K2view GenAI Data Fusion Enables Pelephone to Cut Costs by Reducing Call Handling Times

The first to offer mobile phone service in Israel in 1985, Pelephone quickly became a household word synonymous with cellphone.

Dedicated to innovation and excellence in customer service, Pelephone sought to leverage generative AI (GenAI) technology to achieve greater customer service efficiency and improve the customer experience, by improving operational KPIs, such as first contact resolution rates and average handling times.

To embark on the GenAI journey, Pelephone established an integrated, cross-functional team led by CIO Maya Bachar Gilad, that included members from customer care, data security, and IT, as well as AI, data, and software engineering.

The team quickly realized that to improve the operational KPIs via GenAI would require them to provide their enterprise Large Language Model (LLM) with real-time access to fresh customer data from different source systems, like CRM and billing.

The team identified 4 critical success factors that had to be met:

Cost savings and improved customer experience

With customer satisfaction top of mind, Pelephone wished to improve first contact resolution and reduce the time spent waiting on the line, resolving problems, and onboarding new agents.

Privacy and security

Pelephone demanded that its customer data be protected according to the highest standards of privacy and security.

Scalability

With millions of customers, served by more than 1,000 customer care agents, the ability to seamlessly and cost-effectively scale the solution was crucial.

Prevention of vendor lock in

The selected GenAI framework would need to work with multiple large language models. Pelephone chose Anthropic’s Claude for this project, due to its superior multi-language support, but wanted to be able to use other LLMs as needed.

K2view was selected for the following reasons:

Data privacy guardrails and security

K2view provides a semantic data layer, optimized for conversational GenAI, enabling:

Its data agents to work with secure customer data
The masking of all sensitive data
Exclusive access to the data relevant to the active customer session

Data freshness

Pelephone needed access to up-to-date customer data to resolve customer issues accurately and quickly. For example, a subscriber calling about an Internet outage might require current data from the network, as well as billing, invoicing, and subscription plan details. K2view’s AI data agents retrieve customer data directly from the source, even when it’s fragmented across systems, without overloading the backend systems.

Split-second data retrieval

K2view supports multi-source customer data retrieval in near real time, enabling a GenAI-powered chatbot to operate at conversational latency of less than 200 millisecond response, and at a scale of thousands of concurrent queries.

Platform familiarity

Pelephone had been working with the K2view Data Product Platform for years and knew that it could be trusted at scale with the security and privacy guardrails required for customer-facing workloads. The same tool Pelephone uses to build data services and data pipelines in other use cases like Customer 360, can now be used to create GenAI-powered chatbot.

To gain the trust of the organization, Pelephone proposed a 3-phase GenAI implementation, starting with a human Customer Service Representative (CSR) in the loop:

Screenshot 2025-05-14 at 15.45.36

Rep Assist provides CSRs with the most likely reasons for a customer call – generated by AI during IVR routing, before the call is connected – along with the ability for the agent to get personalized answers to any customer questions in seconds.

The sample CRM screen below showcases the GenAI-powered chatbot embedded into the CRM system – with auto-generated customer insights on the left, and a CSR chat box on the right.
Customer Assist, the next phase, is where a GenAI-powered chatbot communicates directly with the customer, without a human agent in the loop.
Autonomous Agents, the final frontier, make decisions and complete tasks proactively and independently.

For example, knowing that Internet service would be interrupted at a certain time and location due to planned maintenance, autonomous agents would identify affected subscribers and send them a series of text messages and emails, leading up to the event, informing them well in advance of the date and time of the planned outage. The agents could also calculate and apply an appropriate credit to affected subscribers, based on the amount of time that the service was down.

Such proactive measures by autonomous agents would prevent a large volume of inbound calls and digital contacts, while also ensuring customer loyalty and trust.

With the first phase in production, Pelephone and K2view are closely monitoring performance KPIs. When all 3 phases are implemented, Pelephone will be among the first to reap the full benefits of AI for customer service.

Overview

Capabilities

Architecture

Initiative

Industry

Company

Reach Out

News Updates

Education & Training

Resources

Demo

K2view GenAI Data Fusion Enables Pelephone to Cut Costs by Reducing Call Handling Times

Product

Resources

Maya Bachar Gilad

Chief Information Officer, Pelephone

Background

Key Challenges

Cost savings and improved customer experience

Privacy and security

Scalability

Prevention of vendor lock in

Solution

Data privacy guardrails and security

Data freshness

Split-second data retrieval

Platform familiarity

Methodology

Experience agentic AI

Go behind the scenes and see the process of grounding AI agents with enterprise data

Explore more stories