PRISM Ally

Overview

Development stage

DeployedSource: c) Deployed – The use case is being actively authorized or utilized to support the functions or mission of an agency.

High-impact designation

Not high-impactSource: c) Not high-impact

Impact justification

Not reported

Start date

2025-06 (month precision)Source: Jun-25

Withheld from public reporting

NoSource: a) No

Mission

Topic area

Procurement and Financial Management

Operational functions

Search and knowledge retrieval Derived by keyword rules; see methodology

Problem the AI is intended to solve

Assists users with federal regulatory and agency policy questions related to acquisition

Expected benefits

A.I. will help HHS deliver faster, higher-quality public services and measurably improve mission outcomes by cutting cycle times and backlogs, boosting accuracy, and increasing first-contact resolution and customer satisfaction. It will drive cost avoidance and productivity through automation and reuse of shared data/models and code, lowering unit costs per transaction while protecting taxpayer dollars. Built-in accessibility, interpretability, and human-in-the-loop safeguards strengthen equity, fairness, and public trust, with transparent citations, monitoring, and appeal mechanisms. The workforce benefits from targeted upskilling and copilots that reduce manual research and documentation, improving time-to-competency and decision quality. Data quality and interoperability improve via standardized metadata, provenance, and sharing—enabling secure, portable, and interoperable solutions that reduce vendor lock-in and long-term risk. Success will be tracked with concrete metrics such as cycle-time reduction, error-rate and rework decreases, customer experience score gains, cost-per-action savings, accessibility conformance, reuse/adoption counts, training completions, and compliance/incident rates.

System outputs

Ally utilizes a Retrieval-Augmented Generation (RAG) approach to develop answers to user submitted questions. The user submits a query and any prompt instruction needed through the PRISM Ally user interface. Using the query, PRISM Ally performs a vector search of its private knowledge repository to identify relevant information that can provide enhanced context for developing the answer. The user query, prompt, and enhanced context are then passed to the LLM. The LLM considers the information and returns an answer. The utilization of enhanced context provides guardrails for the LLM and helps to increase the accuracy of the answers provided.

Technology

AI classification

Generative AISource: Generative AI: AI that generates new or synthetic content (e.g., images, videos, audio, text, code).

System name(s)

Not reported

Custom-developed code

No

Public source code

Not reported

Sourcing

How it was built

Purchased from vendorSource: a) Purchased from a vendor

Vendor (as reported)

Unison

Vendors (standardized)

None on the standard list

A vendor is the supplier named by the agency. It does not identify the underlying model or AI technology.

Data and privacy

Involves PII

No

Privacy Impact Assessment

Not reported

Authorization to Operate

No

Demographic features

k) None of the above

Training and evaluation data

The PRISM Ally application and private knowledge repository are located within the Unison Cloud, in a FedRAMP moderate environment. The application and repository are maintained by Unison. Regulatory content (e.g. FAR, DFARS, agency supplementals) within the repository is sourced from government authenticated sources (e.g. acquisition.gov, ecfr.gov). Unison updates the regulatory content within the repository with each new regulatory update release.

Federal Data Catalog

Not reported

Governance

0 of 8 minimum-practice questions answered. A blank answer means the agency reported nothing; it does not mean the practice is absent.

Potential impacts and how they were identified

Not reported