1. Context Gathering
Talk to Jetty in chat, upload process docs, call transcripts, or revenue models. Jetty extracts the business model, current pain points, and target outcomes — and asks for the specific things it needs to design well.
Projects is JetStack AI’s AI-driven workspace for building HubSpot implementations from scratch. Instead of importing assets from a source portal, you describe the business outcome — and an AI agent (Jetty) designs a complete blueprint, generates every supporting asset, and deploys it to your destination portal.
Projects is the right tool when you are starting from a business requirement rather than an existing portal. Common scenarios:
If you already have a working portal you want to replicate or extend, use Implementations instead — that workflow imports the existing configuration and replays it on a destination.
Every project follows the same four phases. You move through them in order, with explicit approval points so the AI never deploys anything you have not reviewed.
1. Context Gathering
Talk to Jetty in chat, upload process docs, call transcripts, or revenue models. Jetty extracts the business model, current pain points, and target outcomes — and asks for the specific things it needs to design well.
2. Blueprint Design
Jetty proposes a blueprint: object families, properties, lifecycle stages, automations, and the outcomes each one serves. You can edit it inline, ask for changes in chat, or approve it as-is to advance.
3. Asset Generation
Once approved, Jetty produces every concrete HubSpot asset the blueprint implies — properties, pipelines, workflows, lists, forms, emails, dashboards, reports, and more. Assets appear in the Assets tab as they are produced.
4. Deployment
Review the generated assets, edit anything you want, and deploy. Work is split between the public HubSpot API (properties, pipelines, workflows…) and the Chrome extension (dashboards, reports, snippets…) so the right surface is used for each asset.
A single project can produce a complete HubSpot implementation across the full asset spectrum. Jetty draws on patterns from a large corpus of real HubSpot configurations, so the assets it produces match real-world conventions rather than idealised samples.
| Category | Assets the project can generate |
|---|---|
| Data model | Custom objects, properties, lifecycle stages, association labels |
| Sales | Deal pipelines, deal stages, sales workflows, lead scoring |
| Marketing | Forms, lists (static and dynamic), marketing emails, email sequences |
| Service | Ticket pipelines, support workflows, playbooks |
| Reporting | Dashboards, reports, CRM cards, preview views, goal templates |
| Productivity | Snippets, task templates |
Each asset is fully editable in the Assets tab before deployment. You can ask Jetty for a revision in plain English (“make this workflow only enrol when source = Webinar”) or open the asset’s detail panel and tweak the spec directly.
Projects has two surfaces:
You can switch between them freely. Both surfaces share the same project state, so progress is live in either UI.
| Projects | Implementations | |
|---|---|---|
| Starting point | A business description or process doc | An existing source portal |
| Driver | AI agent designs the assets | You select assets to import |
| Output | Generated blueprint + assets | Imported asset library |
| Deployment | Same destination flow | Same destination flow |
| Best for | New builds, prototyping, standardising | Replication, migration, cloning |
The two workflows share the same destination deployment infrastructure, so anything that runs reliably in Implementations runs reliably in Projects.
Projects is AI-driven for every phase, so it has different operational characteristics than Implementations. Before starting a project, review: