We had the opportunity recently to sit down with Alexei Kounine, Selligent’s VP of Innovation and Solution Consulting, to chat about some of his favorite subjects, from Innovation to Artificial Intelligence in Marketing. Now that we’re more than halfway through 2020, a year that has defied everyone’s expectations, we were curious to see where AI is heading in marketing. Clearly, in spite of a worldwide pandemic and ensuing lifestyle and societal changes, AI is still having an enormous impact and playing an increasingly key role for marketers in their campaigns. Here are some of Alexei’s thoughts as he looks ahead.
Could you describe what our process looks like when it comes to developing features that use AI and innovating in martech?
We’re going through this transformation right now, actually. Usually the development process for features which do not require AI is pretty straightforward: our Product & Engineering teams would code a feature – say a dashboard widget or something that allows you to build a message faster. Then the Product Manager responsible for this feature would put it in the hands of clients as soon as the feature is ready, so they can start using it and provide us with feedback.
AI-based features are different. You need data to actually develop and test the feature. This is because the way AI-based features work is that they do whatever they are “taught” to do. The teaching part means it trains the mathematical model, which is used in the feature as a very important part. We work in close cooperation with our clients to develop the algorithms used in our features and the models which result from training these algorithms, while preserving consumers’ privacy and enforcing all privacy laws, including GDPR and CCPA.
An example would be, if you want to optimize the send time of an email that’s sent in a personal way to every individual in an audience, you need training data to make sure the algorithm actually performs well and sends at the relevant time, while increasing the general open rate of your campaign. Basically, you have three steps in creating the feature: 1) engineers code it; 2) train the feature for specific data for every client; and 3) assessment, to make sure it’s performing to improve a specific metric like optimizing the open rates. The most important thing to understand is that AI will not help you in any way if you don’t have data to train it on. So before you start thinking of using AI features, you need to make sure that you ingest all the interaction data from different channels across your platform. And this continues to be a big focus for us at Selligent.
What changes are you seeing related to data and AI?
One of the things that’s becoming more and more important is that many decisions need to be taken by marketers in real-time. Whether someone purchases something in a brick-and-mortar shop, as stores slowly reopen around the world, or – more likely these days – a transaction or click happens online, all of this needs to be integrated very fast in the platform, and the platform has to automatically react based on this event and the data that is available about this individual consumer. The reaction can be the trigger of a campaign, or simply a change in the consumer’s universal profile. The importance of data being available in real-time is key, but what is even more important are the decisions which need to be taken instantly at a scale of billions of events per day.
Two other things: if you try to get data from as many locations as possible on a specific consumer, you really have to be careful with privacy regulations. So everything that we do at Selligent regarding personal data complies with all applicable privacy regulations such as GDPR, CCPA; things that are very central to the way you store the data and that allow every individual to change it, alter the data, or get access to it.
The fact that you have this data secured and available in real-time will help you run and train algorithms to make your life as a marketer simpler, so you can run personalization in a scalable manner using AI.
What trends around AI and natural language processing are you seeing today? And could you share what our innovation team is working on right now to respond to these emerging ideas?
One of the things I see becoming more important is the fact that, until now, most of the data that we process and most of the predictions we do using data are focused on behavioral data: clicks and emails, page views, interactions, offline/online, and things like that. This gives you the tools as a marketer to do very effective push marketing. You send something, you qualify whether the campaign worked well, and then you iterate.
Pull marketing is becoming more and more important. For example, any client of Selligent has a lot of textual data in the platform, and this data can come from different places; the platform can even be connected to a support center where people send various types of messages. We get feedback from clients in the form of text, so this textual data is becoming more and more relevant. Currently, there are not a lot of tools that we provide to marketers that help them analyze and act upon this textual data.
One of the big things our innovation team is doing now is to build tools that use textual data to help marketers improve everything that they do. This is commonly known as “natural language processing” (NLP) and natural language generation. What this actually means is using textual data in a form that allows you to optimize something; the subject line of your emails, for example. Some people will be more engaged with a specific keyword, or emojis might perform better with some people. NLP is a big topic and we’re developing tools that explore natural language within the Selligent platform.
Of course omnichannel is an important aspect for digital marketers today – to be able to engage with consumers wherever they are. But how do you respond to those who say that email is dying? Is email really dead? Do you see innovation happening on this channel?
People are a bit scared that their expertise in email will not be as relevant in the years to come. But email is still a very important channel and a cornerstone for most companies’ marketing strategies. There’s a lot of innovation that’s taking place in email. One of them is kinetic and AMP email; two technologies that make email more web-like. All of a sudden, you can start purchasing within emails, you can submit forms from within an email, and so on. This changes the role of email. That’s a big thing.
The second one is that deliverability – inbox placement and all the metrics that are attached to email – is becoming more and more important. Anyone who knows a lot about deliverability can help you greatly improve your email marketing. That’s why at Selligent, we have a very skilled deliverability team.
The third thing about email: I speak to many industry analysts and we have a very interesting view on how emails will be read in the future. Today, you read emails with your eyes. But now there are more and more vocal assistants at home or in your car, with Google Home, Alexa, and so on. So marketers are starting to think, “How can my email be more readable by a vocal assistant, so it makes sense for you as a listener when you listen to a voice assistant reading your emails?” The voice channel is something that’s in a bit of an exploration phase, but I think this is one of the use cases that makes a lot of sense. You can have your car reading your emails to you as you drive. It’s a use case that I think more people will start to realize the importance of and start using that way.
It's always interesting to peek inside the mind of an “innovation guru” like Alexei. We can’t wait to see what’s next for AI and marketing – and where Selligent goes with it!