The challenge: 23% of jobs requiring a follow-up call because the original data was incomplete

Keystone Facilities Group manages 220 commercial properties across the Midlands and North of England: office buildings, retail parks, industrial estates, and mixed-use developments. Their helpdesk team of eight handled all tenant maintenance requests by phone.
Each call followed the same pattern: identify the tenant, identify the property, capture the fault details, determine urgency, select the appropriate contractor, and create the work order. Average call time: 9 minutes. Monthly volume: 3,100+ calls.
The helpdesk was the single largest overhead line item in Keystone's operations. Staff turnover was high (the work was repetitive), training took 6 weeks per new hire, and data quality on work orders was inconsistent. 23% of dispatched jobs required a follow-up call to the tenant because the original fault description was incomplete.

The solution: Structured fault capture across all 220 properties, with priority dispatch for safety hazards

Thrumble AI handles all inbound maintenance calls across all 220 properties. The AI identifies the tenant by phone number or name, confirms the property and unit, and captures the fault using a structured diagnostic flow tailored to the most common FM categories: HVAC, plumbing, electrical, access/security, cleaning, and general fabric.
The AI asks targeted follow-up questions based on fault type (e.g., for HVAC: is the unit not heating, not cooling, making noise, or leaking?), captures photos via SMS link where relevant, and creates a fully structured work order with contractor assignment and SLA classification.
Priority faults (safety hazards, water ingress, security breaches) trigger immediate contractor dispatch and a simultaneous alert to the property manager.

The results: The numbers speak for themselves

• 3,100 maintenance jobs dispatched per month across 220 properties.

• Average call duration: 2 minutes 15 seconds, down from 9 minutes.

• Incomplete work orders dropped from 23% to 3%. Contractors now arrive with the information they need on the first visit.

• First-time fix rate improved from 64% to 81%, directly attributed to better fault data captured during the call.

• 94% of calls handled fully autonomously. 6% escalated to a human coordinator for complex multi-trade or landlord-approval issues.

The commercial impact: £280K saved. £145K in avoided repeat visit costs. Deployed across 220 properties in 7 weeks.

£280,000 per year in helpdesk staffing costs saved.
The improvement in first-time fix rates saves an estimated £145,000 per year in repeat visit costs and contractor call-out fees.
Tenant satisfaction with maintenance response improved from 3.3/5 to 4.5/5 across the portfolio.
Deployed across all 220 properties in 7 weeks.

3,100 jobs dispatched. 220 properties. First-time fix rate up to 81%. £280K saved.