E-commerce brand, owner-led

We replaced ticketing SaaS with a custom support system

An owner-led e-commerce brand wanted to focus customer service on common order questions and replace the ticketing SaaS it used for that work.

IndustryE-Commerce
ServicesSaaS Replacement, AI Agents

Support used only a limited part of the SaaS workflow

The brand handled inquiries about shipping status, returns and product features in a ticketing SaaS. Many functions went unused. Staff still switched between tickets, the order system and internal guidance to answer a question. Rules for exceptions also sat in team notes.

Management was considering a custom solution. The support team primarily wanted reliable handovers between automated answers and personal handling. An agent alone would not have been enough. The inbox, history, ownership and resumption of open cases all had to keep working after replacement.

The replacement had to cover the full case workflow

Our task was to assess which SaaS functions in use needed replacing and build the appropriate support workflow. Recurring requests would be answered using approved knowledge and permitted order data. Complaints, unclear identity and departures from the standard process went to people. With its own solution, management also took responsibility for operations and maintenance.

The agent received a complete case handling environment

  • The inbox kept cases together

    New messages were assigned to existing conversations. Repeated deliveries could not create duplicate answers or additional cases.

  • Answers used restricted data access

    The agent received approved product information and necessary order data after checking the association. It could not issue refunds or change contracts itself.

  • Escalations included the conversation history

    The case interface showed the request, sources, draft answer and handover reason. Automated processing stopped as soon as a staff member took over.

  • Operations included error handling and archive access

    Incidents triggered alerts and left open cases visible. Required historical conversations were secured before shutdown and made available under controlled access.

The support team tested handovers and failures first

With support leadership, operations and technical support staff, we reviewed the functions in actual use. We used this to define the replacement’s scope. Rarely used reports were simplified. Shared case history and manual takeover remained firm requirements because inquiries could otherwise be lost.

In working sessions, the team reviewed historical inquiries and deliberately difficult conversations. These included customers sending several messages in succession and follow-up questions about an already escalated order. We also tested a failure of order access. In that case, the agent had to make the missing information clear when handing over to the team.

Migration proceeded channel by channel. During testing, the previous system remained available, but each channel had an explicitly designated system allowed to answer. Before shutdown, we checked open cases, archive access and delivery routes. Technical support staff and support leadership practiced returning to manual handling together.

The custom support workflow took over customer inquiries

The required functions ran in the custom system, and the ticketing SaaS could be retired. The agent handled approved standard requests. The support team took over exceptions with their conversation history and could correct answers at any time. Knowledge maintenance, technical support and error handling had named owners in ongoing operations.

SaaS Replacement needs more than good answers

  • Capture the functions in use

    Check archiving, delivery and ownership too. These functions are easy to overlook in an agent-only demonstration.

  • Handovers must stop automation

    Once a person takes over a case, the agent must not send a competing answer.

  • Your own system needs its own support

    Before replacement, clarify who fixes errors, maintains knowledge and approves process changes.

A first conversation takes 30 minutes.

We discuss which functions of your support SaaS a custom workflow would need to take over.