How Tekion resolves 50% of support chatswith AI

The Challenge

Tekion built a cloud-native platform that runs the entire automotive retail operation: sales, service, parts, accounting, and finance and insurance, all on one system. When thousands of dealerships run their daily business on your software, support is not a cost center. It is the thing standing between a dealership and a day where it cannot close a repair order or deliver a vehicle.

That created a specific set of pressures:

• Business-critical urgency. A blocked workflow is not an inconvenience. If a service advisor cannot close a repair order or a salesperson cannot push a deal through, the dealership stops earning revenue that hour.


• Volume that scales with growth. Every new dealership and every new user added support load. Growing the support team in step with the customer base was neither fast enough nor economical.


• Users who speak dealership, not software. Questions arrive in the language of the shop floor, not the product documentation, which makes conventional keyword search and scripted chatbots close to useless.


• Dealership hours, not office hours. Service lanes open early, sales floors run late, and weekends are peak retail. Support demand does not fit a nine-to-five staffing model.


• A product that ships continuously. Frequent releases meant documentation was perpetually catching up, so agents were often the only reliable source of a current answer.

The Solution: Grounded AI Built Into the Support Workflow

Tekion deployed IrisAgent through a direct API integration into its existing support stack rather than bolting a separate chat widget onto the product. The AI became part of the workflow agents and customers already used.

What was implemented:

• IrisGPT grounded on Tekion's own knowledge. Product documentation, release notes, and historical resolved tickets became the answer source, so responses reflect how the platform actually behaves today rather than a generic model's guess.
• Automatic tagging, routing, and prioritization. Incoming queries are classified by module (service, parts, sales, accounting) and severity, then routed to the team that owns them, which removed a triage step that used to sit in front of every ticket.
• Real-time agent assist. Agents receive grounded answers, the relevant documentation, and similar past resolutions inline, so resolving an unfamiliar issue no longer means searching several systems or asking a senior colleague.
• Automated knowledge generation. Resolved tickets are converted into reusable knowledge articles, so the knowledge base keeps pace with a continuously shipping product instead of falling further behind each release.
• Sentiment detection and escalation. Frustrated users and stalled conversations are surfaced in real time so a blocked dealership gets human attention before the issue becomes an account problem.

The Results

Across more than 300,000 support queries processed on the IrisAgent platform, Tekion has seen:

• A 50% chat resolution rate. Half of all support chats are resolved by AI without a human agent, covering the repetitive how-to, configuration, and workflow questions that used to consume the front line.
• 5x improvement in agent productivity. With grounded answers, similar past cases, and automatic routing delivered inline, agents resolve far more per shift and spend their time on genuinely complex dealership issues.
• More than 300,000 support queries processed, a volume that would have required a materially larger support organization to absorb through headcount alone.
• Coverage that matches dealership hours. Early service lanes, late sales floors, and weekend retail peaks are all covered without staffing every hour.
• Support capacity decoupled from customer growth. Adding dealerships no longer means adding support headcount at the same rate.

Why Automotive Retail Software Teams Choose IrisAgent

• Grounded, not guessed. Every answer traces back to Tekion's own documentation and resolved tickets, which is what makes AI safe to put in front of a workflow where a wrong answer costs a dealership money.
• Domain language handled natively. The AI resolves questions phrased the way dealership staff actually phrase them, not the way the manual is written.
• API-first deployment. IrisAgent integrates into an existing support stack rather than requiring a new customer facing surface or a re-platform.
• Knowledge that maintains itself. Automated article generation from resolved tickets keeps coverage current against a fast release cadence.
• Transparent ROI. Model the impact on your own support volume with IrisAgent's ROI Calculator.