News & Views

Opinion / Beyond the Bot

by Contagious Contributor
Matt Dyke, founder of AnalogFolk, explores how brands can master a range of complex and clever technologies through the AI Stack

h heyerlein

Professor Stephen Hawking has warned that the development of full AI could wipe out the human race. My prediction is far less apocalyptic and perhaps a bit less urgent. I believe that AI will fundamentally transform the way that brands approach marketing. There, I’ve said it, so you’d better be giving it some serious thought.

Yet today, the approach of many marketers to AI has been surface-level, at best. By far, the majority of attention and excitement has been focused on the humble chatbot. Unfortunately, the vast majority of these chatbot experiences are actually quite dumb. That's because the objective, strategy, creative idea and development of these experiences starts and stops at the chatbot interface. Instead, brands must consider the entire stack of technologies and services required to deliver genuinely valuable AI-enabled brand experiences for humans.

According to Gartner, in 2017 marketers will collectively spend more than $250 million on chatbots, virtual agents and other ‘conversational technologies’, with Tractica estimating that by 2025, the value of AI will reach $36.8bn globally. With figures like these, it’s understandable that brands want to jump on board and grab a piece of the AI pie – but this haste is why we’re seeing many bad and few truly effective uses of these complex and clever technologies.

Marketers often mistakenly think of AI as one technology, when in fact it’s a term that can be applied to a set of technologies. These range from speech recognition and natural language processing, through to machine and deep learning capabilities. Rather than thinking of these individually, it's worth thinking about how these technologies stitch together and integrate, through an AI Stack, that can be configured by a brand for a specific task.

The mistake many brands make is starting at the top of the stack, obsessing over the human interface (such as a chatbot), and then rarely moving beyond that point. Instead, they should begin at the foundation and build upwards.

Figure 1: The AI Stack

1/ The Data Pool forms the foundation. This is where brands with the best data will win. Accurate, precise, specific training data sets one application apart from the rest. Many large AI platforms (Google, Facebook) are happy to let people use their services for free, as long as they can then utilise the data to improve their algorithms.

2/ The Development Platform is used to create and host AI services, and expose them within any channel. Selection here will depend on your brand’s existing technology stack and relationships. Don’t be limited to key tech players – any vendor can be included, as can any service that you and your agency create.

3/ AI Services use the data pool to categorise and identify brand assets in real-time. By combining your services, your audience’s experiences will be richer. For brands, a critical part of the AI stack is how they orchestrate the various AI services to ensure a a valuable, seamless experience.

4/ Decisioning how various AI services can be integrated to support single or multiple steps. This is where the brand smarts exists as an algorithm; delivering both connected AI and unique intelligence. This is the brand IP, it must be nurtured and protected as an investment.

5/ Human Interfaces are where people interact with your brand across a number of channels: mobile app, IoT, marketing campaign, even a chatbot. These portals are where users engage and share personal data. The more people who engage with your brand at this layer, the more data they generate which will mean that the AI Stack will learn and deliver a brand experience that is more meaningful.

Rules to Stack by

After getting to grips with the AI Stack, STOP, pause and consider the following when implementing your AI strategy:

USEFUL: Remember never to start with the technology. Start by identifying the human need. The more specific the ‘need’, the easier it is to design a ‘narrow’ AI experience that can perform certain tasks brilliantly and, with usage, will continually improve.

CONNECTED: Experience is everything; Your customers will increasingly expect your applications and systems to connect with others to accomplish key tasks. The future of AI is an AI stack that plays nice with other stacks. Yet, for most organisations this is often an after-thought.

UNIQUE: As brands within the same categories begin an AI arms race, the role of the brand will becomes more prevalent: what is the difference between a Watson-powered jacket selector for North Face, versus Patagonia? Without considering the brand, the end result is a utility that becomes generic to a category. For companies looking to plug into and develop AI technology, the drive to consolidate into a single platform doesn’t just threaten to lock them into an ecosystem, it threatens their brand’s individuality.

OWNED: Ownership of the stack isn’t just about retaining brand experience, it’s also about how your company expertise is best leveraged, communicated and utilised. Multiple providers (Google, Facebook, Apple and IBM) have made one, some, or all aspects of the AI Stack available – these general-purpose off-the-shelf products can’t deliver individualism and whilst they appear to offer a great package, brands need to consider their overall AI strategy before signing up to the API’s of tech giants over whom they have little control.

ENRICHED: The most important advancements in AI will be propelled by deploying technology across the world’s largest and most complex systems. As a brand’s AI evolves, it must enrich every aspect of a business across all channels. By connecting your business data and systems one by one via APIs and making them accessible to the AI stack, you can provide true value to your customers.

An Intelligent Future for Brands

The smart brands won’t see AI as novelty chatbots, to be discarded with the next advertising campaign or PR story. They’ll appreciate that AI is a set of technologies that can add genuine value to people’s lives. As such, it must be seen as an investment in brand intelligence that will benefit a business as a whole. By mastering the AI Stack, these smart brands will be able to develop experiences that are genuinely useful, connected, unique, owned and enrich the whole business. Of course, by then it’s perfectly possible that Professor Hawking’s prediction will have come true and none of this will really matter anyway. See you on the other side.