Want us to set it up for you?
Three phases. Ten days.
AI StackWorks is a product; the Kickstart is the optional managed path. In three phases we understand your workflow, wire your agents into your board, and don't leave until a real feature has shipped.
Workflow discovery
Before we touch a config, we map how work moves through your team today: where it starts, where it drops, and which handoffs break. On this page it sounds simple; in practice it's the difference between an agent setup that lands and one that gets ignored.
Deliverables
- 90-minute intake call with your eng lead
- Current-workflow map (states, owners, typical work shape)
- Pilot-feature selection: one real feature the agents will ship
- Bottleneck audit: the two or three handoffs that drop the most context
- Written scope of what the agents will cover in Phase 2
Expected Outcomes
- A shared picture of how your team actually works today
- An agreed pilot feature the agents will pick up on Day 3
- No surprises about scope when setup starts
Agent setup
Your agents go live, one role at a time. PM agent first, so intake and spec stop being the bottleneck. Engineering agent next, scaffolding and PRs on your real repo. QA agent last, running your tests and posting verdicts in Slack. Every handoff visible on your board, like human teammates.
Deliverables
- Day 3: first agent live (PM or Engineering) on your board
- Day 7: second agent wired with role-based handoffs
- Day 10: QA agent, Slack notifications, approval gates
- Agent configs committed to your repo. You own them
- Runbook + recorded 60-minute training session
Expected Outcomes
- Four agent roles running on your actual workflow, handoffs you can see
- Every handoff visible on your board and in Slack
- Your team knows how to maintain the agents without us
Work complete
The pilot feature we picked on Day 1 moves from intake to merged PR. Through your agents, on your board, visible in Slack. Humans stay in the loop at every gate that matters. We don't leave until at least one feature has fully shipped, with its execution record intact.
Deliverables
- Pilot feature moved intake → merged PR by your agents
- Full execution record for the feature: plan, approvals, tests, PR, docs
- Post-handover review: what worked, what to tune next
- 30-day Slack channel with us for quick questions
Expected Outcomes
- Proof, on your board, that your agents ship — with a full record
- A playbook for the next ten features without our help
- A clean handover and a team that can drive the agents on its own
What's included in the Kickstart
The $5,000 Kickstart is a fixed-scope engagement. Standard integrations and the four standard agent roles are included. Anything new or custom to your organization we scope and quote separately. Think new tools, new roles, or proprietary APIs.
Included (standard)
- Jira, Linear, or GitHub Projects as the board
- Slack notifications + approvals
- Knowledge base reads/writes on Confluence, Notion, or Linear docs
- GitHub or GitLab for PR creation and review
- Claude, GPT, or other major LLMs already in your org
- Four standard agent roles: PM, Engineering, QA, Release
- Up to three workflow state transitions mapped end-to-end
Extra (scoped separately)
- New integrations with internal / proprietary tools
- Custom agent roles beyond the standard four
- Migration from another board
- On-prem or air-gapped LLM deployment
- Training data curation beyond your existing knowledge base
Works with your existing stack.
The Kickstart runs through the tools your team already uses. No migration required.