Generative AI Will Change Your Business. Here’s How to Adapt.

In our last HBR article, “Customer Experience in an Age of AI,” we discussed how one’s personal information could already differentiate branded experiences. With generative AI, the personalization of digital interactions will be even more tailored to what customers want and not how designers imagine it. As the software follows the user, it will take them to places beyond the boundaries of the brand’s products. It must offer solutions for what the customer wants. Help someone with the whole package and guide them on their journey, even if that means partnering with outside partners, rethinking your definition of offerings, or developing the underlying tech and data architecture to connect all the solution components.

Generative AI can generate text, speech, and images. It can also produce music, videos, code, and even video. This capability, when combined with the ability to use someone’s information to customize the what, when and how of interactivity, will make it easier for people and increase the accessibility of the software. The simple question box at the heart of Google and most AI-generated systems, such as ChatGPT or DALL-E 2, will power more systems. Drop-down menus and their inherently limited functionality are no more. You’ll see “What do I want to accomplish today?” and it’ll offer suggestions based on what you have already done in the past, what the system knows about the current context, and what your core goals are. For example, “save money for a vacation,” “remodel the kitchen,” or “manage my family’s meal plans with special dietary requirements.”

Consumers will only care if the software is limited if it has the limitations of the traditional software interface. Changes in how we interact and what we expect will be drastic and democratizing.

The hype around generative AI is centered on its ability to create text, images, and sounds. But it can also generate code to automate action and facilitate pulling external and internal data. It allows users to take a shortcut from a given command to a completed step by generating code. There is no need to go through the entire software menu. You can easily ask questions and analyze the data in an application by asking questions like: “Whom have I not called for the past 90 days?” and “When am I scheduled to be in NYC?”. The query is recognized, a code is created, options are ranked, and the best solution is generated in milliseconds.

It simplifies the way we interact with applications that we are familiar with. This allows more brands to create applications as part of their value proposition. Developing these capabilities is possible based on the weather, traffic, and group of people. Powerful tech companies? Brands that already have relationships with their target categories? New disruptors with a specific focus? Although the game is just beginning, business models and capabilities are already forming.

The Broader Journey, the Broader Boundaries

Building your offering in a world of generative AI and other AI systems that are constantly evolving requires you to take the most comprehensive view possible of your data pool, the journeys. The risks it can facilitate, and how much you want to enable.

Bring together data.

To meet your customer’s needs, you must pull information from across your organization and possibly beyond. Data integration is among the most challenging tasks for many applications and IT departments. AI systems can write code to understand schemas from two databases and combine them into a single repository. This can save a lot of time in standardizing data schema. AI teams must still dedicate time to data cleansing and governance, for example, defining the correct definitions of data features. With AI capabilities, the following steps to combine all the data become more accessible.

Narrative AI offers, for instance, a marketplace to buy and sell data. It also provides data collaboration software, which allows companies to import their data into their repositories with just a single click, in alignment with their schema. Data from all over a company can be integrated and then used to model in seconds.

Combining your data with data from public sources, other AI tools, and data from external parties will improve AI’s understanding of context, ability to predict the question, and have more data to choose from to execute a request.

However, the old rule “garbage-in, garbage-out” still holds. It is crucial to check the accuracy of third-party data with the internal data before it’s integrated into the data set. One fashion brand found, for example, that data on gender purchased from a 3rd-party source did not match their internal data in 50% of cases. So, reliability and authority are essential.

This layer is critical.

The AI must have guidelines to ensure that things beyond its capabilities or inappropriate do not influence it. The need to focus on the rules layer, where experience designers, marketing, and business decision-makers define the parameters the AI should optimize, is heightened.

We set rules for an airline brand using AI to determine the “next-best conversation” with its customers. These rules included which products would be marketed, which copy to use, and which countries could use it.

In the age of generative AI, these constraints are even more critical. Customers will quickly point out when a machine “breaks” and produces nonsensical results. Best approaches start minor and are tailored to specific solutions. Rules can be defined precisely, while human decision-makers can design laws that cover edge cases.

Deliver the entire journey and the specific use cases involved.

Customers will ask for exactly what they want and seek the easiest and most cost-effective method. What is the natural end goal of your customer? How far can your brand go? You can create partnerships to share data and execute actions that will help customers on their journey. Your ecosystem of business relationships is what will make your brand stand out.