AI RFI_Q and A (Part 1).xlsx
XLSX spreadsheet 17 KB Posted
- Solicitation number
- Not on record
About this file
This document is a Request for Information (RFI) from the Air Force Life Cycle Management Center (AFLCMC) seeking market research on artificial intelligence (AI) capabilities for the Department of the Air Force (DAF). The RFI seeks an AI solution that can integrate with Microsoft Office 365 productivity software, providing generative AI capabilities across platforms like Outlook, Teams, and SharePoint. Key requirements include AI-assisted email summarization and draft generation, document creation using structured and unstructured data, meeting transcript summarization, prompt monitoring with SIEM/SOAR alert capabilities, and the ability to operate in an IL4/IL5 secure environment.
Vendors are requested to submit a 10-page Microsoft Word response by 1500 Eastern Time on 30 June 2025, detailing their AI solution's capabilities. The RFI emphasizes the need for comprehensive pricing structures, data control, continuous improvement, and integration with access control systems like Microsoft Purview. Importantly, this is a market research exercise and does not guarantee a future solicitation. The government is open to various AI solutions that can demonstrate potential for IL5 compliance, with the only current restriction being that DeepSeek models are not acceptable. Responses will be reviewed by military personnel, government civilians, and specified government support contractors.
View the file
Other files
| File | Type | Posted |
|---|---|---|
| AI RFI_Q and A (Part 2).xlsx | XLSX spreadsheet | |
| Q and A Template.xlsx | XLSX spreadsheet |
On GovTribe
Work with this file
- Download the original file
- Contacts named in this file
- Similar government files
- Ask GovTribe AI about this file
Text version
Sheet1
| ID | Questions | Answers |
| 1 | Does the submission cover page count towards the 10 page limit? | No. |
| 2 | By when does the AFLCMC intend to issue a solicitation for this initiative? | AFLCMC/HNI does not yet have a timeline of a solicitation to support this type of requirement. This RFI is purely market research to better understand the current marketplace. |
| 3 | By when does the AFLCMC wish to have an operational system? | AFLCMC/HNI does not yet have a timeline for an operational system for this type of requirement. This RFI is purely market research to better understand the current marketplace. |
| 4 | What is the AFLCMC's timeline for implementing this system at an enterprise level? | AFLCMC/HNI does not yet have a requirement for implementing this type of requirement at an enterprise level. This RFI is purely market research to better understand the current marketplace. |
| 5 | What does the AFLCMC attatch value to when considering responses? | This RFI is purely for market research and no specific value for evaluation purposes will be utilized. AFLCMC/HNI is interested to learn about the capabilities that exist in the marketplace to support the scope included within the RFI. |
| 6 | Would you kindly confirm if this is an entirely new effort or is there an incumbent currently providing these services? If so, would you please provide the incumbent contract details? | This RFI is purely for market research purposes of a new requirement. |
| 7 | What will be the contract value for this opportunity? | This RFI is purely for market research; no contract opportunity is currently planned as of this time. |
| 8 | Are there any set-aside considerations for this opportunity? | This RFI is purely market research to better understand the current marketplace. All vendors are welcome to submit a response to the RFI to support decision making on a potential future solicitation. |
| 9 | Is there an existing contract or tool in place for this requirement? | No. |
| 10 | Is this RFI a continuation of a previously released RFI? Or is this a new Requirment? | This is not a continuation of a previously released RFI. This RFI is purely for market research purposes. |
| 11 | Integration & UX: Does the Government envision the solution being primarily accessed through existing Microsoft 365 interfaces (e.g., Outlook, Teams, Word add-ins), a standalone portal, or a combination of both depending on user role? | The Government is looking for a combination of both: A central portal for aggregating data and issuing responses while using in app interfaces for application specific functions. |
| 12 | Architecture & Modularity: Would the Government support a modular, agent-based architecture approach, where specific AI capabilities (e.g., meeting summarization, spreadsheet analysis) can be deployed incrementally or scoped by mission need? | The Government would support a modular, agent-based architecture approach. |
| 13 | Security & SIEM: For prompt monitoring and SIEM/SOAR integration, does the Government prefer alerts in a specific format or schema (e.g., CEF, JSON, STIX/TAXII), and is integration with tools like Microsoft Sentinel or Splunk anticipated? | For the purposes of this RFI response, the Government is interested in integration with tools such as ELICSAR and Microsoft Sentinel. JSON would be a compatible format for alerts. |
