Datangle
Exploring how a data platform can help people move from a business goal to a working infrastructure setup.
- My focus
- Product direction and hands-on development
- 01Understand the business
- 02Propose a setup
- 03Review the plan
- 04Connect and provision
Why this exists
Setting up data infrastructure involves more than connecting a few services. Someone has to translate a business goal into decisions about what to build, what already exists and how the pieces should fit together.
Datangle is where I am exploring that experience. The onboarding journey is a concrete part of the broader product: helping a customer describe their business, review a proposed setup and work through connecting or provisioning the resources it needs.
What I’m building
The infrastructure agent sits inside Datangle. It brings discovery, planning and execution into a connected journey, with a stored onboarding session so the work has an explicit state.
For example, a customer can describe what they want to achieve before reviewing the proposed resources. The plan becomes something they can examine and revise before execution.
Key decisions
- Begin with the business goal. Infrastructure choices need context; a list of connectors alone does not provide it.
- Make the plan explicit. Keep planning and execution as distinct steps so a proposed setup can be reviewed.
- Build inside the existing product. The agent needs to work with the platform’s integrations and account boundaries.
How it works
The onboarding implementation separates session state, plan derivation, plan revision and step execution. That gives each part a specific responsibility and makes the journey more inspectable than a single open-ended conversation.
My accompanying article follows the process from an initial idea through documentation, design and code, including the points where those representations disagreed.
Evidence and learning
There is an implementation in the Datangle codebase, including onboarding sessions, planning and step execution. That is evidence of a build, not evidence of customer adoption or a measured reduction in setup time.
The main lesson so far is that faster implementation does not remove the need to read the proposal carefully, test the experience and revisit the original goal.
Project evolution
Recent decisions and releases, newest first.
- Implementation
Added onboarding execution and access controls
Connected the onboarding service and deterministic step runner with authorization and a feature flag.
- Learning
Documented the idea-to-implementation loop
Wrote up how the infrastructure-agent idea evolved through documentation, design and code.
What’s next
Evaluate the complete onboarding journey: where people understand the proposed plan, where they need more control, and how well the experience handles a setup that cannot complete in one pass.
Related writing
The Loop Got Faster, Not Different documents the thinking and development process.