A team needs a document-answering assistant, with as little custom work as possible. What fits?
Why C is the answer
Agent Bricks Knowledge Assistant
Knowledge Assistant is the Agent Bricks option for answering from documents, so a team needs little custom work.
In detail
- Knowledge Assistant is the Agent Bricks option for answering from documents.
- It suits a team that wants as little custom work as possible.
- Databricks now labels Knowledge Assistant as legacy, and points new low-code agents to Genie Agents.
- Start with a prebuilt option, and customize only if it falls short.
Use case
An illustrative scenario, not a real incident.
An HR team wants staff to ask about its handbook, and it has no engineers free. Knowledge Assistant in Agent Bricks answers from the documents with little custom work.
See it in different ways
flowchart TD
A{"What is the goal?"} -->|"Answer from documents"| B["Knowledge Assistant"]
A -->|"Coordinate several agents"| C["Multiagent Supervisor"]
A -->|"Pull fields from text"| D["Information Extraction"]mindmap
root((Agent Bricks))
Knowledge Assistant
Questions on documents
Information Extraction
Fields from text
Multiagent Supervisor
Coordinates agentsmindmap
root(("Agent Bricks"))
n1["Answer from documents"]
n2["Knowledge Assistant"]
n3["Pull fields from text"]
n4["Information Extraction"]
n5["Coordinate agents"]
n6["Multiagent Supervisor"]Key point: Knowledge Assistant answers questions from documents.
Why the other options are wrong
- A. A fine-tuned model that is retrained on the documents each week
- Weekly fine-tuning is costly, and answers still lag behind the documents.
- B. A hand-written set of rules for each likely question
- Hand-written rules cannot keep up with new questions.
- D. A custom retrieval chain written from scratch in a notebook
- A custom chain can work, but it is the opposite of little custom work.