Assist, a Government Communications initiative

Assist is a tool built by government communicators, for government communicators. Powered by generative AI, Assist is a secure and accessible bespoke conversational tool that empowers users to become even more efficient and effective in their roles through helping to brainstorm, create first drafts and review work. By using Assist, GCS members also benefit from the confidence that its outputs follow GCS policies and standards.

In addition to being able to reference specific GCS documents, responsible use has been the guiding principle of Assist’s development, informed by the GCS Framework for Ethical Innovation and GCS Generative AI Policy. Before using the tool, all GCS members must complete a bespoke ‘AI for Communicators’ training course, designed to upskill and inform them of the safe use of Assist and AI in the workplace. Assist provides users with more than 50 communications-specific ‘pre-built prompts’, which reflect the typical tasks a government communicator might need to do on a daily basis. These prompts span across all seven GCS disciplines, ensuring the tool is tailored to every government communicator use case.

Read the full blog here.

Assist API quickstart

Requirements

• Install make
• Install docker and docker-compose
• Procure an AWS account with Amazon Bedrock.
  • Secure access to Claude 3.7 Sonnet from the region us-west-2 (Assist's current default configuration)
  • Create an IAM user with access to Bedrock, and generate a key for the new user
• Procure an account for 'Insights manager' (previously bugsnag) for logging.

App startup

• Create a .env file from .env.example
  • Populate AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_REGION_NAME, and AWS_DEFAULT_REGION with your Bedrock user's IAM credentials. This will allow your application to access an LLM via Amazon Bedrock.
  • Set the variable IS_DEV=True.
  • Populate AUTH_SECRET_KEY with a secure variable; you send this value with every request to the API for authentication.
  • Populate both OPENSEARCH_INITIAL_ADMIN_PASSWORD and OPENSEARCH_PASSWORD with the same password.
• Run make start from the root directory. This will cause the Docker images to build and launch.
• Visit localhost:5312/docs from your browser to view the available endpoints.

The API is designed to work with the Connect frontend (another service created by Government Communications) though it can be used with other clients.

Generating a response

• Generate an auth-session token by sending a get request to the auth session endpoint with a UUID4 for the user in the header. You can use any valid UUID4 as the user UUID header.
• Generate a new chat with a message by sending a get request to /v1/chats/users/{user_uuid}. The response will contain your completed messaged.