Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.gominimal.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
What Is Agentic Storefront?

Agentic Storefront lets you add an AI chat widget to your website. Customers can ask questions directly on your site, and your AI Agent can answer using your configured protocols, website knowledge, product data, and connected lookup integrations.
Storefront is currently available as a chat widget only. More appearance options and proactive sales experiences are coming later.

​
How Storefront Differs From A Regular AI Agent

A regular AI Agent can draft replies for review, automate replies, or escalate conversations depending on your setup. Storefront is simpler and stricter: it can only automate or escalate. It does not create drafts. If the AI Agent is allowed to answer, it sends the answer directly to the shopper. If it should not answer, it escalates the conversation to your inbox (Minimal AI inbox or helpdesk channel of choice).

​
Instructions Video

​
How To Set It Up

1

Create and configure your storefront

Go to Storefront in the Minimal AI dashboard and create a new storefront.Configure the storefront name, welcome message, colors, escalation settings, and CSAT settings.
2

Add the storefront to an AI Agent

Connect the storefront channel to the AI Agent that should handle shopper conversations.This agent should have the protocols, websites, product sources, and lookup integrations it needs to answer customer questions.
3

Set the AI Agent to automate

Enable automation for the AI Agent.Storefront does not use drafts, so automation must be enabled for the AI Agent to send replies in chat.
4

Set up the automation prompt

Choose one of the automation prompt versions below.Version 1: Conservative AutomationUse this when you want the AI Agent to automate low-risk chat replies and answers backed by real lookup data.
Automate responses in the following situations:

- Opening messages and greetings, such as "Hey", "Hello", or "Hi there"
- Sign-offs and closing messages, such as "Thank you", "Have a great day", or "Let me know if you need anything else"
- Simple acknowledgements and confirmations, such as "Got it", "I understand", or "Noted"
- Any response that retrieved data from a lookup integration, including order lookups, product lookups, website lookups, customer account lookups, inventory lookups, tracking lookups, or any other integration that fetched information to answer the customer's question

The key principle: if the response is either a conversational courtesy or backed by actual data retrieved from lookups, automate it.
This is the conservative setup. It works best when your protocols clearly say which situations can be automated.
Version 2: All-In Chat AutomationUse this when you want the AI Agent to automate as much as possible.
Automate all responses, regardless of topic, complexity, or whether lookups were used.
This overrides protocol-level automation settings and automates everything that is not escalated. Guardrails still apply, but you should only use this when your AI Agent instructions and escalation rules are strong.
For most teams, start with Version 1: Safe Chat Automation.This lets the AI Agent automate:
  • Greetings
  • Sign-offs
  • Simple acknowledgements
  • Product lookups
  • Website lookups
  • Order, customer, inventory, or tracking lookups
  • Other answers backed by retrieved integration data

​
Inbox Behavior

Storefront conversations appear in your inbox. You can review messages, escalations, and CSAT feedback there.
You cannot reply to storefront conversations from the inbox yet. For now, the inbox is used for visibility, escalations, and follow-up.

​
Put Storefront Live On Your Site

Copy your storefront install snippet from the Storefront page in Minimal. Add the snippet to the <head> of your website, right before the closing </head> tag.