An AI-Powered Customer Experience Requires Breaking Down Data Silos
- Posted by Marketing Daitan
- On September 6, 2018
- AI, Architecture, Customer Experience, Data
In 2011, Gartner predicted that by 2020 customers will manage 85% of their relationship with an enterprise without interacting with a human. The convergence of data, devices and technologies like Artificial Intelligence are making the vision a plausible outcome for businesses pursuing Digital Transformation.
Can Artificial Intelligence Truly Meet the Expectations of Today’s Customer?
Today’s consumer expectations cannot be ignored. A customer calling a company on their mobile phone, texting via a message window, or emailing to a support system, expects that company to know instantly who they are without asking. They increasingly expect the company to know their history and the exact status of the product they just ordered seconds ago without even being asked. They expect the company to know whether the repair van is outside their door, or stuck in traffic. Today’s consumers do not expect to have to explain anything. The company should just ‘know.’ AI is believed to be the key to meeting these expectations, and the potential exists—as long as the necessary data is not only accessible, but available in time to deliver the right response.
The Business Reality of Data Silos Creates a Costly Challenge
Fact is, the data exists to do all those things. But it is typically trapped inside multiple disconnected silos that are inconsistently accessible. So, the biggest obstacle to using advanced data analysis is access to the data.
According to Edd Wilder-James in an article, Breaking Down Data Silos, Harvard Business Review, “Since the popular emergence of data science as a field, its practitioners have asserted that 80% of the work involved is acquiring and preparing data. Despite efforts among software vendors to create self-service tools for data preparation, this proportion of work is likely to stay the same for the foreseeable future.”
Data preparation may require many steps to be useful, from translating specific data formats, to handling incomplete or erroneous data. The costs of bad data are high when generating bad or incomplete insights to a business causing it to miss vital information about a customer.
Research by Gartner shows that bad data can cost a typical enterprise more than $13 million every year.
Addressing data accessibility up-front as customer-related events occur and distributing that information is a key component to overcoming the challenge. This means enabling distribution of pertinent knowledge about the real-time status of a customer across departmental silos. When the right information is available to all customer-facing parts of the organization—of course, with proper governance standards maintained— the context about a customer’s current status is more accurate and the whole organization is better able to respond. And, a business’ ability to provide a seamless customer experience increases substantially.
Getting to a Viable Solution
There isn’t an easy solution and most existing systems were architected with the best intentions around privacy and security. Achieving this is a transformative process. As pointed out in the HBR article, it’s hard work that should be approached methodically, progressively in manageable pieces, and fully supported by the C-suite. Because, you must maintain data access to existing business users while implementing changes.
Some of the approaches we have seen work in different scenarios are centralizing applications (not easily done in large organizations, though) and then applying techniques such as Bounded Contexts to allow different departments to have their own particular data attributes while not losing contact with the rest of the organization; building an Event-driven Architecture to distribute relevant events with perfect timing for a meaningful response across the organization; renovating legacy, monolithic applications not built to interoperate with the rest of the company to a more flexible and scalable Microservices Architecture and; deploying Data Lakes (which can be very complex in large organizations) to concentrate raw data for further centralized processing. Such techniques are examples of what organizations can do in order to allow different departments to share data while preserving governance along with their ability to scale independently.
James continues, “To move to the higher value uses (of data) and maintain a competitive edge, we need to lessen the impact of data silos on our businesses. Analyze your business needs, and choose a problem where data (access) could provide a tangible benefit, perhaps in enhancing sales or preemptive incident response. (As change is implemented) each progressive step should build toward an integrated platform for your enterprise data. You don’t want to recreate a whole new set of silos, albeit with advanced capabilities…”
With Useful Data, AI Can Change the Customer Experience
While there is debate as to whether organizations will meet Gartner’s projected milestone of 85% non-human interaction in the next three years, AI-powered experiences have made headway in making the customer’s interaction with an organization faster and more efficient.
It starts with breaking down the silos to make data accessible so AI technology can intelligently leverage the vast data stores accumulated in order to build more continuous experience for customers. For more on this subject, let us suggest reading our latest white paper titled: Smarter Contact Centers Powered by AI. If you have questions or are currently assessing your data management structure, feel free to ping us and let’s discuss.