Customer service · Telecommunications and services companies
AI chatbots for customer service in call centers
An agent that understands what the customer actually wants, resolves the repetitive part, and hands off to a person with the history already loaded.
Team experience: a project delivered by Possition engineers as part of a Globant team. It is not a Possition client engagement; we show it because the same people and the same practice work on your project.
- Client
- Telecommunications and services companies
- Sector
- Customer service
- Product
- Sofía
In numbers
The initial problem
High interaction volume, repetitive inquiries, wait times, customer frustration, and rising operating costs.
The previous process
Service depended on agents for the full volume, including repetitive inquiries, with rigid IVR flows that did not understand each customer's real intent.
The solution implemented
Conversational agents and voice AI with Dialogflow and CCAI: intent understanding, session handling, mission playbooks, handoff to humans, and flow redesign so the bot is not just an IVR with a voice. The design included humanization tests, dialect variants, approval criteria, a war room, and metrics defined from the start, validated with thousands of user interactions.
How we did it
Map of real intents
We analyzed the support volume to separate what can be resolved on its own from what needs a person, and defined playbooks by mission instead of an options tree.
Conversational and voice design
We worked on humanization, session handling, and dialect variants so the agent understands how the customer speaks, not how the script was written.
Handoff with context
When a query goes beyond the agent, the conversation passes to an operator with the history and data already loaded, without asking the customer to repeat everything.
Testing, metrics, and war room
We defined approval criteria and metrics before going live, and validated behavior with thousands of real user interactions.
Integrated systems
- Dialogflow / CCAI
- Telephony and IVR
- Digital channels
The result
Automation of repetitive support with intent understanding, intelligent handoff, and 24/7 coverage for telecommunications and services.
What changed in the operation
- Repetitive queries are resolved without occupying an operator.
- The customer doesn't repeat the problem when a person takes over.
- Support stays available outside call center hours.
- Agent behavior is measured with criteria defined from the start.
Frequently asked questions about this case
How is this different from an IVR with a voice?
An IVR forces the customer to pick options. Here the agent interprets the intent behind what the person says and acts accordingly, even if they don't use the script's words.
What happens when the agent can't resolve it?
It hands off to a human operator with the full conversation history and validated data, so the person picks up where it left off.
How do you measure whether it works?
With metrics agreed before implementation: resolution without human intervention, correct handoffs, and answer quality over real interactions.
Keep exploring
Other work by the team, across different sectors.
TourismMirstravel: an AI agent that answers, quotes, and hands off travel inquiries on WhatsApp
Mirstravel answers and quotes travel inquiries on WhatsApp at any hour; the advisor gets each inquiry already organized in a panel.- 24/7 answers on WhatsApp
- Handoffs with the full record
- Scheduled broadcasts
Professional servicesAI assistants over WhatsApp for firms and professionals
One assistant per practice that takes the initial query, classifies the case, requests documents, and schedules before professional involvement.- One assistant per practice
- Configured by province
- Scheduling and reminders
FinanceArcos Capital: financial data traceability and automated reports
Financial reports generated automatically from Arcos Capital's sources, with every figure traced back to its origin.- Automated reports
- Every figure with its source
- Reconciliation and alerts
Tell us what you need to solve
In a 30-minute discovery call we review the process, the systems and the numbers, and tell you what's worth building, what isn't, and what return to expect. No cost, no commitment.