How to harness the power and the beauty of generative AI

It is hard to believe that when OpenAI launched ChatGPT in November of 2022, it received little fanfare. Less than a year later, generative AI is upending business models, processes and the way we work. 

Generative AI, with its remarkable ability to analyze vast datasets and autonomously generate valuable, human-like insights and content, has the potential to be a powerful operational excellence tool. In fact, a recent report by Mckinsey estimates that generative AI could automate 25 percent of workplace tasks and, when combined with other enabling technologies such as workflow automation, could potentially automate “work activities that absorb 60-70%of employees’ time today.”

According to PEX Report 2024, operational excellence leaders are eager to start experimenting with generative AI but only a small percentage, about 19%, have successfully operationalized at least one project. As one anonymous survey respondent put it: “AI can help with pattern recognition and statistical analysis involved with process mapping and mining to determine opportunities for optimization. There is fascination with ChatGPT’s possible use cases for call center knowledge management and customer-facing chatbots, but I do not think we currently have the risk tolerance for that.”

With this in mind, we could not be more excited to invite you to attend All Access Generative AI in Operational Excellence taking place on November 15, 2023. To give you a preview of what you can expect to learn about, we invited speaker Ulrich Irnich, the CIO of Vodafone Deutschland, to share his thoughts on how generative AI is poised to revolutionize operational excellence. 

Before you read his Q&A, don't forget to register for his fireside chat on Improving service excellence through generative AI at taking place at 9:30am EST on November 15. 

PEX Network: How do you think generative AI will impact operational excellence? 

Ulrich Irnich: It will have an especially big impact on how we interact with customers. For starters, it will help us understand customer intent much faster. Additional use cases in the operations space include change management and incident management. 

PEX Network: Can you share examples of successful implementations of generative AI for operational improvements? 

UI: We implemented generative AI to summarize chatbot chats with our customers and give our agent a fast kickstart to understand the intent in 3 sentences. This is a massive change.

Secondly, we can now produce a complete digital replica of many of Vodafone’s internal support functions using artificial intelligence and advanced analytics. Called a digital twin, it enables analytic models on Google Cloud to improve response times to enquiries and predict future demand. The system will also support a digital twin of Vodafone’s vast digital infrastructure worldwide.

PEX Network: What are some of the biggest barriers to adoption?

UI: Cloud and data are essential to unlock the full potential of AI. We are now in a perfect storm as data, compute power and connectivity are all adding up so that the beauty and power of AI are now being realized. 

The science on AI is already more than 65 years old and, with this perfect storm, it is now possible to operationalize algorithms and machine learning. Having all analytical data in one, big data ocean provides new opportunities for growth hacking and new product innovations. The combination of open APIs, experience layer, data and deep learning algorithms open a new era of customer experience.

PEX Network: What advice would you give to organizations that are considering adopting generative AI for operational excellence but are unsure where to start?

UI: In my view we are in front of three main challenges: to start, deficits in AI education and fear of losing jobs. Secondly, the maturity of data management and cloud readiness is lacking. Thirdly, many companies are still operating in silos and AI requires a holistic approach. Therefore this experimentation should be open, transparent and start with small use cases that your organization can understand. Before you start, make sure you understand why you are using generative AI instead of another solution. 

I would also be remiss if I did not mention AI hallucination, when an AI model generates incorrect information but presents it as if it were a fact. Close monitoring of usage and process supervision, a method that rewards each individual step of a task instead of outcomes, may be helpful in mitigating this issue. In addition, generative AI should not be used on internal company data.

PEX Network: How do you see the future of generative AI evolving in your industry in the next 5-10 years? 

UI: Generative AI will be everywhere. It will be fully integrated into all interactions, process excellence, new AI process modeling and, eventually, will evolve into vertical specific languages.