AI-Powered Software Development and Modernization

Request for Information (RFI) for AI-Powered Software Development and Modernization

Notice ID: AF-A1-RFI-AI-DEV-2025

Issued by: Department of the Air Force (DAF)

Date of Issuance: 2025-12-18

Response Deadline: 2026-01-09

1.0 Introduction and Purpose

The Air Force Deputy Chief of Staff for Manpower, Personnel, and Services (AF/A1) under the Department of the Air Force (DAF) is issuing this Request for Information (RFI) to conduct market research and identify sources with the capability to provide advanced solutions in Artificial Intelligence (AI)-powered software development. This RFI seeks to understand the current and emerging capabilities within the industry for AI-assisted and fully autonomous coding practices for both new development and the refactoring of existing enterprise applications.

The focus of this inquiry is on the application of AI to accelerate development, improve code quality, and enhance the maintainability of software built on or using the following platforms:

• Salesforce (including Public Sector Solutions)
• ServiceNow
• Outsystems
• Angular
• PowerBI

This is an RFI only. This RFI is issued solely for information and planning purposes and does not constitute a solicitation. A contract will not be awarded based on this announcement. All information received in response to this RFI that is marked as proprietary will be handled accordingly.

2.0 Background

The HAF/A1 manages a diverse portfolio of applications critical to our mission. A significant portion of these applications are built on platforms such as Salesforce, ServiceNow, and Outsystems, with front-end interfaces often developed using Angular. Our objective is to modernize these systems and accelerate the delivery of new capabilities.

We recognize that traditional software development methods, while proven, are often resource-intensive. Recent advancements in AI, such as AI-assisted coding (often termed "vibe coding") and the potential for autonomous AI agents, present an opportunity to revolutionize our software development lifecycle. We are seeking to understand what is within the "realm of the possible" for leveraging these technologies to meet our development and refactoring needs.

3.0 Scope of Interest & Required Information

The HAF/A1 is interested in understanding the industry's capabilities in the following areas. Respondents are requested to provide detailed information addressing the questions below.

3.1 AI-Assisted and Autonomous Coding Capabilities

1. General Approach: Describe your company's experience and approach to using AI in software development. Please differentiate between AI-assisted coding (where AI acts as a co-pilot to a human developer) and fully autonomous coding (where an AI agent can take a set of requirements and produce a functional application with minimal human intervention). How does your approach in AI use shift under complex architecture?
2. AI-Assisted Coding: What is your company's experience with this development paradigm? What tools and methodologies do you employ? What areas of success and limitations have you experienced?
3. Autonomous Agents: Do you have capabilities for fully autonomous AI agents that can perform end-to-end software development tasks, including requirement analysis, coding, testing, and deployment? If so, please provide examples of:
   • How you handle monitoring and oversight.
   • Describe the technology and provide examples of projects where this has been successfully implemented.
4. Platform Experience: Detail your experience in using AI to develop and/or refactor applications on the following platforms. Please provide specific examples for each.
   • Salesforce: Experience with Apex, Lightning Web Components (LWC), and refactoring into the Public Sector Solutions (PSS) model.
   • ServiceNow: Experience with ServiceNow's platform development, including custom applications and workflows.
   • Outsystems: Experience with AI-driven development or refactoring of applications within the Outsystems low-code platform.
   • Angular: Experience with generating, testing, and refactoring Angular components and applications using AI.
   • PowerBI: Experience with AI assistance in PowerBI, such as DAX language to perform basic visualizations.

3.2 Tools and Technologies

1. AI Models and Tools: What specific AI models (e.g., GPT-4, Claude 3.5, Gemini) and AI coding assistant tools (e.g., GitHub Copilot, Cursor, Cody, Tabnine) does your company utilize and how do you incorporate use specifically with PII/PHI data?
2. Proprietary vs. COTS: Do you utilize commercial off-the-shelf (COTS) AI tools, or have you developed proprietary AI models and platforms for software development? Please describe your technology stack.
3. Low-Code/No-Code Integration: How do your AI capabilities integrate with low-code/no-code platforms like Outsystems or the declarative aspects of Salesforce and ServiceNow?

