AI anchored customer service transformation

In this episode of EY.ai Unplugged, Dinesh Mishra, Partner, and Leader – Digital and Emerging Technologies, EY, shares how AI is transforming customer service. He explains how AI can reduce customer service costs by over 50% by automating complex tasks such as collections and onboarding.

Dinesh outlines key success factors, including customizing AI models for quality, minimizing latency, optimizing infrastructure costs, and designing AI-supervised teams. He also emphasizes the importance of integrating AI with existing enterprise technology to maximize its impact.

Key takeaways

Cost reduction – AI can automate high-effort processes, significantly lowering customer service costs.
Model customization – Tailored AI models drive higher quality and customer engagement.
Cost optimization – Strategic use of compute and infrastructure improves efficiency and ROI.
Agentic teams – Structuring AI-supervised agents by task and skill enhances compliance and control.
Enterprise integration – AI systems must integrate seamlessly with existing tech stacks for maximum effectiveness.

For your convenience, a full text transcript of this podcast is available on the link below:

Dinesh: Hi, welcome to EY.ai unplugged series. I am Dinesh Mishra, Partner and Leader, Digital and Emerging Technologies, EY. I will be speaking to you about AI anchored customer services transformation. With advancements in AI tech, you can design to have zero human customer services team today. If you are spending more than 20% of your customer services budget on personnel cost, you are spending too much already.
Many companies have already embarked on programs that will reduce their customer services costs by over 50% for pretty complex tasks, ranging from collections on late payments, optimizing investment portfolios, car purchase processes, customer onboarding into loyalty programs, and more.

When you think of AI anchored customer services transformation, you have to consider the following:

1. Quality of responses that will drive engagement and adoption

You need custom training of models be it LLMs, SLMs, specialized models or quantitative language models as adoptions evolve. This will determine the depth and robustness of contexts the model has, which in turn will determine the quality of responses. This is where a lot of time and effort will be spent by companies outside of the core tech companies that are churning out these language models. Think of it as software packages that still need to be tailored to be successful in a specific deployment scenario.

2. Optimize for speed— speed will drive efficiency and natural interaction

Latency of responses has to be minimal and as natural as engaging with the human. The biggest factors determining AI response latency are hardware, mainly GPUs, model size, query length, network speed and inference optimization.

3. Optimize for cost

This drives your speed of scale and your return on investment. You have to optimize for queries, consumption and compute. For doing this, the language model and version used databases and sources, and compute infrastructure have to be optimized. You have to balance cost, quality and performance for your specific use cases. The solution you consider must have the ability to query the fit for purpose data source using the most efficient language or data model for that query.

4. Design agentic organizations

Take inspiration from human forms of organization that will most naturally fit into the rhythm of your overall enterprise. AI agents are organized by common task and hence skillset pool grouped into pods supervised by AI for managing agents, further supervised by functional AI managing agents, each having appropriate access rights, security clearances and oversight or override rights. There will be control, compliance and performance management agents that will be constantly checking for compliance and performance, and a human or humans monitoring all of this in real time on dashboards.

So, to summarize, AI anchored customer services transformation will need to consider model contextualization, latency minimization, cost optimization, and design agentic organizations —all in one comprehensive solution.

None of this is going to be plug and play from a tech perspective, at least at scale and with enterprise readiness. There will be further value to be added by helping businesses leverage as much of their existing tech investments in a bring-your-own-technology model, or BYOT, where the AI orchestration layer, will sit on top of all existing tech from Microsoft, Salesforce, SAP, Oracle, Amazon, Google and the like to achieve business outcomes and objectives.

This will be the new and critical role that business and technology consultancies will need to play for all businesses undertaking AI anchored services, functional or enterprise-wide transformation. Thank you.

If you would like to listen to our podcasts on the go:

Amazon music

Spotify

SoundCloud

May 2025

Podcast

Episode 18

Duration

4m 52s

Customer experience

The EY Customer Experience solution can help your business integrate CX programs that help deliver sustainable long-term value. Discover more.
Read more
Artificial Intelligence Consulting Services

Our Consulting approach to the adoption of AI and intelligent automation is human-centered, pragmatic, outcomes-focused and ethical.
Read more
Technology transformation

Discover how EY's technology transformation team can help your business fully align technology to your overall purpose and business objectives.
Read more

In this series

EY refers to the global organization, and may refer to one or more, of the member firms of Ernst & Young Global Limited, each of which is a separate legal entity. Ernst & Young Global Limited, a UK company limited by guarantee, does not provide services to clients.

Insights

Highlights

Services

Spotlight

Industries

Case studies

Careers

Spotlight

About us

Top news

AI anchored customer service transformation

Key takeaways

In this series

Presenters