| 14 | IL4/IL5 Readiness: Can vendors who are actively pursuing IL4/IL5 readiness — with a defined roadmap, security controls, and past FedRAMP experience — be considered viable, or is current ATO/authorization a strict requirement? | For the purposes of this RFI response, the Government is interested in any current or demonstrated potential capability to obtain and maintain an ATO. |
| 15 | Functional Priorities: Are there specific functional areas (e.g., email summarization, documentation drafting, spreadsheet analysis) that the Government sees as highest priority for near-term prototyping or phased deployment? | For the purposes of this RFI response, the Government is interested in the functional areas mentioned as these are sought by AF users. |
| 16 | Will services be required for development and integration of this requirement? | This RFI is purely for market research purposes to understand the current marketplace and a requirement for services for development and integration has not yet been determined. |
| 17 | If services will be required will personnel be onsite or remote? | This RFI is purely for market research purposes to understand the current marketplace and a requirement for services onsite or remote has not yet been determined. |
| 18 | Will services personnel require clearances and base access to assist with this requirement post award? | This RFI is purely for market research purposes to understand the current marketplace and a requirement for services personnel has not yet been determined. |
| 19 | Will this solution be deployed on existing infrastructure or is the USAF interested in an infrastructure offering to support this requirement? | The Government is open to infrastructure offerings that are IL5 and compatible with AFNET and military dependent O365 tenants. |
| 20 | Will the proposed solution be deployed on NIPR , SIPR and potentially AF JWICS or reside only at an unclassified level to afford a solution such as copilot the opportunity to pull data from external resources? | The proposed solution will be on NIPR. |
| 21 | Integration Expectations with Microsoft Ecosystem: Could the government clarify whether solutions must natively integrate with all listed Microsoft applications (e.g., Office 365, Outlook, Teams, SharePoint), or if modular integration (e.g., Teams only) will be considered sufficient? | This RFI is purely for market research purposes to understand the current marketplace. Information about a native integration or modular approach may be provided within the RFI response to assist the Government with understanding the pros/cons of each model. |
| 22 | Structured and Unstructured Data: Can the government define or provide examples of the types of “structured and unstructured data” currently used/stored within local and cloud systems that the AI solution should reference for document generation? | Structured and unstructured data includes emails, documents such as Word, PDF, presentations, and spreadsheets. |
| 23 | Meeting Transcripts and Chat Summarization: What platforms are used to generate these transcripts (e.g., Teams only)? Will contractors be expected to provide transcription services or only summarization? | Currently, Teams is the primarily supported platform for virtual meetings to generate transcripts. The contractor can provide summarization of the transcripts generated. |
| 24 | Alerting Capabilities for Monitoring User Prompts: What are the government’s expectations for granularity and real-time monitoring in prompt alerting? Is there a list of preferred SIEM/SOAR platforms for integration? | The government expects prompt alerting to follow existing policy. For the purposes of this RFI response, the Government is interested in integration with tools such as ELICSAR and Microsoft Sentinel. |
| 25 | Hosting and Security Requirements: Are there preferred cloud service providers (e.g., AWS GovCloud, Azure Government) for hosting in IL4/IL5 environments? If a vendor does not currently have an ATO, will a path-to-ATO approach be acceptable, and should it be detailed in the response? | The Government is open to CSP solutions that are IL5 and compatible with AFNET and military dependent O365 tenants. For the purposes of this RFI response, the Government is interested in any current or demonstrated potential capability to obtain and maintain an ATO. |
| 26 | Comprehensive Pricing Structure: Will the government provide a notional user base (e.g., # of users, access frequency) to help tailor fixed vs. variable cost models? Are usage-based pricing models (e.g., token consumption for LLMs) acceptable for consideration, or does the Government have a preference for licensing? | The Government does not have a preference at this time, as this RFI is purely for market research purposes to better understand the current marketplace. Information about differing pricing models may be provided within the RFI response to assist the Government with understanding the pros/cons of each model. |
