L1 support, lifted to senior level.
In under a week.
A Medical Equipment Retailer's L1 support team was drowning in MTTR while senior reps churned. The company shipped a no-code customer-support agent on Dataworkz: ingesting internal SharePoint and Epicor knowledge, surfacing answers with citations, and giving junior staff senior-quality resolution. Live in production in under a week.
What this team shipped on Dataworkz.
- Customer
- Medical Equipment Retailer
- Industry
- Medical equipment retail · mobility & accessibility products
- Use case
- Customer support knowledge agent for L1 ticket resolution
- Stakeholder
- Initiated by the company President, GenAI capability was not in-house
- What it ingests
- SharePointEpicor
- Tooling
- RAG BuilderAgent HubSaaS deployment
- Pain on entry
- Inexperienced L1 team · high MTTR · senior-rep churn
- Status
- Live
From customer question to closed ticket, without escalation.
A customer asks about equipment setup, troubleshooting, or maintenance. The L1 support rep receives the ticket through the existing workflow.
Dataworkz retrieves the right context from SharePoint and Epicor, returns a synthesised answer with citations, and surfaces it inside the support workflow.
The junior rep delivers a senior-quality response: no escalation, no knowledge-base search, no paging a senior rep.
Three problems compounding each other.
Tickets were taking longer to resolve.
The product is mobility and accessibility equipment for end users who need answers that day, not three escalation steps later. Resolution time was moving in the wrong direction.
Junior reps escalated; seniors were leaving.
L1 didn't have the institutional knowledge to resolve issues without escalation. The senior reps who did have it were churning, and every departure made the knowledge gap deeper.
The President wanted a GenAI solution. The skills didn't exist.
The team had the support workflow and the knowledge base: but no in-house skills to prototype, test, and deploy a GenAI application. Hiring an ML team was not the move at this scale.
The capability they needed existed. They just needed a way to ship it without building the platform first.
A customer-support agent that reads the company's own knowledge base, answers with citations, and gets better every week.
The team built the entire workflow inside Dataworkz's RAG Builder: wizard-driven, no pipeline code, deployed as SaaS. The agent reads from the systems the support team already worked in, and outputs into the workflow the support team already used.
Ingest
Knowledge base documents · ERP records · equipment manuals · maintenance histories pulled in continuously.
Index
Documents chunked, embedded, indexed. Pick the model, the embeddings, the vector store. No pipeline code.
Answer
Support rep asks a question or receives a ticket. The agent retrieves context, synthesises an answer, and includes citations to source docs.
Improve
Every interaction captured. New use cases added in under a day. The agent gets better as the team uses it.
Hire and train more L1. Build a GenAI solution in-house. Or ship a wizard-driven support agent in a week.
Adds cost, not capability.
Doesn't fix MTTR, adds throughput on a workflow that's already slow. Doesn't survive senior-rep churn, every departure restarts the training loop. The bill goes up faster than service quality does.
30% MTTR improvement. Live in under a week.
Wizard-driven RAG Builder. The company's existing knowledge, SharePoint + Epicor, becomes the agent's source. No-code. SaaS deployment. The support workflow stays the same; the answers get better.
Six-to-twelve-month engineering bet.
Six months to stand up an ML team. Another six to ship something. Mid-market companies don't have ML platform engineers sitting on the bench, and the support deadline isn't moving.
The team picked Dataworkz because the agent shipped against their actual workflow, SharePoint, Epicor, the support team's existing tooling, without requiring them to become a GenAI engineering team to keep it running.
MTTR down. New use cases in a day. Live in a week.
| What they measured | Before | After |
|---|---|---|
| Mean Time to Resolution (MTTR) on customer support tickets | Baseline | 30% improvement |
| Time to add a new support use case (new product, new policy) | Weeks | Under 1 day |
| Time from kickoff to first production deployment | A platform build cycle | Under 1 week |
| Capability lift for junior support staff | Escalate or wait | Resolve at L1 with citations |
| Build path | Hire ML team · ship | No-code wizard · ship |
Outcomes from production deployment at this customer.
You don't need a GenAI team to ship a production support agent.
If you're a peer mid-market service organisation: retail, equipment, after-sales support, field service, three things matter from this deployment.
Your knowledge base is already the moat.
SharePoint, Confluence, an ERP, a CRM, a folder of equipment manuals, whatever your senior reps actually rely on becomes the agent's source. The work is making that knowledge available in the support workflow, not generating new content.
Ship before you build.
Wizard-driven, no-code RAG agents go from kickoff to production in days. A six-month internal build is the wrong answer to a problem that's costing you customers this quarter.
L1 stays. The answers get better.
The agent doesn't replace your junior staff. It hands them senior-quality context so they can close tickets without escalating. The team owns the relationship; the agent owns the lookup.
Your support workload isn't waiting for a GenAI build cycle.
Book a working session with the customer-support team. We'll walk through the workflow in this case study against the knowledge base and ticketing tools your team already uses.