Flint AI Assistant

Flint is an autonomous AI agent for Zendesk that reads incoming requests, decides what to do, and responds on its own — with enterprise-ready security and compliance support built in.

• Handles repetitive support requests automatically

• Pulls from ticket context, your knowledge, and connected sources

• Lets you review and improve responses through feedback

• Helps agents work faster by suggesting replies, next steps, and translations when needed

What Flint Does

• Reads each ticket and understands what the customer is asking

• Pulls from connected docs, help content, and backend tools like Shopify, Magento, and other API-connected systems

• Reduces repetitive ticket handling for agents

• Lets you refine future responses with simple prompt instructions

• Helps agents work faster with reply suggestions, next steps, and translations when needed

• Grounds each reply in connected knowledge and available data

• Keeps responses more consistent across similar tickets

What That Unlocks

• Faster first responses

• More consistent answers across the team

• Less repetitive manual work for agents

• Easier scaling for repetitive support flows

• Better team productivity

• Lower support costs as volume grows

• Better customer experience

• Customer-service best practices built in

• Stronger data security and compliance support (GDPR, ISO 9001, ISO 27001)

How It Works

1. A customer request comes in

2. Flint reads the ticket context and surrounding details

3. It checks connected knowledge, docs, and backend systems

4. It determines the best response for that type of request

5. It replies to the customer automatically

6. If needed, you can refine future behavior with one prompt

7. Flint applies that guidance to similar requests going forward

8. If it cannot solve the ticket, it escalates to an agent

Common Use Cases

• Handle repetitive customer questions automatically

• Answer support requests using connected knowledge and backend data

• Refine tone, structure, and response rules with simple prompts

• Support agents with suggestions, next steps, and translations when needed

Set it up, connect your knowledge and systems, and let Flint handle repetitive requests while your team focuses on the exceptions.

Data Disclosure

Data accessed from your Zendesk account:* Agent profile: name, email, role, user ID* Ticket data: subject, description, status, field values, comments (cached up to 24 hours)* Help Center articles: title, body, section, category (indexed as embeddings)* Webhooks: list, create, update

Data sent to third parties:* OpenAI — messages and ticket context for AI inference. OpenAI does not train on API inputs.* Google Gemini — messages and ticket context for AI inference. Gemini does not train on API inputs.* Stripe — agent email, display name, token usage counts for billing.

Data we do not collect: End-user personal data is not stored. Ticket text is processed transiently and not retained beyond 24 hours.

Retention: Auth tokens expire within 24 hours. Chat threads deleted after 24 hours. Help Center embeddings retained until deleted via app settings.