| 27 | Data Control and LLM Training Concerns: Will the Government require vendors to demonstrate how user prompts and data will be isolated from public model training corpora, and if so, what level of technical detail is preferred? | This RFI is purely for market research and no demonstrations will be requested at the time of RFI response submittal. |
| 28 | Access Control Integration: Does the Government currently use Microsoft Purview extensively, or are vendors expected to offer integration plans as part of their AI solution’s data protection strategy? | The Solution would need to work with Purview's access controls to ensure data security. |
| 29 | Industry Day or Demonstration Opportunities: Does the Government intend to host an industry day, virtual Q&A session, or request technical demos prior to releasing a formal RFP? | AFLCMC/HNI does not yet have plans for an industry day or potential future solicitation. This RFI is purely market research to better understand the current marketplace. |
| 30 | Could the Government provide insight into the primary metrics for success? For example, will the program's success be measured by user adoption rates, specific time savings on administrative tasks, or other key performance indicators (KPIs)? | This RFI is purely for market research and no specific metrics for evaluation purposes will be utilized. AFLCMC/HNI is interested to learn about the capabilities that exist in the marketplace to support the scope included within the RFI. |
| 31 | Can the Government provide a rough order of magnitude for the anticipated number of users during a potential pilot phase and a potential full-scale deployment? | AFLCMC/HNI does not yet have a requirement for a potential pilot or full-scale deployment. This RFI is purely market research to better understand the current marketplace. |
| 32 | Does the Government have a preference for, or any restrictions against, specific Large Language Models (LLMs)? Furthermore, will the Government require the ability to update or swap the underlying LLM as technology evolves, or will that be managed by the solution provider? | There is no preference for an LLM outside of blacklisted LLMs ex. DeepSeek. There is no preference for whether the Government or the solution provider maintains the LLM as long as the LLM is updated in a timely fashion. |
| 33 | The RFI requires the ability to export alerts to existing SIEM/SOAR solutions. Can the Government identify the primary SIEM/SOAR platforms currently in use within the target environment (e.g., Splunk, Microsoft Sentinel)? | For the purposes of this RFI response, the Government is interested in integration with tools such as ELICSAR and Microsoft Sentinel. |
| 34 | Does the Government have a preference between predictable, fixed-cost models (e.g., per-user, per-year licensing) versus variable, consumption-based models (e.g., token-based pricing) to align with DAF budgeting processes? | The Government does not have a preference at this time, as this RFI is purely for market research purposes to better understand the current marketplace. Information about differing pricing models may be provided within the RFI response to assist the Government with understanding the pros/cons of each model. |
| 35 | Does the Government expect AI to process content only within the email clients (e.g., Outlook), or can email content be redirected to secure cloud repositories like Box for processing and compliance enforcement? | The Government expects email content can be redirected as long as data remains within an IL5 environment and USAF maintains control of the data at all points. |
| 36 | Can you specify which SIEM or SOAR platforms are currently deployed or preferred for receiving AI-generated prompt or policy violation alerts? | For the purposes of this RFI response, the Government is interested in integration with tools such as ELICSAR and Microsoft Sentinel. |
| 37 | How does the Government define “suspicious prompt patterns” and what thresholds or rules would trigger alert generation? | Suspicious prompt patterns is defined as prompts that attempt to bypass security controls either imposed on the LLM or access controls such as exfiltrating data from documents they don't have permissions on. Thresholds will be dynamic. |
| 38 | Can AI tools access stored email content for summarization and insight generation, or must AI actions occur only in real-time or in-transit workflows? | Yes, the AI tool can access stored email content for summarization and insight generation. |
| 39 | Will DAF mandate integration with Microsoft Purview for labeling and classification, or can equivalent (we would propose Box Shield) policies be used to enforce access control and data residency? | The solution would need to work with existing access controls. |
| 40 | Is the requirement for draft email generation tied strictly to Microsoft Outlook, or can external AI tools suggest drafts stored in Box folders for later user use in Outlook? | There is no current requirement, but a goal for email draft generation is to create real-time drafts for the needs of the user. Referencing templates may not have the flexibility required without direct integration to Outlook. |
