Personalized onboarding
Turn the first bottleneck into an agent.
We will map where Friday should run, identify an approved open model, review your source stack, and convert one verified workflow bottleneck into a narrow, permissioned agent specification.
15 minutes · Video callWhat personalized onboarding covers
Deployment boundary
Choose customer VPC, private cloud, or on-premise infrastructure and document who operates each layer.
Model policy
Select an approved open model, inference runtime, and the controls that keep prompts and outputs in boundary.
Source map
List the systems, accounts, scopes, and minimum permissions needed to understand the chosen process.
Agent guardrails
Define the agent objective, allowed actions, denied actions, audit requirements, and human approval steps.
What you leave with
The first conversation is designed to produce a useful technical and operating outline, not a generic product tour. You leave with a proposed deployment boundary, the sources required for the first brief, a testable definition of the workflow bottleneck, and a narrow agent-pilot specification your security and operations teams can evaluate. The outline also identifies the decisions and technical owners needed to move the pilot forward.
A good first workflow
The best pilot is frequent, measurable, and currently expensive in human coordination. Examples include recurring triage, missing-context follow-up, cross-system status reconciliation, release-readiness checks, or routing work to the correct owner. High-impact or irreversible actions remain behind explicit human approval.
Workspace or privacy questions
Use the same booking link and include the topic in your scheduling note. For privacy requests, include the email address associated with your Friday account so the workspace can be located and the request can be verified.