Banking analytics modernization

Step 2: Cloud-agnostic modeling

Agent configuration

Snapshot of the agent profile for this stage.

Agent: Modeling Agent

Goal: Derive cloud-agnostic service objects from normalized datasets and docs.

Input: Normalized staging data

Expected output: Canonical services catalog

System prompt guidance: Translate provider-specific SKUs into canonical service types while preserving performance requirements and assumptions.

Available tools: javascript_eval

Current status

Live overview of this step execution.

Step state:in progress

Tools available: 1

Observed tool calls: 3

Activity events: 3

This is a base observability view. In future iterations it can stream real-time logs and tool outputs.

When this step is done, pass its `output` folder content to the next step's `input`.

Step agent chat

Chat with this step's LLM agent to inspect reasoning and outputs before automation is finalized.

Start a conversation with this step agent. Ask it to explain assumptions, mappings, or outputs.

No model response yet.

Tool calls

Each call made (or queued) by the step agent.

Tool	Purpose	Timestamp (UTC)	Status
sql-runner	Aggregate compute/storage/network metrics from staging tables.	2026-03-03T09:25:33Z	completed
entity-classifier	Classify resources into VM, DB, object storage, and networking objects.	2026-03-03T09:31:02Z	running
assumption-tracker	Store unresolved ambiguities and confidence levels.	2026-03-03T09:33:11Z	queued

Agent activity log

Human-readable log trail for this step.

Built temporary SQL views for EC2, RDS, S3 and VPC datasets.

09:24:02

info

Generated 42 cloud-agnostic service candidates.

09:29:40

success

Still resolving GPU profile for 4 VM workloads.

09:32:27

warning