| 41 | Must all user data (emails, prompts, AI outputs) remain within a specific Government enclave, or are FedRAMP High commercial environments (e.g., Box Zones, Box Shield) acceptable if no training occurs on user data? | All user data must remain within an IL5 environment. |
| 42 | Can AI-generated summaries of Teams/Outlook meeting transcripts be stored in Box (fed ramp high) for downstream review and compliance logging? | No |
| 43 | What level of external API interaction is expected or permitted—are event-based webhooks, metadata tagging APIs, or scheduled script access acceptable for integrations with collaboration platforms like Box? | API connections are limited to IL5 gov cloud. |
| 44 | Are you seeking a full-stack solution with an LLM included, or expecting vendors to connect to an existing/approved LLM service? | There is no current requirement, but the Government is interested in what a full-stack solution could include. |
| 45 | Are any LLM providers (e.g., Azure OpenAI, Anthropic, AWS Bedrock) already approved or in use within your IL4/IL5 environment? | For the purposes of this RFI response, the Government is interested in any current or demonstrated potential capability for LLM providers to be used within an IL5 environment. |
| 46 | Do you require on-prem or enclave-hosted LLMs, or will cloud-based Gov-authorized endpoints (e.g., Azure OpenAI IL5) suffice? | For the purposes of this RFI, Gov Authorized IL5 endpoints would be sufficient. |
| 47 | Will the solution be expected to transcribe audio from meetings (e.g., Teams calls), or will it operate on already-transcribed text provided by existing tools? | For the purposes of this RFI, DAF365 Teams has the option to transcribe text and export for the format already. If the tool can use that, it could be enough to suffice. |
| 48 | If transcription is required, is there an existing capability (e.g., Microsoft Teams transcription, Azure Speech Services) that the AI solution should consume and analyze? | Microsoft Teams offers transcripts of text that the AI solution can analyze. |
| 49 | Which DAF user groups (e.g., FM, S1, cyber, acquisition) will use the AI assistant? | There is no current requirement, but a AI Assistant could be utilized by various DAF user groups. |
| 50 | What types of documents or emails are expected to benefit most from AI summarization or generation (e.g., taskers, SITREPs, travel requests, awards, evaluations)? | The Government does not have a preference at this time for documents or emails that will benefit most from AI summarization. |
| 51 | Are there common follow-on steps after reviewing these documents—such as routing, approvals, or data entry—that could also be automated? | Potentially, but the focus of this RFI is not on the automation of processes. |
| 52 | Do users often copy/paste from SharePoint, Teams, or Outlook into other systems? If so, which systems? | For the purposes of this RFI, we do not have a list of other systems at this time. |
| 53 | Are any of these document processes governed by templates, policies, or workflows that need to be enforced or standardized? | For the purposes of this RFI, we do not have a list of what these other documents could be at this time. |
| 54 | Would users benefit from having an AI assistant that not only generates content but also suggests next steps (e.g., notify supervisor, attach file, start form)? | For the purposes of this RFI, the Government would be interested in this capability. |
| 55 | Are you targetting specifically Copilot, Power Automation and Power App based solutions, or other RPA and LLM Inference platforms as well? | The Government is conducting Market Research to identify what LLM Inference platforms exist in the marketplace, not targeting a specific tool. |
| 56 | Can you provide some details on the existing SIEM/SOAR solutions already in use? | For the purposes of this RFI response, the Government is interested in integration with tools such as ELICSAR and Microsoft Sentinel. |
| 57 | For API and third party expandability, do you have any specific target platforms in mind? | The potential goal is to allow for the integration of other components or to allow the solution to connect to other applications in an expanded format. This might also include having Government developed solutions that reference back to the provided solution via the API to use some of its capabilities. |
| 58 | Is there is perferred LLM Inference platform (Azure Open AI, Ask Sage, etc)? | The Government does not have a preference for LLM inference platforms. The Government is interested in platforms that are IL5 or can become IL5 compliant, and compatible with AFNET and military department O365 tenants. |
| 59 | Are there any government-imposed constraints or preferences regarding the type of AI model(s) or specific vendors that may be used for this requirement? | The Government has the constraint that DeepSeek models may not be used for this requirement, but for the purposes of this RFI is interested in all models and capabilities currently exist within the market place. |
File details come from the government source that posted it. Updated .