3.3 Quality and Maintainability

1. Code Quality: How do you ensure the quality, reliability, and maintainability of AI-generated code? Describe your processes for code review, testing, validation, and monitoring of the new/modified workflow. 
2. Intellectual Property: How do you address intellectual property, licensing, and ethical concerns related to code generated from models trained on public repositories? 
3. Debugging and Refactoring: Describe your process for debugging and refactoring complex, AI-generated codebases. How do you manage technical debt in an AI-driven development environment? 

4.0 Security, Compliance, and Network Integration

1. DoD Network Experience: Please list any AI development tools (e.g., AI code assistants, model APIs) that you have experience using within a DoD or DAF network environment. For each tool, please specify if it has an existing Authority to Operate (ATO), at what Impact Level (e.g., IL2, IL4, IL5, IL6), and how you incorporate use specifically with PII/PHI data.
2. Accreditation Strategy: Describe your experience and methodology for achieving an ATO for software development tools and AI models within the DoD Risk Management Framework (RMF). What is your strategy for accrediting solutions that rely on commercial cloud-based AI models?
3. Data Security: How does your solution ensure the protection of Controlled Unclassified Information (CUI) when utilizing AI models? Address data residency, data-in-transit, and data-at-rest encryption, particularly when interacting with commercial AI services. 
4. On-Premise & Air-Gapped Capabilities: Does your proposed AI development solution have the capability to operate in a fully on-premise or air-gapped government cloud environment (e.g., an IL5 environment)? If so, please describe the architecture and any limitations.
5. AI Model Security: What is your process for vetting and securing the AI models used in your development workflow? How do you mitigate risks such as AI model poisoning, data leakage, other supply chain vulnerabilities inherent in pre-trained models, and ensure resiliency to adversarial attacks? 

5.0 General Questions

1. Productivity and ROI: How do you measure the return on investment (ROI) of using AI coding assistants? Provide any available metrics or case studies that demonstrate productivity gains (e.g., reduction in development time, increase in features delivered). 
2. Impact on Code Quality: What is the quantifiable impact of AI-assisted development on code quality? How do you measure metrics like bug density, code complexity, and maintainability before and after AI adoption? 
3. Developer Experience and Skill Shift: How does the adoption of AI tools change the role of a developer? 
4. Adoption and Integration: What are the biggest challenges organizations face when integrating AI tools into their existing software development lifecycle (SDLC) and workflows? How do you recommend overcoming them? 
5. Governance and Accountability: In a development process that includes AI-generated code, how is accountability maintained? How do you monitor the new/modified workflow? Describe the governance frameworks you recommend for reviewing, validating, and taking ownership of AI-produced code. 

6.0 Response Submission

Interested parties are requested to submit a response of no more than 15 pages, addressing the questions outlined in Sections 3.0, 4.0 and 5.0. Responses should be submitted electronically in Microsoft Word or Adobe PDF format to melissa.mattis@us.af.mil and corey.collins.12@us.af.mil no later than 17:00 EST on 2026-01-09.

Please include the following in your response:

• Company Name, Address, and Point of Contact
• CAGE Code and DUNS Number
• Business Size and any applicable socio-economic classifications.
• Recommended NAICS and PSC

7.0 Disclaimers

• This RFI is for market research purposes only and does not constitute a commitment, implied or otherwise, that the Department of the Air Force will take procurement action in this matter.
• The HAF/A1 will not be responsible for any costs incurred by any entity in furnishing this information.
• Responses to this RFI are not offers and cannot be accepted by the Government to form a binding contract.
• All information submitted in response to this announcement is voluntary; the HAF/A1 will not pay for information requested nor will it compensate any respondent for any cost incurred in developing information for this RFI.