With Matt Porter of Qbase, we explore below some Frequently asked questions about Artificial Intelligence and Machine Learning in Database Marketing. Read through below, and also reach out to us if you need answers to a different question from our experts.
What is Artificial Intelligence (AI)?
There are two simple elements which make up AI. Moreover, they tend to be confusing. Therefore, let me define for you how I see AI because there are two clear and different approaches.
Definition of Machine Learning (ML)
Machine learning (ML) is the implementation or usage of computer algorithms that improve automatically through experience. But ML is not artificial intelligence. This despite many companies jumping on the current bandwagon and claiming that their tools use artificial intelligence.
The fact is, they are using ML and calling it AI. ML is where a computer churns through multiple algorithms to produce a result. Furthermore, it can even tell you what the best result is.
Definition of Artificial Intelligence (AI)
Artificial intelligence (AI) is where the computer then uses the best result and makes a decision. In short, it takes action in part due to that result without any human involvement.
Let us apply this to the world of marketing. Companies are holding terabytes of data these days and it is forever growing. A large part of this data is customer communication and interaction across multiple platforms like:
- social media;
- their website;
- phone call logs;
- postal campaigns;
- customer feedback and;
- complaints, to name a few!
ML is able to analyse across all datasets, detecting various failures and successes. AI would take the learnings from ML and apply those to optimise current and future marketing campaigns or change existing campaigns, allowing companies to respond more quickly to their customers’ behaviour.
Combining ML and AI makes marketing more sophisticated, allowing us to interact with people on a more personal basis, treating them as an individual rather than part of a larger segment.
How do I actually use Artificial Intelligence (AI)?
I am quite lucky in that Qbase is partnered with several technologies that have embedded ML and AI into their tools, which make my job easier and very exciting. Two partnerships I work particularly closely with is the Apteco Marketing Suite and Sisense.
The Apteco Marketing Suite is a campaign selection and analytics platform, where I can perform exploratory analysis on datasets to use in predictive models. Through recent developments in the FastStats tool, I can identify from transactional data, what behaviours are important and help classify those transactions into meaningful groups by a few clicks of a button. In other tools, this usually requires a lot of code writing or time developing tests.
Models can be set into a state of ‘always learning’, where the model coefficients are constantly re-evaluated. Apteco has developed the Best Next Action module which then analyses patterns of transactional data and provides recommendations on what to do next. This can be across different levels of transactional data and provide recommendations on the next channel, message, or product.
Adding PeopleStage for Artificial Intelligence and Machine Learning in Database Marketing
Combining the analytic power of FastStats with another Apteco product, PeopleStage, audience parameters can be used to make decisions on which model is best for that individual, indicating which journey the individual should take and creating a truly personalised interaction.
Business Intelligence platforms help businesses to understand what is happening through insight and analytics. These platforms provide access to data on a large scale and can distribute insight to a wide proportion of the workforce.
What inevitably happens when you share insight, is follow up questions. Accordingly, when you have many people throwing questions your way you need a way to quickly answer the low hanging fruit so you can get to the more complex and often interesting insight. With typical BI tools people can find the answers themselves, by being able to drill down into the dashboard, but this is limited.
Adding Sisense for Artificial Intelligence and Machine Learning in Database Marketing
However, what Sisense can do is go beyond the filters in a dashboard and allow you to search within the entire data model that sits behind the dashboard.
But where this becomes really exciting is Sisense’s use of AI in their Natural Language Querying ‘ask me anything’ button, which allows you to write a question as if you were talking to a human being and Sisense will go and explore the data model and present you the answer. It’s like “Ask Jeeves” has been reincarnated but for insight!
Where is the AI you are asking here? Well if you look under the hood, you will see that the Natural Language Query algorithms are building a library of terminology and mapping them to the fields within your data model, continuously learning which is the most appropriate mapping to the data. Different departments and people often use their own terminology. Over time, the NLQ learns and builds that library for you.
How should you use Artificial Intelligence (AI)?
There is AI and ML embedded in a lot of tools today, but people are either scared of using it thinking it will replace them one day, they don’t know how to use a particular module or tool, or they are not even aware that the functionality exists.
I see the use of AI and ML only making my job more interesting and saving me time. It is helping me surface insights quicker and take action by suggesting what to do next. It is helping me churn through large datasets to help me understand what is useful and help me to prioritise where to invest my time. But at the end of the day, it is helping me maximise my skills rather than replace them.
You should do so, too. Just 22% of marketers are using AI (according to Forbes and Salesforce). Therefore, there’s definitely an opportunity for you and your organisation to benefit from it.