Want an AI-enhanced customer experience? Start with Your Proprietary Data

As businesses navigate the frontier of customer experience (CX) whilst grappling with the myriad changes brought about by generative AI, they are gradually coming to terms with the increasingly important role of enterprise-grade AI.

Renowned research firm, Gartner, predicts that 80% of customer service and support organisations will apply generative AI in some form to improve agent productivity and CX.

Furthermore, a recent poll by the same firm found that 38% of leaders see improving CX and retention as the primary focus of generative AI investments.

So clearly, the enterprise expectations for AI-enhanced customer experiences are high.

However, there is a gap between these aspirations and actual realisation.

While there are no doubt different ways to solve this problem, at Toku, we have a distinct perspective on the matter:

We’ve constantly advocated for enterprises to understand their own business data and leverage it to elevate their customer experience.

We touched on the importance of this in an earlier blog post, How Data-Enriched Communication Channels Can Deliver Empathy and Efficiency at Scale.

Now that we’re talking about enterprise-grade AI for CX initiatives, the need for data, in particular proprietary data, is becoming much more pronounced.

Recently, I had the privilege of attending the CCAS 19th Regional Contact Centre Symposium where our Founder and CEO, Thomas Laboulle and Head of AI, Isuru Rajakaruna, shared their thoughts on the promise of AI in business, its current challenges, and real-world examples of how enterprises can use AI to transform their customer experiences.

Watch the full speech below.

In this article, I’ll share some key takeaways from their enlightening session.

Before diving in, let’s address a common misconception.

ChatGPT is not the beginning of AI

AI, as a term, is broad and encompasses a variety of technologies.

While recent breakthroughs in generative AI, such as ChatGPT and Bard, have garnered significant attention in the public consciousness, they represent just one facet of the AI universe.

It’s essential to recognise that the debut of generative AI models doesn’t mark the inception of artificial intelligence.

The journey of AI can in fact be traced back to the 1980s when several of the core concepts used in today’s deep learning applications were introduced.

These early endeavours laid the groundwork for multiple innovations.

Notably, in 2016, AI played a crucial role in optimising the cooling of Google’s temperature-sensitive data centres, enhancing efficiency by 40%. Just two years later, in 2018, Google DeepMind’s AlphaZero achieved world-class performance in chess, underscoring the diverse potential of AI applications.

Three ways that AI is set to transform CX

The current slew of generative AI tech promises to bring sweeping changes to the CX industry.

Why? Because generative AI can accomplish three important things.

1. Efficiency in Customer Support: AI can swiftly resolve straightforward enquiries, allowing human agents to dedicate their attention to more intricate issues. This means customers receive timely solutions for simple queries, while more complex concerns are addressed with the depth and nuance they require via the human touch.
2. Scalability of Operations: Beyond efficiency, AI offers scalability by streamlining and automating business processes and operations. As a result, businesses can expand operations smoothly without incurring additional costs, laying the groundwork for sustainable growth.
3. Personalised Experiences: By continuously learning from new linguistic data, AI can adapt and refine its interactions to align with individual user preferences and nuances. This means every communication can be customised, ensuring it resonates on a more personal level. Such bespoke interactions can foster greater brand loyalty, as customers appreciate and value an experience that feels uniquely personalised to them.

AI challenges in CX for enterprises

AI might have opened up numerous opportunities for CX leaders to turn their customer experience into the profit centre of the business.

However, it has also introduced some new challenges:

1. Unreliable Outputs

A notable example in 2023 underscores the pitfalls of AI’s unpredictability. A lawyer utilised ChatGPT for legal research in a personal injury lawsuit. Tragically, the AI tool generated fictitious cases that the lawyer unknowingly presented in court. Such instances of AI “hallucinations” not only mar the reputation of professionals but also jeopardise the very integrity of the brand.

In a customer support scenario for example, AI tools might provide incorrect or misleading solutions, resulting in customer dissatisfaction and potential loss of trust in the brand.

Get monthly nuggets of wisdom for all things customer experience in your inbox

2. Model Bias

Models like Text-to-Image Generative Models exemplify the inherent bias in AI systems. Analysis of its outputs reveals amplified stereotypes, with racial and gender disparities more exaggerated than real-world biases. For instance, it often portrays women of colour in menial roles while white men are predominantly depicted as CEOs.

Think about the implications in customer engagement – left uncontrolled, AI could push stereotyped content or offers based on biases, alienating diverse customer groups and perpetuating harmful, biased narratives that drive a wedge in society.

