# R&D & Product (/use-cases/rd-product)



Product and engineering teams use Helio to go from a problem statement to a shipped, monitored feature without switching tools for every step. The same AI teammates that write code and review PRs also draft the PRD, publish a page as an Artifact, and patrol every deploy after it goes out.

<LivePreview component="TaskBoard" />

## What an AI product and engineering teammate does [#what-an-ai-product-and-engineering-teammate-does]

* **Writes code and reviews PRs**: takes a feature spec or bug description and produces working code; reads diffs and leaves structured, specific feedback
* **Turns a problem into a PRD**: asks the clarifying questions a spec needs before writing starts, then produces a document the team reviews and edits together
* **Publishes a page as an Artifact**: turns a pitch into a landing page, published as a stable, org-shareable link you keep iterating on
* **Patrols every deploy**: confirms a release is actually live, not just reported as shipped, and flags uptime or certificate issues before you notice
* **Runs as a full product team**: hire product, research, design, code-development, and code-review experts into one channel and let them hand work to each other in sequence or in parallel

## A typical workflow [#a-typical-workflow]

**You have a feature to build.** Describe it in your #engineering channel, assign it to your AI engineer, and they'll start working in the channel. They'll write the code, run tests, and open a PR for review.

**You're kicking off a new project.** Give your PM teammate the problem, not the solution, and they ask what's missing before writing starts. Once the PRD is ready, hire the rest of the team into the same channel and let the PM break the work into pieces the others can pick up.

**You need a page to share internally, fast.** Give any teammate the pitch, and they draft the copy, build the page, and publish it as an Artifact, a stable link anyone in your org can open. After every deploy, a separate automation confirms the new version is actually served before you find out the hard way.

## Recommended setup [#recommended-setup]

| Teammate              | Template                                                                                                             | Key skills to add         |
| --------------------- | -------------------------------------------------------------------------------------------------------------------- | ------------------------- |
| Engineering           | Code Development Expert                                                                                              | GitHub, Web Search by Exa |
| Code review & quality | Full-Stack Code Review Expert                                                                                        | GitHub                    |
| Product               | Product Management Expert                                                                                            | Notion                    |
| Full product team     | Product Management Expert, Deep Research Writer, UI Designer, Code Development Expert, Full-Stack Code Review Expert | One per role              |

Browse the full catalog in [Skills & Plugins →](/connect/skills-and-plugins).

**Trust level suggestion:** Standard for most engineering work. You'll want to approve actions like merging PRs or pushing to main.

## Getting started [#getting-started]

<Steps>
  <Step>
    Hire a **Code Development Expert** teammate using the template.
  </Step>

  <Step>
    Connect **GitHub** so they have access to your repos.
  </Step>

  <Step>
    Add them to your **#engineering*&#x2A; or **#development** channel.
  </Step>

  <Step>
    Assign your first task: "Write tests for the auth module" or "Fix the bug in issue #42."
  </Step>
</Steps>

Starting a project with more than one role from day one? See [Organize your own tech team](/use-cases/rd-product/ai-team) instead.

## How these work [#how-these-work]

* **One entry point**: every scenario starts by telling an AI teammate the goal in a channel, no code, no commands to memorize.
* **Human / AI split**: you set the goal, give constraints, decide, and review; the teammate breaks it down, executes, produces the object, self-checks, and reports back.
* **Every step verifiable**: the output is a durable product object (an automation, a document, an AI teammate) with a stable link you can keep iterating on.
* **Irreversible actions confirm first**: sending externally, publishing publicly, or deleting, the teammate explains and waits for your go-ahead.

## Playbooks [#playbooks]

<Cards>
  <Card title="Launch a landing page" href="/use-cases/rd-product/landing-pages" description="Publish a page as an Artifact: a stable, org-shareable link you keep iterating on." />

  <Card title="Monitor every deploy" href="/use-cases/rd-product/deploy-watchdog" description="Check that a deploy actually shipped, then patrol uptime and certs." />

  <Card title="Write a PRD" href="/use-cases/rd-product/prd-writer" description="Turn a problem and a goal into a PRD engineering can build against." />

  <Card title="Organize your own tech team" href="/use-cases/rd-product/ai-team" description="Hire product, research, design, code-development, and code-review experts into one channel." />
</Cards>

## Next steps [#next-steps]

<Cards>
  <Card title="Hire an AI teammate" href="/ai-teammates/hire-an-ai-teammate" description="Set up your first AI software engineer." />

  <Card title="Connect GitHub" href="/connect/connect-tools-to-ai-teammates/connect-github-to-ai-teammates" description="Give your teammate repo access." />
</Cards>