Just Carpets’ support team handled thousands of tickets each month in Trengo.
While tickets had labels and tags, there was no certainty they reflected what customers were actually saying.
The company couldn’t confidently answer:
Are tickets correctly categorized?
Do the tags match real customer intent?
Which issues take the most time to resolve?
Without analyzing the actual conversation content, leadership only had a partial picture of the workload. There was also a strong suspicion that mislabelled tickets were hiding the most valuable automation opportunities.
The idea of building a new SaaS platform for AI was stalled — there was no validated dataset to guide the investment.
We began with a tooling discovery phase, working with support staff to map workflows.
All ticket data, including customer and agent messages, internal notes, and existing tags was exported into a single structured dataset.
Step 2 – Semantic Content Analysis
We deployed a custom LLM agent to review every ticket, reading the actual text to:
Verify if tags matched the ticket’s true content.
Suggest primary and secondary reclassifications where mislabels were found.
Detect product and service mentions for accurate filtering.
Score sentiment progression across the conversation.
Identify recurring triggers through root cause analysis.
Step 3 – Automation Opportunity Mapping
With accurate tagging restored, we quantified complexity, urgency, and resolution times for each ticket type identifying the top candidates for automation based on ROI potential.
We delivered a clear, lasting solution to detect exactly what customers’ tickets are about — not just what they are tagged as.
This live analysis tool processes every new ticket in real time, reading the actual conversation to generate a deeper, more actionable report than Trengo’s native analytics.
This includes:
Continuous validation and improvement of tag accuracy.
Agent performance tracking across response quality, sentiment handling, and resolution speed.
Conversation sentiment scoring to detect issues early.
Trend detection to surface new or rising support topics before they become major problems.
By combining automation with an always-on analysis engine, Just Carpets’ support team now has a permanent intelligence layer ensuring they can make data-backed decisions, keep the automation roadmap current, and adapt quickly as customer needs evolve.
By combining content-level analysis with automation and AI, Just Carpets reached better conclusions then ever.
Resulting in the following clarification and mechanisms:
Reliable tagging and classification, ensuring accurate reporting and AI training.
Reduction in repetitive manual tag work, freeing the team to focus on complex cases.
Ongoing operational intelligence via the new AI analysis tool, surpassing the insights available in Trengo’s built-in analytics.
The result: faster customer support, more confident data-driven decisions, and a scalable automation roadmap — all without building a costly new platform.
Key Results
Reduction in repetitive analytics work
More insights into the actual problem -> solutions
Accurate estimations of the correct tags
