Brett Burkhart over at Shift recently gave a talk on how you can use AI and ML as a part of your marketing strategy. It’s a pretty cool topic, so we wanted you to hear his points, as well.
First off, maybe you’re not familiar with what AI and ML are. These acronyms go hand in hand, standing for “Artificial Intelligence” and “Machine Learning”. When you hear these terms it’s easy to assume that:
- Only giant companies can use them
- They spy on people to learn and steal information from them
- They’re all about robots taking over the earth
But, those assumptions couldn’t be further from the truth. Artificial intelligence is “the capability of a machine to imitate intelligent human behavior” and machine learning is “utilizing algorithms based on sample data, to make predictions or decisions without being explicitly programmed to perform the task”.
Sometimes it’s helpful to look at examples of these technologies that you’re probably using today, but not calling AI & ML. Here are a few examples of these technologies that you’ve likely encountered:
Amazon making product recommendations at what seems like the very moment you’re interested
Pandora learning the music you like and making recommendations you didn’t know existed
Nest thermostat noticing trends in your activity and adjusting your home’s systems
Examples of AI in business:
Process automation: The art of applying AI technology to your everyday processes to bolster productivity & enhance efficiency.
Ex: Inventory management
Cognitive insight: Using algorithms to detect patterns in large volumes of data and interpret their meaning.
Ex: Fraud detection
Cognitive engagement: Engaging employees and customers using services like natural language processing, chatbots and machine learning.
Ex: Customer service
Now that we’ve covered what AI and ML can do, and what they’re being used for, let’s dive into how they can be used in marketing, specifically. Below are the facets of these technologies that we’ll cover:
- Chatbots & Conversational AI
- Split Testing vs. A/B Testing
- Product/Content Recommendations
- Visual Search & Image Recognition
- Social Listening & Sentiment Analysis
- Dynamic Pricing
1. Chatbots & Conversational AI
These work from a pre-programmed library of responses to questions (sort of like an FAQ), and learn as they encounter new questions, terminology and user responses. A good and experienced AI chatbot is difficult to detect from a human.
2. Split Testing vs. A/B Testing
A/B testing has been around forever (showing two options to an audience to see which version resonates more), but split testing using AI allows for many more combinations of information (headlines, content, imagery, CTAs, etc). It can learn to show only the combinations, which are getting the highest response rate.
3. Product/Content Recommendations
Amazon knows everything about us because their system’s paying attention to a plethora of activity we take. Clustering & interpreting of consumer data, paired with profile information & demographics, allow AI-based systems to continually adapt and offer recommendations tailored in real-time.
4. Visual Search & Image Recognition
Can be used to personalize the shopping experience using visual search technology. It also allows for recommending of relevant products based on how they look (similar or complementary style, color, etc) rather than solely on past behavior or purchases.
5. Social Listening & Sentiment Analysis
With AI, brands can perform sentiment analysis on social conversations. It enables them to identify and understand attitudes towards their brand and/or products. It also allows them to spot potential issues and take action before they become too widespread.
6. Dynamic Pricing
Can be used to compare businesses’ product pricing against competitors to determine if their pricing is too high, about the same, or too low. For example, AirBnB helps property owners determine the price that they should list their property based on geographic location, listing features, local events, photographs and reviews, as well as market demand and timing.
Currently, AI’s copywriting skills for marketing purposes are mostly limited to more basic tasks, such as email subject lines and product descriptions. In February, OpenAI which is backed by Elon Musk, revealed it had created an artificially intelligent text generator called GPT2 which mimics a journalist’s style very, very well. The generator is still not very capable of creating original ideas, though.
Where Should You Start?
No matter what you’re thinking of doing with a new technology like artificial intelligence, it’s best to work through a controlled process. This will help ensure you’ve checked the necessary boxes for a successful and valuable outcome. Here are the recommended steps to take:
- Get Familiar With AI — Take the time to become familiar with what modern AI can do and some examples of what is being done.
- Identify the Problems You Want to Solve — Begin exploring areas you, or your clients’, organizations feel there is untapped opportunity, and evaluate whether AI might be a potential solution
- Prioritize the Value — Assess the potential business and financial value of the various possible AI implementations you’ve identified
- Acknowledge the Internal Capability Gap — Identify the difference between what you want to accomplish and what you have the organizational ability to actually achieve within a given time frame
- Bring In Experts and Set Up a Pilot Project — Start building and experimenting! (experts could be internal or external)
- Form a Taskforce to Integrate Data — Before implementing machine learning or artificial intelligence, clean up your data sets to ensure the quality of the information you’ll leverage
- Start With a Small Amount of Data — Start simple, use AI incrementally to prove value, collect feedback, and then expand accordingly
- Roll Out Full Scale — With value proven through pilot projects, you can now confidently roll out the full solution for you or your clients’ organizations to reap the rewards
To get (and stay) ahead of your competitors, put AI to work so that you can spend less time gathering & organizing information and more time acting on it.
Will there be uncertainty in the process? You bet. But is there uncertainty in what the future holds if you don’t take action to learn and progress? Absolutely. Exploring new technologies and new ways of doing your work is essential for growth and continued success. Take control of your organization’s future by educating yourself on what’s possible, ideating on what’s applicable, and innovating on what’s valuable.
Want to talk through an idea or the steps to test your idea and bring it to life? Let’s chat!
Read More »