Artificial Intelligence. Cutting Through the Hyperbole.

Each year at every marketing event there are a select number of hot topics presented as the next BIG THING. Think back to 2010, when everybody was talking Facebook Marketing, Blogs, Mobile Responsive Design, etc. Most of these concepts have been adopted and are now accepted and integrated into typical marketing plans.

So, it’s the same today with Artificial Intelligence (AI), you can’t go to an event without somebody presenting on the future of AI, how it will influence every facet of the marketing mix and how savvy marketers should use it to super-charge their customer engagement. In addition, the big platform vendors are lining-up to introduce you to their humanised AI personalities - think Sensei (Adobe), Einstein (Salesforce) and Watson (IBM).

It’s the killer combination of big data, cloud technologies and data processing speeds that deliver on their promises of real-time personalisation built around dynamic audience profiles. In reality, these AI propositions from the vendors are generally just a re-packaging of a bunch of their existing technologies made to appear more whole and tangible. But, hey, that’s marketing for you…

As partners of a number of enterprise platform vendors, including Adobe, Oracle, Marketo, and Salesforce, we guide our clients through the minefield of platform assessment, recommendations and deployment. As such, it’s important for us to have a perspective on their future-state MarTech stacks and how new technologies, such as AI, can align to their strategic objectives and transform their overall marketing performance.

Like a Sheepdog

If you speak to 100 different marketers, you’ll get a hundred different answers on the definition of AI and what it means to their business. In my mind, AI covers a broad range of different technologies that seek to mimic the application of human intelligence to solve a given scenario (often way faster and more accurately than a human being could ever achieve). So, that means solutions that cover machine learning techniques, voice and image recognition, and semantic search.

One of the areas of confusion around AI is that it is way off from being more human than human. Intelligent systems are still far from convincingly passing the Turing test, a historical benchmark for ascertaining the tipping point of AI. But, that misses the point. The goal should only be to out-perform humans in certain areas of capability. A good analogy I heard the other day, is that a sheepdog is way better at herding sheep than the shepherd. Somehow it innately understands the needs and motivations of these herd animals. Sheepdogs are amazing point solutions. But, we wouldn’t then expect the same dog to compose a symphony or calculate the trajectory of the moon landings. And, that’s the way we should look at AI.

Cutting Through the Hyperbole

All good so far. But, how can marketers cut through the hyperbole surrounding AI and apply its use throughout the B2B and B2C customer lifecycles? After all, it doesn’t come cheap and the time to implement and optimise should not be underestimated. In addition, AI lives on data, so if you haven’t centralised your data sets around a common data model, it will be nigh-on impossible to deliver results. This is one reason why CDPs are now on every marketers shopping list and why vendors like Oracle are in a rush to get theirs to market.

Personalisation at Scale

One of the most obvious capabilities of AI is to enable marketers to have real-time humanised conversations with buyers which are hyper-personalised and optimised for any marketing channel. Conversations that use data to empathetically reflect their customers' pain points, goals and ambitions. Machine learning can process billions of data points to establish the most effective times to make contact and what words in subject lines or CTAs on the website are most likely to be effective. They’ll just do this in the background with minimal input and tweaking required from the marketer.

This will mean your organisation can finally have true one-to-one conversations with 5 million of your best customers and prospects. So, whether it’s website, email, mobile app or in-store kiosk, AI will determine the most appropriate message and timing to push the consumer along their purchase or loyalty journey.

Content Creation & Optimisation

While AI can’t write a compelling speech, insightful white papers or in-depth blog posts, it is able to pick elements from a data-set to structure a convincingly human article. This capability can be used for reporting on data-focused information e.g. financial analysis, sporting fixtures, etc.

But, where AI comes into its own is in the area of curated content built around individual audience members. This technique has historically matched people to products e.g. customers who bought product A typically buy product B. Or, subscription services that use machine learning algorithms to make recommendations to drive usage of the service. Think how Showtime makes recommendations based on your preferred genre of programs, viewing preferences and devices, etc. (By the way, we built that capability for Showtime using Salesforce Marketing Cloud…)

Propensity Modelling & Predictive Analytics

Propensity modelling is a commonly stated capability of AI. After all, these models are built on processing huge volumes of historical data to build models that make accurate predictions about the real world. And, this is AI’s sweet-spot, crunching billions of data-points to arrive at some actionable recommendations.

As marketers, we can use this capability to predict the behaviour of customers, such as when are they likely to convert, the optimal price point to get the sale (and optimise margin) and whether a cross-sell or up-sell is most likely to be effective. And, the more relevant and accurate the data we can apply to these models, the better the results.

In a B2B context, these predictive models can use machine learning to score prospects and establish whether they’re “hot” and ready to pass to sales. We’ve been doing this for years with platforms like Eloqua and Marketo, where we apply a score to certain implicit (behavioural) and explicit (profile) data points. But, it has always been fairly binary e.g. If XX job title, give 50 points, if downloaded white paper, give 25 points. Now with AI, we can super-charge the scoring models to add a layer of intelligence that further enhances our ability to generate a steady stream of ultra-qualified leads to that ever-demanding sales team.

Programmatic Media Buying & Ad Targeting

Most marketers have adopted the capabilities of programmatic media buying and the efficiencies and accuracy it can bring, With AI, we can use machine learning and propensity models to increase relevance for audience members and reduce wastage. In addition, by processing historical data, AI can determine which ads performed best on specific websites, and at certain points along the customer journey. And, to ensure they keep coming back to your website, AI can establish the right time and place to retarget prospects on a third-party website with the optimal ad creative to generate an interaction or a click.

Chatbots & Customer Service

Securing new customers is far more expensive than keeping and providing an excellent experience for your existing ones. AI will enable brands to determine which customers are showing characteristics most associated with churn. As such, we can target them with customer support to drive product/service usage, as well as offering incentives and offers at the right time to drive subscription renewals.

And, if your customer service costs are going through the roof as users find it ever-easier to get in contact with you, Chatbots are now coming into their own. They can convincingly mimic human intelligence and act as the front-line of your Customer Services team to field questions, process orders and direct customers to the right support channels. What’s more, there are a number of open chatbot development platforms, which means it’s relatively easy and cost effective to build one of your own. Then, give it a human name to make it more credible/approachable. Something like "Charlie"...

Summary

So, when it comes to AI, there’s a lot to consider. There’s no doubt it will transform the way marketers approach and execute their marketing strategies. Hopefully spending more of their time on the creative stuff, while Watson, Einstein, Sensei, et al, crunch the data. But, if you need help navigating the minefield that AI is becoming, how to avoid the hyperbole and how to build an achievable AI roadmap, drop me a line. After all, Verticurl are global partners with Oracle, Pega, Marketo, Adobe, Salesforce and IBM, so we’ve got the inside track.