AI Features
Use AI to diagnose repairs from photos, forecast inventory needs, draft customer messages, and run an always-on AI receptionist chat.
AI Features
QuantumRepair includes several AI-powered tools designed to save you time and help you make better decisions. All AI features have sensible fallbacks — they assist your team, they do not replace judgment.
AI Diagnostics
Where: Ticket detail page → "AI Diagnose" button, or automatically during new ticket creation
What it does: Given a description of the issue (or a photo of the device), the AI analyses the damage and returns:
- The most likely damage type(s)
- Severity rating (minor, moderate, major, severe)
- Suggested parts to fix it (matched against your actual inventory catalog)
- Estimated repair cost range
- Confidence level
How to use it:
- When creating a new ticket, describe the issue in the "Reported issues" field
- Optionally upload a photo (see Vision Intake below)
- Click AI Diagnose
- Review the suggestions — accept all, accept some, or dismiss
- Accepted parts are added as line items on the ticket
The AI draws from your parts catalog when suggesting items. If a part is in stock, it notes that. If it is low stock or not in catalog, it flags that too.
AI Vision Intake
Where: New ticket wizard → photo upload step
What it does: You upload a photo of the damaged device. The AI analyses the image and identifies damage that is visible — cracked screens, bent frames, water damage indicators, missing components. This is especially useful for:
- Giving accurate quotes before the tech has fully inspected the device
- Creating a photographic record of the device condition at intake
- Speeding up the intake process for technicians
How to use it:
- On the new ticket form, click Upload photo
- Take a clear, well-lit photo of the damaged area
- Upload it — the AI processes it and fills in the damage fields automatically
- Review and confirm the AI suggestions before saving the ticket
Tips for better results:
- Use natural lighting — avoid strong shadows or reflections
- Include the full device in the first shot, then a close-up of the damage
- For screen damage, include the powered-on screen if possible
AI Inventory Forecasting
Where: Dashboard → AI page → Inventory Forecast section
What it does: The AI analyses your sales history and current stock levels to predict which products will run out first and when. For each product it shows:
- Current stock and reorder point
- Predicted demand over the next 30 days
- Days until stockout (how many days until you run out at the predicted demand rate)
- Recommended reorder quantity
- Risk level: High Risk (stockout in under 7 days), Monitor (7–14 days), Healthy
- Confidence score — how reliable the prediction is based on your data history
- AI reasoning — a plain-English explanation of why it is recommending a reorder
How to use it:
- Go to the AI page from the sidebar
- Click Run forecast
- Review the forecast table — sort by risk level to see the most urgent items first
- For any item marked "Should reorder", click Create purchase order to automatically draft a PO for that item
For best results, run the forecast after you have at least 30 days of sales history. Shops with less history will still get recommendations, but with lower confidence scores.
AI Chat Widget (AI Receptionist)
Where: Your public website (via the embedded chat widget)
What it does: An AI chat assistant appears on your website and answers common customer questions 24/7, including:
- Service pricing
- Availability for booking
- Part availability ("Do you have an iPhone 14 Pro battery in stock?")
- General repair timelines
- Your shop location and hours
Beyond answering questions, the AI can automatically book appointments — it collects the customer's name, phone, device type, and preferred time, then creates the booking in your system.
When a visitor asks something the AI cannot confidently handle, it escalates to your inbox and sends a notification to your team.
See the Booking & Chat Widgets guide for setup instructions.
AI Draft Replies (Inbox)
Where: Inbox → any conversation → AI draft button
What it does: When you are replying to a customer in the inbox, clicking the AI draft button generates a suggested reply. The AI draws on:
- The conversation history
- The customer's linked repair tickets
- Your shop name and context
The draft is a starting point — always review and personalise before sending.
AI Tone Settings
Where: Settings → AI Tone
You can configure how the AI communicates with customers. Options include:
- Professional — formal, concise, no slang
- Friendly — warm and approachable
- Casual — relaxed, conversational
This affects the chat widget's responses and the inbox draft replies.
AI Copy for the Storefront
Where: Settings → Storefront → any text field with the "AI" sparkle icon
When editing your storefront content (tagline, service descriptions, meta description), click the AI sparkle icon next to any text field. The AI generates copy based on your shop name, services, and location. Use it as a starting point and edit to match your voice.
Understanding AI limitations
- The AI assists, it does not replace — always review AI-generated parts suggestions before adding them to a ticket. The AI does not know the repair history of a specific device.
- Vision intake is not 100% accurate — some types of internal damage are not visible in photos. Treat the AI diagnosis as a starting estimate, not a final quote.
- Forecast accuracy improves over time — the more sales history you have, the better the predictions get.
- The AI does not store conversation history between sessions — each new chat widget session starts fresh.
Provider and reliability
QuantumRepair uses Anthropic's Claude as the primary AI model, with an automatic fallback to OpenAI when needed. A monthly spend cap prevents runaway AI costs. All AI calls are rate-limited to protect your account.