AI customer experience: Doing AI versus actually being an AI organization

There’s always been a subtle difference between “doing” and “being,” which is important in the context of AI customer experience.

“Doing” digital things means random acts of investment here and there in websites, mobile apps, databases or even business automation platforms that allow for multiple functions to use digital tools.

Once “doing” becomes second hand, the next step might be toying with advanced digital solutions to accelerate automation or points of personalization.

“Being” digital means these tools and channels, the automations and digital systems, become the backbone of operations and data. High-fidelity customer signals get dropped into the center of decision-making processes.

Being digital means embracing lots of change.

Digital organizations had to fundamentally shift operations, reorient teams, and bid farewell to old processes in exchange for digital workflows and operational strategies that delivered top line growth and bottom-line efficiencies and savings. This is the difference between doing and being. And yes, change – and the appetite for it – is often the biggest challenge.

Now we find ourselves in the throes of another massive shift as we enter this age of AI. Here’s the truth nobody wants to say out loud: If you decide to BE an AI organization, you will fail if you don’t change.

That’s what we’re here to discuss: what to change first.

The chicken and egg debate

When it comes to realizing the benefits of AI customer experience strategies across sales, service, commerce and marketing, which comes first: The Platform or the Data Egg?

In CX circles, plenty of players will do many things with AI. There will be bursts of use cases and wins thanks to automation and the capacity to create and generate personalized content and assets at scale.

But for those organizations that choose to not just do but actually BE an AI-powered CX organization, other considerations must be made:

• Is the generative AI being discussed available or it is planning to be a beta sometime soon?
• Is the data available and ready to further train and enhance the AI model’s understanding of the business and of the customer?
• Is the AI fit for general enterprise use?
• Is it trained to focus functionally?
• Is it constrained to functional walls or is it empowered to connect across the business to truly make a difference for the customer and, in turn, for the bottom line?

While the egg and chicken debate will rage on, in the case of AI and customer experience, the answer feels a bit easier to navigate because history has shown the platform and the composability of that platform must be established first. Otherwise, there’s no place from which models can pull data, let alone create paths to high-fidelity signal across our businesses, our ecosystems and our customers.

Without a solid platform and framework for the workflows and automations, it will work for a brief, glorious moment, but then quickly begin to give way under the pressure.

AI customer experience: 3 questions to ask

Here are three questions every organization should ask when navigating this journey across CX in search of becoming that AI-empowered enterprise.

1. Composability: Do the tools and solutions across the entirety of the CX ecosystem flex and connect to establish a holistic foundation for today and tomorrow’s CX delivery?
2. Access: Are digital dams creating unintended data droughts?
3. Availability: Are AI-empowered processes available today, or are they a promise for a later day?

These three questions are intertwined to the point that being an AI enterprise demands all three be addressed.

Intelligent customer experience: Definition, benefits, examples

Your business – connected, insightful, and adaptive: Discover the power of intelligent CX.

Composability, CX, and AI

This isn’t about the composability or connectivity of a handful of components. The modern customer journey can’t afford loosely connected tools hoping that APIs can save experience. This is a question about the fundamental architecture on and across which we intend to build our CX delivery systems.

While it may have been suitable in the past for tools to sit side by side, with casual handoffs of workflows connecting functions like sales and service to commerce, when we add AI and the demands of generative AI on top of these loose connections, the house of cards falls down.

Platform composability will be key to the operational success of CX’s ability to reach beyond the constraints of functional tools that only optimize the experience of that single function. Architectures that anticipate the scalability and reusability of assets don’t stop at the often-parroted mantra of “one and done.” They go beyond to expect that the one application or asset created isn’t just shared, but accelerated and optimized as it’s reused and repurposed.

Composable frameworks allow organizations to lean into modern tools for workflows and automation without being held back by legacy complexity or customizations.

AI depends on data access

What used to be thought of as “functional silos” have transformed into digital dams, blocking the flow of data across organizations, making it impossible for AI to consume what it really needs. AI doesn’t just thrive on data; it quite literally needs data to survive.

