Case Studies / How AI Agents in Customer Support Cut Softorino’s Team Workload by 44%

Case Study

How AI Agents in Customer Support Cut Softorino’s Team Workload by 44%

By Bogdan Pecanac, Service Delivery Manager

For app companies that need to stop backlogs from building up

Updated on October 30, 2026

Services

Customer Support
Technical Support
Human & AI Team
Support Workflows

About the Company

Softorino builds small, easy-to-use apps for Mac and Windows. Its tools, like WALTR PRO and AltTunes, help users transfer files to iPhone and iPad and manage iOS data from a computer. As ticket volume grew, so did the support queue. A busy shift could end with a backlog big enough to spill into the next one. Customers also reached out by phone, chat, and email at all hours. Softorino needed 24/7 L1 customer and technical support that could keep pace. They turned to SupportYourApp to build a team that pairs live agents with AI. This AI agent case study covers how that setup works and what it changed.

Challenges

With volume climbing, Softorino needed round-the-clock support that could clear requests as fast as they came in, on every channel.

Growing ticket backlog

The team handled 800 to 900+ requests a month. In busy periods, a backlog could build fast enough to overwhelm a shift. Unresolved requests then rolled into the next one and waited longer for a reply.
Calls, chats, and email 24/7

Customers called, chatted, and emailed about account questions and technical issues at any hour. Covering L1 customer and technical support on all three channels around the clock meant every channel needed someone on it, on every shift.
A fixed CSAT target

The project runs on an agreed CSAT threshold of 85%. Holding that score month after month gets harder when queues grow, since longer waits show up quickly in satisfaction scores.

Solution

SupportYourApp brought in two AI agents for calls and chats, kept a live team on email, and routed complex cases from AI to people.

Project knowledge base

Working with Softorino, the team compiled regular client updates, unique procedures, and request-handling guidelines into one functional knowledge base. The AI agents were configured on it. Now, the knowledge base is adjusted as needed when something changes on the project.
AI ticket resolution on calls and chats

SupportResponse handles every chat and SupportVoice answers every call, 24/7. Both were trained on the project's current requests, so they answer from the questions this project's customers actually ask.
Workflows by request type

The team built structured workflows for specific requests and scenarios. Each one sets out how that kind of case gets handled, so recurring requests follow the same path every time.
Escalation to the email team

Complex cases the AI agents can't resolve are escalated to the shared team that handles email. Live support agents cover email around the clock and take over anything that needs a person's judgment.

Achieve faster ticket resolution with AI agents on every call and chat

Talk to an Expert

Results

Once both AI agents were live, Softorino's queues stopped building up between shifts, and the team got more done in each one. Shorter queues helped improve CSAT with SupportVoice and SupportResponse handling the routine load on calls and chats.

85%

CSAT, consistently above the agreed threshold

67%

more requests resolved per shift

~500

cases a month reach the team after AI, down from 800–900+

Bogdan Pecanac

Service Delivery Manager

With seven years of experience in customer support and one year focused on client management, Bogdan is a solution-oriented leader who prioritizes team morale and performance. He relies on analytical insights and strong relationship-building to drive results. Outside of work, Bogdan trades team management for world-building as a Game Master for tabletop titles like D&D, providing his friends with a creative escape from the mundane. When he’s not crafting adventures, he can be found cycling, reading, or taking day trips into nature to enjoy a good book in the fresh air.

Posted on October 30, 2026