It’s easy to get excited about the use of artificial intelligence (AI) - such as predictive analytics and machine learning - in B2B marketing. It seems like only a few years ago that virtual assistants using speech recognition, customer service chatbots and machine learning-optimized recommendations seemed like the stuff of science fiction.
That said, we're still in the early stages. It may be two decades since a computer beat chess Grandmaster Garry Kasparov but that doesn’t mean brands aren’t still facing a steep learning curve when it comes to putting AI tools to use. There has been plenty written about Microsoft’s disastrous ‘Tay’ experiment and more human-acting chatbots have a long way to go, particularly when it comes to their linguistic and natural language shortcomings, and the difficulty integrating them with other business systems.
Until a time where even the most obnoxious Turing-style grilling fails to rattle a customer chatbot, can B2B organizations really make use of Artificial Intelligence and related technologies for marketing?
Predictive account selection
If you look at your existing clients, these accounts are likely to have many things in common – even if this isn't immediately obvious on paper. Using predictive analytics, data drawn in from all your systems can be leveraged to look at the behavioral patterns and attributes that all your best customers share. Not only can this be applied to your existing prospect database to find the hottest prospects, such predictive lead scoring techniques can be used to look beyond into the wider market to help you cherry pick new ones who demonstrate an inclination for your services.
Will this replace the same account-based marketing decisions made by human intuition? The results will likely answer that question, but predictive decisions can be made at scale, spot patterns that might not be obvious to the untrained eye, and judged without bias or questionable hypotheses.
Marketers are already using behavioral, transactional and demographic data to shape and refine web and email content during a buying journey. Predictive analytics can be added to the mix to further customize the customer experience.
Predictive personalization observes online behavior to determine how likely future behavior (such as making a purchase, completing a form or searching a specific web page) might be. This data can then be used to serve prospects content and/or services most likely to connect with their interests. Helping you deliver on the ‘right time, right message’ adage, this ‘hyper personalization’ can even be used to recommend individually titled email subject lines, as well as automate the optimal message sent time.
With the belief that virtual assistants could be smarter, careful consideration is needed before you place one in a customer-facing role. Yet behind the scenes, AI technology will be welcomed by B2B sales reps. Like Siri or Alexa, AI sales assistants can automate and streamline the administrative elements of their work, providing a single interface from which a rep can request LinkedIn profiles, create tasks and log events within a CRM, or demand updates on existing deals. Other examples can use AI tech to qualify inbound leads, engage them in conversation and schedule calls with reps if a lead indicates that they are ready to take things further.
Automated ad buying and placement
Programmatic advertising has been used for several years to automate the buying, placement and optimization of creatives into the most appropriate digital space. Plus, Data Management Platforms (DMPs) can tailor these adverts depending on time, location or even weather factors. However, AI is introducing even greater levels of automation. For example, by optimizing cost per click bids for more effective ad spend, using machine learning round-the-clock to monitor changes in market conditions, and to dig deeper into behavioral data for even more precisely targeted ad placements.
The uses of AI in B2B marketing are already fairly mind blowing and what we’ve covered is still just a fraction of a much wider range of uses, including curation, fraud detection, product pricing, sales call transcriptions and analysis, predictive search and so on.
Of course, it’s easy to get ahead of ourselves as even the cleverest deep learning algorithms can be hampered by the limitations of humans. Like more ‘old fashioned’ marketing techniques, they can only achieve what you tell them. In the same way feeding bad data into a campaign will give you bad results, give AI inaccurate, incomplete or biased datasets to work with and you could be facing a disaster. Hence why Tay went rogue when using the darkest depths of Twitter as her knowledge base.
So, before getting carried away living out our Sci-Fi dreams, we need to get the foundations sorted first with a thorough data management and governance system and a Single Customer View. It might not be as exciting as the prospect of a computer doing the parts of your job you don’t like, but it’s easily a more achievable route to better marketing.
Maintaining a cross journey conversation with your customers
Customers engage with brands across many touchpoints and many channels. Yet, while any one of these touchpoints might provide a positive experience in isolation, customer journey optimization is about maintaining this experience as they travel from one point of engagement to the next. Learn:
- What does a customer journey consist of and why isn't it linear?
- How do you measure customer intention?
- Why do you need customer microsegments?
- What are the benefits of cross journey communication?
- What is required for effective cross journey optimization?
This article originally appeared on the B2B Marketing website.