Email vs. AI: Will AI take over email or compliment it to make it even more effective?
15 Feb 2019
Written by Nick Crawford, Principle Consultant at Twist Consultancy and member of the DMA Email Council.
On the 13th Feb the DMA hosted its first “Big Debate: What is the future of email?” event.
As part of the discussion we considered:
Email vs. AI: Will AI take over email or compliment it to make it even more effective?
As a speaker for ‘team email’ I wanted to share with you some of the discussion and provide a plain speak explanation to AI, prompted by conversations post event. Many of the audience confessed to being slightly confused on the topic at the moment, but aware that it’s a term being used more and more at the moment.
So what is AI (artificial intelligence)?
It’s any task performed by a programme or machine that can “demonstrate at least some of the following behaviours associated with human intelligence: planning, learning, reasoning, problem solving, knowledge representation, perception, motion, and manipulation…”
OK, but what is AI in relation to email marketing?
Many of the platforms now include an AI offering in their solution, for example dotdigital’s engagement cloud offers AI enhanced engagement or RedEye’s boost to predictive analytics.
One of the most advanced platform solutions is provide by Ometria .Looking at what their ‘intelligence layer’ delivers provides a great summary of how AI can make a difference to us marketeers.
Collaborative filtering: The ability to create relationships between people and products. Once restricted to recommendation based on past purchase history, using AI now allows prediction based on ‘what people like me like’ or ‘you might like this because you are like’
How does this help? Well it provides content relevance without explicit historic interactions with a product or category.
Temporal correlations: The ability to not only recognise time between interactions, I’ve just bought a TV so won’t be likely to buy one again soon, but much more smartly, the ability to link associated products. I may well want more content or streaming ideas. Managing the huge amount of product combinations to achieve this is only possible with AI.
Re-enforcement learning: The classic self-learning training which optimises decision algorithms. These support all kinds of predictive models, send frequency and even automated channel choice. And as the algorithm model is tested over time, the machine learns, refines and optimises.
What marketeer wouldn’t want these tools at their disposal? As Alastair Jones CTO at Ometria summaries “We view the real power of AI in retail marketing as a kind of ‘assistant’ helping retailers make sense of the data and enabling personalisation decisions at a hugely increased scale”
The power to use and link multiple data points across multiple moments is where our human brains can often struggle to grasp the enormity of what’s possible. As an aside it also removes the natural human bias we incur in any decision making process.
So with this great computing power is there still a place for human’s?
Yes. I believe that great marketing is a blend of science and art. For all of the processing power of AI it is still a narrow learning approach. A specific task using a specific set of parameters.
‘Thinking’ is more than this. It provides context and interpretation. As Alastair also notes, we have all been given an Amazon recommendations that ‘feels’ wrong. Marketing is as much an emotive experience as it is a pragmatic one.
The context and journey we take, the brand and product interactions offered and delivered still need that human element. After all we are creating an experience. However with the right AI tool kit, the marketing bar is raised.
Please login to comment.
Comments