3. Regulatory Compliance

The rapid proliferation of generative AI applications poses a regulatory conundrum. With no global oversight framework, multinational companies will have to grapple with fragmented rules across markets for the foreseeable future.

Especially when using generic public data-based models, enterprises will find it challenging to ensure compliance, because they do not have control over these models. For example, AI has immense potential to scale personalised marketing, but companies may inadvertently breach data privacy regulations across different regions because the AI model API or the open-source model they’re using failed to comply with rules in certain countries.

This could lead to legal ramifications, damaged customer relations, and in turn, a tarnished brand image.

Enhancing customer experience with proprietary data-based AI models

Proprietary data is going to be the black gold for enterprises as they start using AI for improving CX.

Here are some reasons why we believe it offers an unprecedented opportunity for businesses to gain a competitive edge.

1. Superior accuracy

When AI models are trained on business-specific data, the precision and relevance of their outputs are significantly elevated. The use of generic models, especially in the case of generative AI, often falls short in delivering contextually accurate results.

In comparison, AI models, such as ALPACA (A Large-scale Pooling Architecture for Content Analysis) which is based on LLAMA (Large Language Model Architecture) and further fine-tuned with use-case orientated data, excel in accuracy.

2. Tailored customisations and continuous improvement

Proprietary data ensures that AI solutions are tailored to fit the unique processes and requirements of a business. One of the inherent limitations of generic models is they cannot always accommodate the dynamic needs of businesses because they lack the context in which the company operates.

In addition, proprietary data allows businesses to be more agile, by continuously retraining models incrementally. This way, businesses can leverage a continuous improvement cycle, ensuring that their AI solutions remain relevant and effective over time.

3. Upholding privacy and security

As data regulations become more stringent, the sustainable use of proprietary data requires that businesses maintain compliance with the guidelines developed from country to country. By securing customer data and employing bespoke AI solutions, companies can ensure that their AI-driven initiatives are both compliant and within the bounds of emerging regulations.

AI in action – Real-world breakthroughs in customer experience

Let’s explore some really exciting applications of enterprise-grade AI for improving customer experience. You might already have noticed some of these at play in your contact centre operations!

1. Improving agent training and coaching

Advanced Agent Training with Simulators: Traditional training methods for agents can only go so far. With AI-driven simulators, agents are exposed to a myriad of scenarios, enabling them to navigate complex situations and refine their problem-solving skills. This immersive training experience fosters adaptability and ensures they are prepared for any challenge.

Performance Benchmarking: AI offers granular analytics, facilitating the comparison of agent performance against established benchmarks. This not only promotes a culture of excellence but also helps identify areas necessitating further training or resources.

Automated Feedback Generation for Agents: Continuous feedback is vital for growth. AI systems can evaluate agent-customer interactions and generate actionable feedback, expediting the learning curve and fostering a culture of constant improvement.

2. Harnessing customer insights and analytics

Automated Detection of Conversational Events: AI can meticulously scan conversations, pinpointing areas where agents might miss the mark, such as inadequate greetings or seeking irrelevant information. By highlighting these events – and escalating them to supervisors in real time – businesses can take corrective measures, elevating the overall service quality.

Talking to Data, Instead of Analysing It Manually: Sifting through customer data is now effortless with AI. For instance, businesses no longer need to rely only on standard matrices. Instead, they can model custom scenarios and the analytics engine can then fetch the answers for them directly.

Agent Performance and Rating: By analysing myriad interactions, AI provides a comprehensive view of agent performance. This not only aids in recognising top performers but also in understanding areas of improvement.

Capacity Planning and Prediction: AI’s predictive analytics empower businesses to anticipate call volumes, adjust staffing needs, and optimise resource allocation, ensuring that customers aren’t kept waiting.

Start using proprietary data for AI-enhanced customer experience

As AI technologies become more and more accessible, it’s the rich, proprietary data owned by businesses that will make or break your CX strategy.

Why? Because proprietary data is the bedrock of a bespoke AI for CX strategy (which for reasons discussed in this article, is the only feasible strategy for most enterprises for the foreseeable future).

In our experience, businesses are already sitting on a goldmine of unique data, accumulated through years of interactions and analyses. However, not all can harness its transformative power.

Some have knowledge gaps when it comes to AI, which presents a golden opportunity for businesses to leapfrog the competition by partnering with AI consultants and experts.

They can refine your AI models to be not just effective, but bespoke, perfectly aligning with both company objectives and customer expectations. The future of enterprise CX hinges on this customised approach, making AI not just a nice-to-have, but an integral, transformative force in how you engage customers.