Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/arainey2022/myaskai-docs/llms.txt

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

When a customer is stuck, they often reach for a screenshot rather than words. My AskAI’s image reading capability means your agent doesn’t have to ask the customer to describe what they’re seeing — it can look at the image directly, understand the context, and give a relevant, informed response. This removes a significant source of friction from support conversations and leads to faster, more accurate resolutions.

What Image Reading Enables

Your AI agent can process images that are shared in:
  • Chat messages — inline images or attachments sent during a live conversation
  • Email bodies — images embedded directly in an email
  • Email attachments — image files attached to a support email
Once an image is received, the agent processes it using a vision-capable language model and incorporates what it sees into its response — no manual tagging or human review required.
Images are not stored on My AskAI’s servers. When you review conversations in Improve → Inspect & Logs, you will see a text description of what the image contained rather than the image itself.

Use Cases

A customer shares a screenshot of an error message or a broken UI state. Rather than asking the customer to type out the error code, your agent reads the screenshot, identifies the error, and provides the relevant troubleshooting steps immediately.
A customer uploads a photo of the item they received to request a return or exchange. The agent can identify the product, assess visible damage from the image, and guide the customer through the returns process — or escalate to a human agent with the image context already captured.
A customer shares a photo of a damaged parcel or defective product. Your agent can confirm the damage is visible in the image and trigger the appropriate resolution workflow, such as issuing a replacement or routing to your logistics team.
A customer shares a photo of a device, cable setup, or configuration screen. The agent uses the visual context to give more precise setup guidance rather than walking through every possible scenario.
A customer uploads a photo of a receipt or printed order confirmation. The agent can read the document, extract relevant details, and use them to answer the customer’s query — for example, verifying a purchase date for a warranty claim.

How It Works

My AskAI passes images shared in conversations to a vision-capable large language model (LLM) alongside the conversation history. The model analyzes the visual content — reading text, identifying objects, interpreting diagrams — and returns a description that is used as context when generating the agent’s reply. Up to 3 images can be read and processed within a single message or reply.

Channel Availability

Image reading is available on the following helpdesks and channels:
Helpdesk / ChannelImage Reading Available
Intercom (chat)✅ Yes
Intercom (email)✅ Yes
Zendesk Messaging✅ Yes
Zendesk Tickets / Forms✅ Yes
Freshdesk✅ Yes
Freshchat⚙️ In development
Gorgias (chat)⚙️ In development
Gorgias (email)⚙️ In development
HubSpot (chat)⚙️ In development
HubSpot (email)⚙️ In development
Slack❌ Not available
Microsoft Teams❌ Not available
API❌ Not available

Pricing

Image reading incurs an additional usage cost of $0.02 per reply where the feature is used. You are only charged when an image is actually read and processed — conversations with no images attached are unaffected.

Using Images in Tasks and Guidance

Image reading integrates with My AskAI’s automation features:
  • Tasks — You can reference images within task definitions. For example: “Request an image from the user of the damaged parcel before proceeding” or “If the user shares an image, transfer the conversation to the returns team.”
  • Guidance — Guidance rules can trigger on image content. For example: “If someone shares an image of a broken box, escalate the conversation to a human agent.”
Combine image reading with human escalation rules to automatically route conversations where a customer has shared evidence (such as a broken product photo) directly to the right team, pre-loaded with the image context.

Limitations

  • A maximum of 3 images can be processed per message or reply.
  • Very large image files may be resized before processing, which can reduce the legibility of small text.
  • Handwritten text in images is recognized on a best-effort basis and may not be reliably transcribed.
  • Images shared in Slack, Microsoft Teams, or via the API are not currently processed.
  • The images themselves are not stored on My AskAI servers; only a text description is retained in conversation logs.

Build docs developers (and LLMs) love