Trusted by
Cyber Essentials
99.5% Uptime SLA
120+ UK Businesses Deployed
24/7 Multi-Agent Coverage
QuestGates, a UK loss adjusting group, deployed AI to automate motor claims extraction across 300+ claims with a 1.8% error rate - outperforming manual entry, catching missed SLA breaches, and building the case for board-level AI expansion across their entire portfolio.
Trusted by
The 30-second version
The carousel you saw, or the one you'll see. The full case sits below.
Keep scrolling for the full case.
Key Outcomes
Vs Human Baseline
Loss adjusting runs on claims volume. To process the same workload manually, a UK firm would need a dedicated claims-entry team, not as a one-time cost, but every quarter, every year.
UK claims-handler median total package roughly £35-45k fully loaded. Manual email-to-CRM extraction averages roughly 10 minutes per claim. Annual scale assumes 50,000 claims across a multi-office adjuster. Industry manual-entry error rate from operational benchmarks; AI rate measured across 300+ tested motor claims.
The Challenge
Loss adjusting is a domain where every claim carries legal, financial, and compliance weight. QuestGates operates as a group with multiple brand pillars, including the recently acquired Brown Swords, processing thousands of claims annually across motor, property, and legal divisions.
The challenge was not just building AI - it was earning trust in a sector that runs on exactitude. Manual data entry was slow, inconsistent, and invisible to management. Claims were falling through the cracks. And with 25 active projects and a lean team, the business needed digital workers that multiply capacity rather than add complexity.
Email forwarding between brands loses origin metadata - requiring dedicated per-brand inboxes
Three-way coordination with external provider for API access, authentication, and test environments
QGLaw operates under stricter compliance rules requiring separate data handling protocols
The Solution
QuestGates' engagement began with a focused proof of concept - automating motor claims data extraction - and has rapidly expanded into a multi-agent vision spanning claims processing, legal compliance, CRM integration, and self-service automation. Each digital worker is designed to multiply the existing team's capacity, not replace it.
Incoming · claim emails
Different senders, formats, words.
Outgoing · CRM record
Same shape, every time, measurable.
Extracts structured claim data from incoming emails - policy holder details, incident location, third-party information, and loss adjuster assignment fields. Validated at 1.8% error rate across 300+ claims, outperforming manual entry and catching three claims the human team missed entirely.
Connecting claims analyst output directly into QuestGates' Cube CRM via API. Integration with NetMonkeys for Swagger documentation, JWT authentication, and AI-flagged claim tagging in the live environment.
Email sorting and prioritisation for QGLaw, which operates under stricter compliance requirements. Categorises incoming correspondence, flags urgency levels, and routes to appropriate case handlers - with legal compliance review built into the design.
Extension of the motor claims model to property insurance - covering low-value, standard, and high-net-worth claims. Same CRM integration pathway, same extraction methodology, applied to a different asset class with its own field requirements and escalation rules.
See it in action
Audio captured from a live AIOS Workforce session. The Motor Claims Analyst Agent reading an incoming broker email, classifying severity, validating against Cube CRM, and writing a structured record.
The Results
AI extraction outperformed manual data entry across all tested motor claims
Claims the human team never processed - turning a quality test into an SLA compliance discovery
Discrepancies between AI output and CRM records traced back to human input errors, not AI mistakes
After Monday's session I realised this is far more impactful than I originally thought. I was wondering if you would be willing to do a similar meeting but to our board? For all the will in the world, I won't convey the full breadth of opportunity the Implement AI platform offers.
Overall it looks brilliant. There are also three claims within that set that we didn't process - which is a win for you guys to capture something we were out of SLA for.
Our investors were also interested in learning more - but selfishly, I want us to have a head start before the news travels through their entire portfolio.
There's lots of different scenarios we have to deal with because the client dictates how we interact with data. Having a multi-tool for that is amazing. I've been noting down all the different use cases I know already exist in the business that we've been pushing away.
Leadership
Executive Chairman & Co-founder
Award-winning technology entrepreneur who has provided technology services to clients ranging from SMEs to FTSE-100 companies across sectors from retail to defence.
pierslinney
CEO & Co-founder
Combined expertise in technology and clinical fields. Experienced innovator in deploying exponential technologies during platform shifts.
aalokshuklaWe connect to your systems and show you where accuracy, speed, and revenue are leaking - before deploying a digital workforce to close the gaps.