From training large language models leveraged for generative AI to the AI algorithms powering recommendations, data sits at the center of everything. What used to suffice as “good enough for machine learning responses” simply don’t satisfy most organizations’ threshold for acceptable response, let alone satisfy the customer’s demand for accuracy and context.

Imagine a customer visiting a chatbot that promises to share updates on that customer’s most recent order. If that chatbot can’t seamlessly connect to multiple commerce, supply chain, product and back-end ERP solutions, the answer will be limited and the experience meaningless.

Today’s customer expects the bot to know everything from availability of product to exact location of the shipment and anticipated arrival time. This expectation demands that the dams, especially those that have been unintentionally erected between functional tools, be demolished, or at minimum, cracked to let that water known as data spill out.

---

---

A beta in need of data: Is AI available now or just promises?

The blunt reality of many of the generative AI tools touted throughout 2023 is that they were promises: great experiments of use cases of AI model applications. Simply, they are a beta in need of data.

This promise of AI often hinges on a vendor gaining access to enough data to train the models properly and appropriately. In the race to take advantage of commercially available models like OpenAI’s ChatGPT, questions around ethical use, data privacy and safety, and even accuracy were set aside in the name of innovation.

But now, as organizations set their sights on the impact, results, and effectiveness of these tools in action, new questions are quickly emerging asking if teams and customers are truly better off with these new solutions. Once again, it’s important to weigh if an organization is going to BE an AI-empowered organization or just offer a few workflows, automations or experiences that are incrementally better with advanced AI models and applications.

For example, in the case of AI in sales solutions, we need to consider whether sellers are more effective and efficient with AI tools, or if they’re just faster in one aspect of their jobs. In order to truly transform the work of selling, AI tools for sales must have the composable architectures that enables connection to cross-enterprise systems, bringing the data needed by these AI models closer to the seller’s work and workflows.

If the data from ERP can’t be brought closer to the data from CRM, AI tools will be unable to crunch and crank to identify friction or opportunity.

But – and this is a big but – if you make the choice to BE an AI enterprise, you’re also making a commitment to be a data organization. The two march hand in hand. So the real question is: have you established solid ground on which both CX and AI don’t just walk, but run?

This is where the question of composability rises to the top and focuses our answer that yes, the platform does need to come before the data-shaped egg.

AI for customer service: Faster fixes, happier agents

AI for customer service can improve the agent experience, speed resolutions, and boost customer satisfaction.

AI customer experience: Rising to the top 

How is this becoming reality? SAP is an example of a vendor that has made this hard turn, and in the course of doing so has become an AI organization. The first step started several years ago when the entire CX portfolio was unpacked, rearchitected, and relaunched. The decision was to ensure the composable architecture needed for AI was ready to act in the service of CX.

Rebuilt from the ground up to purposefully empower data, workflow, and automations to work in the service of selling as opposed to the functional silo of sales, SAP Sales Cloud is focused on empowering selling to happen anywhere across the organization while also helping sales teams to engage far more effectively and contextually with their customers.

Similarly, SAP Service Cloud is focused on how exceptional service based on the customer’s context can be delivered anywhere across the organization, integrating data that comes from anywhere across the customer’s journey.

Because of this dedication to composable tools on composable architectures, the acts of selling and servicing aren’t limited or constrained. But more importantly, these tools won’t need a massive overhaul to onboard new innovations in AI and intelligence.

This is why when SAP’s CEO Christian Klein announced a massive investment into AI and said AI was far more than hype for SAP, but was actually going to redefine how work is done from finance to selling, many of us in the analyst world weren’t surprised.

In reality, SAP becoming an AI enterprise itself has been on the roadmap for years, even if that wasn’t the articulation. SAP had to rebuild itself, rebuild the SAP cloud, and fully commit to composability as a strategy to develop a far more solid, flexible and agile foundation.

Without that shift, any move towards AI would only represent doing; it could never constitute being.