Our sister agency OMD, recently published a report on the state of artificial intelligence in Finland. Key takeaways from the report was that only one fifth (21%) of the Finnish population understands what an AI application or device is. At the same time, Artificial Intelligence has become the new buzzword in marketing. But how many marketing managers know how it could be used in their marketing strategy? In an earlier blogpost, we discussed what marketeers need to know about AI and machine learning. But since the topic is usually quite misunderstood and hard to grasp, we wanted to take a more practical approach to the matter.
You can read a summary of OMD’s report on artificial intelligence in Finland here.
A brief introduction to AI – What you need to know?
Like the idea of human flight, the idea of Artificial Intelligence (AI) is not new. For most of our history, we have wanted to fly like the birds. And for most of our history, this seemed impossible. Nowadays, human flight is so commonplace that we take its marvel for granted. This will be the case with AI as well. The idea of AI has been around for centuries but the invention of the electronic computer around the 1950’s AI and its implementations really started to take off. In the 6 decades since then, progress has been made in fits and starts. With the most recent advances made in technology and theory, we are now again at a place of rapid progress in this area. In fact, with things like google translate and customized recommendations on websites such as Amazon and Netflix, we are now using what can be called AI on a daily basis.
AI can come in two main flavors, strong AI and weak AI. Strong AI refers to the full-blow replication and surpassing of human intelligence. Weak AI aims at simply applying AI techniques to solve a narrower range of problems e.g. accurately figure out what ad should be shown to a user on your website. Strong AI is still far away, but Weak AI is very much a reality now. Machine Learning in particular, is a form of AI that is already in use on a daily basis.
Some basics: Our handy AI-cheat sheet for Marketeers
Feeling confused? Don’t worry, we have some basics explained for you.
Machine learning – Machine Learning originated in the 1980’s to give computers the ability to learn and solve problems in certain limited settings. Currently, the terms AI and Machine Learning are often used interchangeably.
Neural Networks – A very flexible machine learning technique that can be used to solve complex problems
Deep Learning – a set of methods that implements large (deep) neural networks with unsupervised learning, huge successes recently (look at a picture and describe it using natural language). More successful than experts in identifying diseases, but the inner workings are so complex that unlike the human expert, the algorithms cannot tell you why they made the particular prediction.
Clustering algorithms – Finds pieces of data that are similar to each other and puts them into separate groups. This is a familiar concept to marketers in the form of segmentation.
Decision trees – As the name suggests, this is a way of dividing data into various groups by using a tree-like decision process. Related to clustering.
Interested in new marketing technologies?
How to utilize AI as a part of your marketing strategy with the RACE-framework
All above is good and fine but how can companies actually use AI as a part of my marketing strategy? Here is one quite practical approach from Robert Allen on how brands can use artificial intelligence as a part of their marketing planning.
Reaching customers – Attract visitors with a range of inbound techniques
- AI generated content (wordsmith)
- Smart content curation (recommendations)
- Voice search
- Programmatic Media Buying
Act – draw visitors in and make them aware of your product
- Propensity model (sounds like machine learning)
- Predictive modelling – predict if customer will convert (e.g. PCM)
- Lead scoring – score whether a lead is worth pursuing for a sale
- Ad targeting – machine learning models can learn from past data to target ads more effectively
Convert – nudge interested customers into becoming customers
- Dynamic pricing – provide sales to only those who need it
- Web and App personalization
- Chatbots – mimic customer interactions (getting more and more usable for smaller companies as well)
- Re-targeting – get customers back based on history
Engage – Keep your customers returning
- Predictive customer services – predict churn based on history
- Marketing automation – who to contact, when to contact customers, what phrases to use based on machine learning
- 1:1 dynamic emails
5 real examples of brands utilizing AI right now (That you are probably already using)
- Music: Spotify Discover Weekly. I listen to this every Monday and throughout the week, it kind of gives good vibes to the coming week in the form of personal music recommendations, who knew that behind this is a deep learning algorithm?
- Video entertainment: Netflix Recommendations. Just finished an awesome series and want to watch more? Want to find more British comedy series? No problem. Streaming and entertainment services like Netflix have developed highly sophisticated recommendation services, thanks to artificial intelligence.
- Media: Yle Voitto the news robot. A robot is writing content for Yle News. Wait what? You heard right. The biggest news company in Finland is using a robot to create content for their sports section on the web. So far, the robot has been successfully writing articles on ice hockey.
- Mobile: Google Voice Search. Voice searches are getting more popular and works well (at least on android phones). In the future, accurate voice recognition is expected to be more commonplace.
- Social media: Snapchat facial recognition: Been snapping with some funny filters lately? Instagram and Facebook have also introduced their own facial recognition tech into their platforms.
This are just a few examples of how artificial intelligence is already changing how consumers interact with brands and services. As you can see, the possibilities are endless and the tech is within reach. As many of the examples above illustrate, artificial intelligence is becoming simpler than we tend to think. It is not always about human interaction robots or self-driving cars. Artificial intelligence is already here, right now.
You can check out the full summary of OMD’s Report below (Finnish only)
Do you want to stay up date with the latest news on data driven marketing and technology?
Authors: Nissanka Wickremasinghe and Rasmus Hilli