Lindy builds a support agent with Slack escalation
Janine Heinrichs used Lindy to build a support inbox agent that answers from a knowledge base and sends unknown questions to Slack; live ticket results were not reported.
unite.ai
Original prompt
Shared by the source author. See source notes for context.
monitor my support inbox and help answer customer support questions using a knowledge base
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- Janine Heinrichs
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Lindy generated the configured support agent in about a minute. The article reports no live ticket handling, accuracy measurement, or customer outcome.
- Human involvement
- Heinrichs connected the Google and Slack accounts and retained the option to inspect and edit the agent before testing or publishing. The source does not report a live ticket test or an approval process for outbound replies.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Heinrichs selected Lindy’s customer support email suggestion and kept its prompt for knowledge-based replies and Slack escalation.
- Lindy generated the agent and displayed its workflow steps in the editor.
- She connected Google and Slack, reviewed the workflow nodes, and prepared it for testing or publishing.
Janine Heinrichs wanted to see how quickly Lindy could turn a plain-language request into a customer-support agent. She chose Lindy’s “Customer support email” suggestion, which supplied a prompt asking for an agent to monitor a support inbox, answer from a knowledge base, and send a Slack direct message when it could not answer.
Heinrichs left the generated prompt unchanged and submitted it. Lindy created the agent in about a minute and showed its steps in the interface. She then connected her Google and Slack accounts, opened the generated workflow, and found that she could edit the agent step, model, and other nodes before testing or publishing it.
The source documents the creation and configuration of a support agent, not a live support operation. It does not report real tickets handled, answer accuracy, or a customer outcome. That makes the useful result a quick starting workflow: use a knowledge base for grounded replies and route unknown questions to a human in Slack. Anyone copying it should test representative tickets, verify the escalation path, and keep replies in draft until the knowledge source and permissions are checked.
Prompt
Original prompt
Source: Unite.AI · Janine Heinrichs
Frequently asked questions
What did the generated support agent do?
Its prompt asked it to monitor a support inbox, answer questions using a knowledge base, and send a Slack direct message when it did not know an answer.
How long did Lindy take to create it?
Heinrichs reports that the agent was generated in about a minute, then she connected Google and Slack.
Did the source report results on real customer tickets?
No. It documents building and configuring the agent, but not a live test or customer support outcome.
More details
About this account
Heinrichs supplied the goal and accepted Lindy’s generated prompt. Lindy created a multi-step support agent configured to use a knowledge base and escalate questions it could not answer through Slack.
Opened Janine Heinrichs’s dated first-person review and read the step-by-step build. The published prompt is quoted as an exact excerpt under 25 words; the article reports agent creation and account connections, not live support results.
Credit: Janine Heinrichs — Unite.AI
Source published
Added to Agents In Use
Source checked
Tools and topics
Lindy · Google account · Slack · Support knowledge base
customer support · knowledge base · Slack escalation
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.