AI Conversational Platforms: please don’t call them “bots”

At Microsoft we are working on many Artificial Intelligence (AI) conversational platforms, transforming the way companies engage their customers. When building them, one of the most important success factors is providing these AI with the right personality: right for your company… and right for your customers.

Please remember, in case your AI conversational platform had a strong “personality”, never call him or her “bot”. Otherwise, you may have quite a “personal” response… :-)

It’s important to know the difference between conversational AI and conventional chatbots (or just “bots”). Chatbots serve up canned responses to anticipated requests and statements. The systems are often isolated applications, quite simple, and not integrated with company platforms or data. This means they can’t understand context or reply without that additional information, which can make them too one-dimensional to offer the kind of support customers demand today. Also, bots usually listen your customers just with the intent to answer any specific question, without understanding much of it, and definitely do not generate much added value, insights, or augmented intelligence for your company.

Conversational AI platforms listen with the intent to understand

Conversational AI platforms, on the other hand, are deeply integrated into information systems, can communicate across most corporate channels, including text messaging, voice interfaces, corporate websites and social media, and – most important than anything else - they listen your customers with the intent to understand, generate added value, and enrich your company with insights and augmented intelligence coming from your customers and your employees.

For the reasons above, successfully deploying conversational artificial intelligence (AI) is like no other digital business-process upgrade. In fact, it’s not an IT upgrade in the conventional sense; conversational AI does nothing less than usher sophisticated robotics into the front office. The surest route to project failure would be taking this fact for granted.

Where these cross-channel AI systems, designed to interact naturally and fluidly with internal users and/or customers, are most like traditional business is in how short-sighted decisions can damage your company. What should you keep in mind, therefore, when deploying conversational AI?

Let’s see the WHY, WHAT, and HOW of building a successful AI Conversational Platform.

The WHY

Conversational AI’s “persona” will embody your company’s values and reflect its competence by communicating accurately and efficiently with uncounted, hard-to-predict human beings. As a matter of facts, Conversational AI will become your Brand’s New Face: Like switchboards and websites before it.

A synthetic voice created with a personality that is too at odds with your brand can be just as disastrous as a tone-deaf advertising campaign. Sometimes, creating “just another bot” can really be the most devastating initiative you can undertake, for your brand.

Progressive Insurance, for example, has capitalized on its popular, quirky advertising mascot “Flo,” creating the Flo Chatbot on Facebook, which chirpily banters and aims to sell policies. Also, Vodafone has created TOBi, with a personality which represents Vodafone’s values, while Telecom Italia (TIM) has rolled out a new AI digital Assistant called Angie which features a comic book heroine.

The WHAT

Your Conversational AI must integrate all Key Data Sets you have within and outside your company. CXOs today need to move past lip service about data’s value. Savvy executives don’t act on partial data, and the same is true for successful customer-facing AI systems. When deploying conversational AI, ensure that the system has access to the following data sets—otherwise, its effectiveness will be limited:

Also, when comparing the two interface options, European and Latin American telecommunications firm Telefonica realized it needed to go beyond the levels of service conventional chatbots deliver. The result was a system that provides its customers, who live in six countries, with a broad and rich data set across several channels, including home devices, mobile devices and social media. Like Progressive, Telefonica used off-the-shelf software tools to build, test, deploy and deliver its natural-language solution.

Think of data as fuel for your AI system. Assuming your firm’s data is as safe as possible, give your AI systems unhindered access to every database that they need to perform tasks successfully. No one wants to be on an airliner that is short of jet fuel, and companies can’t afford to be in the same disastrous situation with their AI systems.

The HOW

When building a Conversational AI, you should really mind the tech behind it. There are facets of a company where technical issues can be fixed without customers ever knowing there was a problem. Deploying conversational AI is not one of them, because its shortcomings will be your brand’s shortcomings. An AI system must be bulletproof.

Part of bulletproofing is saturation integration. Little happens if conversational AI platforms aren’t extensively integrated with and through ERP and CRM systems. Successful implementations have real-time access to all relevant corporate databases (cloud-based or otherwise), amounting to gigabytes if not petabytes of data. Ideally, their interfaces—synthesized voices—are realistic enough to keep users focused on the task, not the software.

In addition, as a practical recommendation, you should preferably start small, when deploying conversational AI. Make the first implementation small and focused. Early, smaller deployments can serve as building blocks, helping enterprises adapt to new policies and procedures as well as internalize lessons learned from the deployment of what is, after all, very complex and unique software.

Be willing to pivot, as well. Perhaps the most counterintuitive point is that CEOs have to be ready to break what their firms have just built. If a chatbot you’ve implemented isn’t working correctly or in tone, reevaluate. Begin again if necessary, because it will be one of the most visible aspects of your brand.

To conclude, conversational AI is going to be part of a branding arms race in every industry, and the winners will be the CEOs who exceed customer expectations and market standards using AI. Have the right WHY, WHAT, and HOW, and you will build a brand that, among other attributes, will promote stronger, longer-lasting relationships with customers. If done right, AI Conversational Platforms, at least early on, will look like magic to customers, setting the company apart in an important way.