What does AI mean for marketing?The Buzz
AI has become a mainstay topic in the media since the arrival of ChatGPT in November 2022. Reactions have been mixed from hailing it as the greatest arrival of the century to fearing it could put jobs at risk.
So what’s the fuss all about?
Artificial Intelligence is the process of leveraging computers and machines to mimic problem-solving, decision making, and other tasks which previously would have required human intervention, or human intelligence, to complete.
The concept has been around since the 1950s and is seen in a number of different technologies – driverless cars, face ID on your phone, predictive text, Google maps, Netflix recommendations…
What is generally being talked about in the media at the moment, however, sits into a few pots.
Machine Learning, which in itself is an umbrella term, is short-form for the process of creating a computer system that is able to learn and adapt using algorithms to analyse and make predictions based on patterns of learning. The machine runs the process again, and again, and again, making tiny adjustments until it finds the optimal outcome, based on the challenge it’s been set.
Reactive AI operates only on the data it’s given and doesn’t have functional memory, so its capabilities are limited (though it has still created chess champions). In marketing, we use this type of AI for recommendations and SPAM filters.
There are also Limited memory machines which store data temporarily, meaning the AI can learn new strategies and perfect its decisions – think self-driving cars.
Large Language Model (LLM) & Natural Language Processing (NLP)
ChatGPT, Open AI’s artificial intelligence chatbot, alongside Bard, Google’s equivalent, combine two types of AI:
Natural Language Processing (NLP): which uses data around words and definitions to allow computers to convert text or voice into commands (think Google search results or predictive text).
Large Language Model (LLM): Which is designed to understand and create text that mimics human conversation, supporting users to answer questions and complete small tasks like composing emails, planning holidays, and creating other forms of copy.
From a marketing perspective, there are pros and cons to AI. Here are a few:
AI can be a fantastic time saver for automating tasks:
- Email sends – these can be pre-scheduled to deliver in the right time zone for the recipient for instance, ensuring they receive your information at a time they’re ready to read it
- Automated responses – acknowledging receipt of a query and managing expectations regarding response times
- Social media scheduling – plan ahead so you always have a consistent drip of content
- Lead scoring – assess a potential customers readiness to convert by setting up automated scoring to their behavioural cues
- Chat-bots to handle routine questions – freeing up your time to delve into the more challenging asks
AI can analyse huge amounts of data and help you spot trends that might have been difficult to identify manually. This is a great short-cut to making smart strategic and tactical decisions quickly.
AI tools can be a great cheat for adapting communications using clever automations to anticipate customer needs and interests:
- Respond to website activity
- Trigger communications following abandoned carts
- Make recommendations based on previous purchases to anticipate your customers’ needs
- Serve content more likely to get them to convert based on previous purchases
This new technology is far from perfect and often you still need a human to sense check what it’s doing.
They both give disclaimers to this effect:
ChatGPT: “While we have safeguards in place, the system may occasionally generate incorrect or misleading information and produce offensive or biased content. It is not intended to give advice.”
Bard: “Bard will not always get it right. Bard may give inaccurate or inappropriate responses.”
AI can only respond to the data it’s given. Get that wrong and your algorithms might be a little skewed. AI algorithms tend to be biased so they need to be checked and not just accepted as correct.
Chat GPT and Bard are amazing at creating copy about almost anything. However, the copy it creates is more like a first draft, especially if the subject matter is niche or nuanced. It will definitely need a second view by an expert.
It’s also not quite as creative as a human, not yet at least, and can only recreate or mimic something that already exist. For something truly original, humorous, or emotive, you need something that emotes.
The world of AI is fascinating, and the technology is progressing all the time. The arrival of ChatGPT signified a huge leap forward in the ability for computers to imitate human interactions, however it’s still a watching brief.
Marketers need to be careful about the lack of regulation around burgeoning technology and be mindful that they are still responsible for AI output, meaning it all needs to be checked and double checked.
We’ll be watching closely what happens next.