> Source: https://mywaypoint.ai/docs/platform
> Documentation index: https://mywaypoint.ai/llms.txt

# How Waypoint fits together

**Waypoint AI changes your connected systems into agent-ready data, Chip answers questions & builds on it, pipelines run the work on their own & the results land in the systems your team already works in, Salesforce, Jira, Slack, ServiceNow & the rest of your stack.**


## For

- [How Chip Works](https://mywaypoint.ai/how-chip-works)
- [Chip Build guide](https://mywaypoint.ai/docs/build)
- [Key terms](https://mywaypoint.ai/docs/terms)

## In

- [Salesforce](https://mywaypoint.ai/docs/integrations#salesforce)
- [Jira Cloud](https://mywaypoint.ai/docs/integrations#jira-cloud)
- [Slack](https://mywaypoint.ai/docs/integrations#slack)
- [ServiceNow](https://mywaypoint.ai/docs/integrations#servicenow)

## Your workspace

Everything Waypoint does for your team happens in one workspace: the data, the people & everything Chip builds. Inside it the work is organised into spaces, one per use case, so a colleague who opens the renewals space finds the renewals tracker, the digest that runs on Monday & the note that says why it was built. Your workspace is yours alone: its data is held apart from every other workspace's & every query stays inside it.

## Your systems become agent-ready data

Connect a system once & everything in it, every case, ticket & account, becomes data Chip, your pipelines & your dashboards work on directly: typed, current as the source changes & ready to join with your own tables, with nothing to export or re-key. Ask for a dashboard over your open cases & the data is already there. The system you connected stays the system of record.

## Chip, in two modes

**Ask** answers: a question in, an answer out with every source linked so you can check it, & nothing in your data changes. **Build** makes things: a prompt in & a table, the screen your team opens, an automation or a document out, with the plan shown before anything is created. Both work over the same data, so what a pipeline files tonight is what Ask answers from tomorrow & what Build puts a dashboard over next week. [Ask](https://mywaypoint.ai/docs/ask) & [Build](https://mywaypoint.ai/docs/build) each have their own guide.

## Pipelines run the work

A pipeline is work that gets done without anyone watching. It starts from something that already happens: a case created in Salesforce, a ticket updated in Jira, a message in Slack, a schedule, a row that changed, a button someone pressed or a call to a URL. Its steps read & write your data, label what arrived, ask an agent for judgement & finish the job in your systems: the issue created in Jira, the channel told in Slack, the email sent, the case updated in Salesforce or ServiceNow. You stay in control: a pipeline starts as a draft & a draft never fires, you test it with sending switched off & you switch it on when it is right. Open any run afterwards & read exactly what happened, step by step.

Chip builds automations as pipelines, so what Chip built from a prompt & what you drew by hand are the same thing in the same editor. [The pipeline builder](https://mywaypoint.ai/docs/platform/pipeline-builder) has the detail; [Agents as steps](https://mywaypoint.ai/docs/platform/agents) & [Labels & routing](https://mywaypoint.ai/docs/platform/labels) cover the two kinds of step that carry judgement.

## What you get

- **Work done in your systems.** An issue created & commented in Jira, a channel told in Slack, a case updated in Salesforce or ServiceNow, an email sent from your own domain.
- **Dashboards & screens over your data.** A list, a kanban, a calendar, a dashboard of numbers & charts, an explorer whose grouping the reader switches. Each view has its own address, so a link hands someone the screen.
- **Data you can query.** The fields extracted from a contract, the summary filed every Monday, the label on every case: rows & records you chart, filter & ask Chip about.
- **Documents.** What Chip writes down about every build (the intent, the overview & the plan) & the files you give it to read, so anyone can see why something exists.

## Where it runs & whose model

Waypoint runs in Waypoint's cloud or self-hosted in your own environment, the same product either way, so your data stays where your security team wants it. Chip works with the model provider you already use: an administrator adds it once & every agent step & labeling step runs on it. [Where it runs](https://mywaypoint.ai/docs/platform/deployment) has the detail & [Security](https://mywaypoint.ai/security) describes the controls.

## Read next

- [The pipeline builder](https://mywaypoint.ai/docs/platform/pipeline-builder): triggers, steps, branches, drafts & versions.
- [Agents as steps](https://mywaypoint.ai/docs/platform/agents): the step that reads, looks things up & decides.
- [Labels & routing](https://mywaypoint.ai/docs/platform/labels): policies that label, routes that follow the label & how accuracy is watched.
- [Where it runs](https://mywaypoint.ai/docs/platform/deployment): Waypoint's cloud or yours, bring your own model.