Selected work · Local home services
A trust-first website and request-triage system for a family-run roofing company that needs to turn urgent searches and referrals into organised, credible enquiries.
The business tension
Homeowners are anxious, often on a phone, and rarely know the contractor’s terminology. Northmark needs enough context to respond responsibly, but a long or diagnostic form would either lose the customer or create unsafe confidence.
Reduce uncertainty before asking for commitment.
The experience connects homeowner observations, safe photo capture, human review, scheduling, and owner-approved follow-up. It never diagnoses roof condition or implies automatic dispatch.
What the diagnosis changed
People had to translate a worrying situation into an internal service category before they could ask for help.
Project evidence, process, and human accountability were separated from the quote journey.
Calls and forms arrived without consistent address, timing, photo, or ownership context.
Experience walkthrough
The sequence follows the complete customer and operational journey. Each explanation shows what changed, why the decision was made, and what customer or business moment becomes easier.
Recognition, proof, and next step
A concise homeowner promise, real-work photography, situation-led pathways, explained process, and project evidence appear before the request. Orange is reserved for action while the overall tone remains calm and field credible.
Observation-first mobile triage
People select what they can see—leak, visible damage, storm change, or ageing roof—without being asked to diagnose it. Immediate-danger guidance exits to local emergency help and the workflow fails closed to human review.
Useful request context
Address, property type, timing, known age, water entry, access, and contact details are progressively disclosed. Every field has a practical reason and uncertain answers remain valid.
Optional evidence capture
The interface asks only for photos taken from a safe location and explicitly says not to climb. Progress, success, failure, retry, captions, and removal are visible while successful files stay preserved.
Accuracy and consent
Original observations are shown back without reinterpretation. Contact consent is scoped to the request, marketing remains separate, and the next state is described as human review rather than instant diagnosis.
Decision-based proof
The story pairs the homeowner concern with inspection focus, selected scope, materials, and care notes. Before and after imagery supports the decision narrative without claiming a real client outcome.
Operational overview
The queue combines workload signals, review states, completeness, region, ownership, and next action. Charts appear only where comparison improves prioritisation and the next decision.
Facts before interpretation
Original observations, uploads, timing, location, and activity remain traceable. Staff can assign ownership or request clarification without overwriting what the homeowner submitted.
Safety checkpoint
Reported facts and unknowns are separated. Scheduling stays locked until a reviewer records a safe next action and an accountable outcome, preventing an urgency flag from becoming an automated diagnosis.
Route-aware coordination
Route bands, staff-set duration, travel buffer, message preview, and conflict handling sit together. A person confirms the slot; the system only organises constraints.
Transparent waiting state
The customer receives a reference and a clear ‘waiting for review’ status—not a false appointment promise. Additional information appends to the record without replacing the original request.
Owner-controlled communication
Templates, timing, approval, stop-follow-up, and audit history are explicit. The system makes considered follow-up easier while keeping every send and opt-out decision human-owned.
Success signals
These signals connect the experience to customer behaviour and day-to-day operations.
Reflection and boundary
The key boundary is that software may organise observations but must not assess structural safety. The next step is to test the language under stressful mobile conditions, validate photo recovery on weak connections, and make sure every request stays visible only to the responsible team.