidence to a privately hosted AI model. Responses include citations and direct links to underlying SOWs and scoping records, so users can examine the source material and verify the information.
05 / Early results
Less time spent on
research and proposals.
Users reported less time spent researching previous work and assembling proposal materials, alongside more accurate, contextually relevant responses and proposal content.
Research & proposal preparation
Illustration of the maximum reported time saving
Normalized illustration of “up to 85%” savings; not measured hours or an average across users.
Promising early signals
The reported 2× perceived accuracy reflects users’ own assessments. The 2× proposal win rate is a 100% relative improvement in an initial sample of 10 processed opportunities.
These preliminary results do not establish causation or predict future performance.
Reduced proposal cost
Junior business development staff can research technical capabilities that previously required senior engineers.
Answers on the fly
Account managers can address client questions in real time without waiting to consult the development team.
Experience that stays accessible
Employees can draw on the firm’s documented project history, beyond what any one person can remember.
The bigger opportunity
Make project knowledge
available across the business.
Thoughtful data engineering, private AI infrastructure, and evidence-grounded retrieval made decades of accumulated experience accessible and verifiable. The company’s historical records became an enduring resource for business development.
Explore what AI could do for your business.
Let’s talk