For service businesses with phones that ring, technicians on the road and receipts in glove boxes: a proposed workflow where every call is logged, every job assigned and every receipt reviewed. Today the call-log and driver-tracking parts run as internal pilots; jobs, expenses and the customer workspace are planned modules we scope with you first.
How the service businesses workflow runs
- An enquiry or call arrives and is logged by branch and extension Call Logs · Pilot
- It is matched to the customer (Core, planned) Core · Planned
- A callback or job is created and assigned (callback queue in development; Jobs planned) Jobs · Planned
- The technician completes it with photos and time (Jobs, planned) Jobs · Planned
- Receipts are captured against the job (Expenses, planned) Expenses · Planned
- The owner reviews history and reports Insights · Pilot
Where AI helps: Planned: AI summaries of calls and jobs and drafted follow-ups for approval. Today Call AI produces transcripts and rule-based tags only. Status changes and customer messages stay with your staff.
Modules in this workflow
Proposed workflow: every call is logged (Call Logs, pilot), missed calls become callbacks (in development), jobs are assigned and timed (Jobs, planned), drivers and technicians are tracked (Track, pilot) and receipts are reviewed instead of lost (Expenses, planned). We scope the planned parts with you before anything is quoted.
One business workspace: customers, tasks, staff and history in one place.
Purpose, features and dependencies
Customer details, tasks and history live in different places, so follow-ups slip and nobody has the full picture.
- Three staff accounts with permissions
- One workspace / location
- 500 CRM contacts with lead stages
- Notes, tasks and activity history
- CSV import and export
Availability: Planned.
Starting allowance: Three staff accounts · One workspace · 500 contacts · 1 GB storage · 50 AI drafts/month
Where AI helps: AI drafts customer replies and summarises long notes on request. It never changes customer records, prices or stock by itself.
One stock figure shared by every channel you add.
Purpose, features and dependencies
Quantities drift between the shop, the marketplaces and the supplier feed, which causes oversells and lost sales.
- Supplier CSV and feed imports (working, internal)
- Reconciliation of marketplace stock with reports (working, internal)
- Central stock ledger per SKU and location (in development)
- Barcode receipts, picks and adjustments (in development)
- Low-stock alerts from your rules (in development)
Availability: In development. Needs: Core.
Starting allowance: 1,000 SKUs · One location · One supported supplier feed
Where AI helps: AI can flag unusual movements and draft reorder suggestions from history once the ledger exists. Stock levels only change through recorded movements, never from an AI guess.
One marketplace connection: listings, stock, orders and price alerts.
Purpose, features and dependencies
Listing updates, stock changes and orders are handled in each seller portal separately, which takes hours and still leaves gaps.
- One marketplace connection per module (Amazon.ca and eBay.ca verified; Walmart.ca read-only pilot)
- Catalogue and listing updates
- Stock quantity sync from supplier availability
- Order import with status
- Price and listing alerts (daily competitor report on eBay; pricing-issue list on Amazon)
Availability: Pilot. Needs: Core. Enables: Guarded Repricing.
Starting allowance: One seller account / country · 1,000 SKUs · 500 orders per month
Where AI helps: AI drafts listing titles and bullet points from your product data for you to approve. Publishing is always a human action.
Price changes above a floor you control.
Purpose, features and dependencies
Manual repricing is slow, and unguarded automation can price below cost.
- Cost, fee and shipping-aware floor price (working, internal)
- Featured Offer monitoring where the API provides data (weekly)
- Run-by-run change history with rejected-row reports
- Maximum price / ceiling rules (in development)
- Pause per SKU, group or account (in development)
Availability: Pilot. Needs: Core + Sell.
Starting allowance: Requires Sell · 1,000 repriced SKUs
Where AI helps: AI can suggest rule changes from history, but prices never move below the configured floor. ORCA does not promise Buy Box or Featured Offer placement.
Every call and missed call in one branch-aware view.
Purpose, features and dependencies
Missed calls are not tracked, so enquiries are lost and nobody owns the callback.
- Incoming, outgoing, answered and missed calls
- Search by phone number or customer name
- Filters by branch, extension and date
- External customer calls separated from internal calls
- Callback queue with owner and status (in development)
Availability: Pilot. Needs: Core. Enables: Call AI.
Starting allowance: One supported provider · Five extensions · 1,000 external calls per month
Where AI helps: Hold and wait metrics are shown only where the provider supplies those events; nothing is inferred from recordings alone.
Transcripts and call reasons for recorded calls; summaries in development.
Purpose, features and dependencies
Listening back to calls does not scale, so details and promises are lost.
- Searchable transcripts (local transcription, quality limited)
- Call reasons (rule-based tagging, in development)
- Brief summaries (in development)
- Suggested next actions and draft follow-ups (planned)
- Human approval before any customer message
Availability: In development. Needs: Core + Call Logs.
Starting allowance: Requires Call Logs · 500 audio minutes per month · Additional audio C$0.10/minute only by agreement
Where AI helps: Summaries and sentiment are aids for review, not facts about a person. Nothing is sent to a customer without approval.
Driver assignments, delivery status, proof of delivery and live location.
Purpose, features and dependencies
Dispatch runs on phone calls and paper, so customers and staff cannot see where a delivery is.
- Driver assignments and stop lists (working, internal)
- Delivery status per stop with photo proof of delivery
- Live location, trips and replay from supported devices and the Traccar phone app
- On-shift-only location (in development)
- Delivery history per customer
Availability: Pilot. Needs: Core.
Starting allowance: Three drivers · 1,000 completed stops per month
Where AI helps: Location comes only from devices and apps where permission was granted. On-shift-only visibility is in development; ORCA does not promise always-on background tracking.
Work orders with technician time, photos and history.
Purpose, features and dependencies
Work orders live on paper or in chat, so time, parts and photos are hard to find later.
- Work orders with status
- Technician assignment and timing
- Photos and notes per job
- Parts and allocations
- Job history per customer
Availability: Planned. Needs: Core.
Starting allowance: Three technicians · 500 jobs per month
Where AI helps: AI can summarise a job from notes and photos for the customer record. Status changes stay manual.
Receipt capture with review, categories and export.
Purpose, features and dependencies
Receipts pile up, get lost, and are entered late and inconsistently.
- Receipt capture by photo or upload
- Extracted date, supplier, total and tax for review
- Approval and categories
- Duplicate flags
- CSV export
Availability: Planned. Needs: Core.
Starting allowance: 100 single-page receipts per month
Where AI helps: Extraction is a proposal that a person reviews. Expenses is not an accounting or payroll service.
Owner dashboards and exception reports; AI questions planned.
Purpose, features and dependencies
Data sits in several tools and nobody has time to assemble it.
- Owner dashboards per channel (working, internal)
- Profit and billing reports (working, internal)
- Exception reporting: missing costs, suppressed listings, pricing issues (working, internal)
- Saved reports (planned)
- Bounded AI questions over approved reports (planned)
Availability: Pilot. Needs: Core.
Starting allowance: Dashboards and exception lists for connected modules · Five saved reports (planned allowance) · 100 bounded AI questions per month (planned allowance)
Where AI helps: Forecasts are only offered where enough history exists; AI answers, once available, are limited to the approved reports.