> ## Documentation Index > Fetch the complete documentation index at: https://docs.rippit.com/llms.txt > Use this file to discover all available pages before exploring further. # Pipeline examples > Prompts to start a pipeline from: enrich every record, report and alert on a schedule, research the outside world, and combine all three. Each prompt below is something you can paste into Rippit chat, or give to your own assistant with "use Rippit to" in front. Change the source, the schedule, and the destination to match your workspace. Rippit asks about anything it still needs, then builds the [pipeline](https://docs.rippit.com/pipelines/overview.md) as a draft for you to test. ## Enrich every record Add the columns your source systems never recorded, on every row, and keep them filled as new rows arrive. ```text wrap theme={null} Add three columns to every Zendesk ticket: the main reason the customer contacted us, the product area, and a 1 to 5 frustration score. Keep them filled in as new tickets arrive. ``` ```text wrap theme={null} On every Gong call, flag whether a competitor was mentioned, which one, and quote the sentence where it came up. ``` ```text wrap theme={null} For every conversation our AI agent handled, judge whether its answer was correct, wrong, or no answer, and re-judge when a human agent replies afterward. ``` Once the columns exist, ask about them in chat ("which contact reason grew fastest this month?"), chart them on a canvas, or build the next pipeline on top of them. ## Report and alert on a schedule Anything you would want to know regularly, delivered where your team already works. Each prompt names a period that matches its schedule, so every run reports on its own window and nothing is repeated. **Understand what support is dealing with** ```text wrap theme={null} Every Friday at 4pm, post one Slack report to #support-leads summarizing the biggest themes from the past week's support conversations, with links to three example conversations for each. ``` ```text wrap theme={null} Every morning, post an urgency summary to Slack, including a brief update when nothing urgent is found. ``` ```text wrap theme={null} Every Monday, email each team lead the coaching opportunities from their agents' conversations last week, one section per agent. ``` **Watch revenue and churn risk** These depend on what your workspace brings in. Sales call transcripts and account data open up the accounts and calls angle. With support conversations alone, stick to the sentiment and complaint versions. ```text wrap theme={null} Every Monday, email me the accounts that raised a complaint last week that we did not resolve, with the conversations behind each. ``` ```text wrap theme={null} Every week, summarize the objections that came up in calls that did not close, and post it to the sales channel. ``` ```text wrap theme={null} Every morning, alert #cs-escalations about accounts with three or more negative conversations in the past day. Stay silent when there are none. ``` **Track what the product gets wrong** ```text wrap theme={null} Every day, post a summary of suspected product bugs with conversation links to Slack for triage. ``` ```text wrap theme={null} Every Friday, summarize feature requests and product friction from customer conversations and post a digest to the product channel. ``` ```text wrap theme={null} Every Monday, tell me whether complaints about checkout went up or down against the four weeks before. ``` Notice that some prompts say what to do when nothing matches. That is the difference between a quiet week and a pipeline that silently stopped working. ## Research the outside world Bring in what your own data can't tell you, and keep it current. ```text wrap theme={null} Every Monday, search the web for news about each of our 50 largest accounts: funding, layoffs, leadership changes, and product launches. Save each finding as a row with its link and date. ``` ```text wrap theme={null} Every week, ask ChatGPT, Claude, and Gemini which customer support QA tools they recommend for a mid-size SaaS company. Record each answer and its citations, and tell me whether we were mentioned. ``` ```text wrap theme={null} For each company in my target accounts table, find the head of support and the head of customer experience, and save their name, role, and profile link. ``` See [External research](https://docs.rippit.com/pipelines/research.md) for what each search method returns and which providers it can use. ## Combine them The steps compose, so one pipeline can carry a job from raw data to a finished action. Describe the whole job and Rippit works out the chain. **Churn briefs for each account owner** ```text wrap theme={null} Score every support conversation for churn risk. Every Monday at 8am, for each account with two or more high-risk conversations in the past week, look up its owner and renewal date in our accounts table, search the web for recent news about the company, and email the owner a short brief with the evidence. ``` What Rippit builds: 1. **AI columns** on conversations: a churn-risk score and the reason behind it. 2. A **lookup** that adds each conversation's account owner and renewal date. 3. A weekly **rollup** with one row per account, kept only when it has two or more high-risk conversations. 4. An **agent run** on each of those rows that researches the company and writes the brief. 5. An **action** that emails each brief to the account's owner at 8am. **Issues from bug reports** ```text wrap theme={null} Every morning, open a Linear issue for each product area where more than one customer reported a bug yesterday, with a summary, the affected accounts, and links to the conversations. ``` The [overview](https://docs.rippit.com/pipelines/overview.md#what-you-can-build) walks through one more: bug triage with account context. > [card] [Pipeline building blocks](https://docs.rippit.com/pipelines/steps.md) > > What each kind of step can do. > [card] [Build, test, and run](https://docs.rippit.com/pipelines/build-and-run.md) > > Test safely, check the cost, and go live. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.