Case study: automation in a contact center cut handling time by 30%

Starting point: a contact center at the edge of capacity

The protagonist of this case study is a mid-size contact center that serves several services-industry brands. The team is about 40 agents, working shifts in a 7-day-a-week model. The main contact channels are: phone, email, web forms, and site chat.

At the start of the project, the team was facing typical problems:

• long average handling time (AHT) for simple cases – up to 6–7 minutes,
• peak-hour call queues reaching dozens of customers,
• a large volume of repetitive questions about case status and basic information,
• no consistent ticketing system – some cases were tracked in spreadsheets, others "in agents' heads".

Leadership knew that without customer service automation and better process organization, the contact center would have to keep growing headcount. So the project goal wasn't only to raise quality, but also to deliver concrete time, resource, and money savings.

Rollout scope – which processes did we automate?

Together with the customer, we mapped the key areas where contact center process automation could deliver the fastest results:

• Call automation – an intelligent IVR and a simple voicebot to handle the most common questions.
• Email request handling automation – the Debesis ticketing system, categories, SLAs, and reply templates.
• Automatic notifications – email/SMS at key moments in the process (request acknowledgment, status change, case closure).
• Integrations with back-office systems – CRM, billing system, and customer portal.

Call automation – from IVR to voicebot

The first stage was rebuilding the helpline. The previous voice prompt boiled down to a "wait for an available agent" message. Every call therefore went to a person, even if the customer just wanted to check a balance or case status.

We designed a new scenario in which call automation covered:

• an IVR menu routing calls to the right queues based on topic,
• integration with the CRM system – the system identified the customer based on phone number,
• a voicebot handling the most common questions (balance, payment due date, case status).

As a result, 23–27% of calls were fully handled by automation, without an agent involved. For more complex cases, the voicebot passed the call along with a note to the right queue.

Ticketing system and customer service process automation

In parallel, we rolled out the Debesis ticketing system, which became the "single source of truth" for all cases. Every call, email, or customer form turned into a ticket with a number, status, and an owner.

The most important elements of customer service automation in this area:

• rules routing requests to the right queues (sales, billing, complaints),
• automatic case acknowledgments with stated wait times,
• reply templates for repetitive questions, populated with CRM data,
• escalations – if a ticket doesn't get a reply within a defined time, the system informs a supervisor.

Automations that speed up customer service – concrete examples

Three types of automations worked particularly well in the project:

1. Automatic call summaries on the ticket

After each phone call ended, the system automatically appended the recording, customer data, and basic contact parameters to the ticket. The agent only added a short note. The time saved on each case was about 30–40 seconds.

2. SMS reminders instead of phone follow-ups

Some calls were about reminding customers about payments or missing documents. We set up rules in which the system automatically sent SMS or email reminders a few days before the deadline. The number of inbound calls in this category fell by over 40%.

3. An intelligent priority queue

Previously, agents saw all cases in a single list. Now the system orders the queue by priority, wait time, and customer value. As a result, agents handle critical requests first, rather than the ones that "happened to land at the top of the list".

Results – savings of time, resources, and money

Three months after rollout, we measured concrete results of the contact center process automation. Below are the key numbers (an average over three consecutive months, compared to the baseline period before the project):

On top of that, the repeat-contact rate for the same case (FCR – first contact resolution) improved by 9 percentage points. Customers rarely have to call again, which immediately translates into less helpline load.

What was key to the customer service automation success?

Customer service automation isn't about "throwing a bot on the helpline". In the project described, three elements ensured success:

• Thorough analysis of contact reasons – before designing automations, we collected a list of the 20 most common reasons for calls and requests.
• Simple start, fast iterations – we started with a few IVR scenarios and the simplest templates, and only then added more layers of automation.
• Team buy-in – agents had a say in the wording of messages, templates, and rules. As a result, customer service process automation was perceived as support, not a threat.