The marketing leaders joined host Steve Mann for an episode of Leadtail TV’s Counterpoint B2B program.
The fact is, AI is everywhere. We encounter it when Google completes our search phrases for us, when Spotify queues up a playlist of songs we love, and when social media platforms serve us content.
Machine learning is being built into many of the products consumers use every day.
For B2B marketers, algorithms are becoming part of the way we understand and interact with our audiences.
There was a time when email or social media marketing would have been considered the shiny new thing. The shiny new things of yesterday are the established tools of today. And, while content robots are probably not going to be churning out marketing copy anytime soon, as AI matures it will very likely become another tool we see as commonplace.
Know What You’re Looking For
The question for B2B marketers today is how to most appropriately and effectively leverage artificial intelligence in its current state of development.
For Pam, that means understanding your objective. “As a marketer, you need to understand what you are using AI for, and understand your process flow, and understand where you can inject the AI to actually help you.”
Perhaps the most fundamental mistake marketers can make with artificial intelligence is deploying it everywhere it’s available. It’s an easy way to feel cutting edge while not getting a lot of useful information. Having a clear objective means fully understanding how your marketing organization works, knowing the processes and workflows inside and out.
Starting with this internal “map” makes it easier to see where AI can create value for the marketing team. One area Pam and Doug agree artificial intelligence makes a lot of sense is analytics.
Doug shares an example from his own experience. On a website with hundreds of thousands of visitors, only a fraction of them were identified as customers of leads. Doug saw an opportunity to use the tens of thousands of data points available from the known website users to get a better idea of who was using the site. “Because we had these tens of thousands of people … we were actually able … to run an AI against that, and then make inferences about the hundreds of thousands of people that we didn’t know.”
Doug and his team were then able to create custom content based on the AI analysis and offer users a better, more effective customer experience on the website. Artificial intelligence was a useful tool for this project, but it was Doug’s understanding of his organization and existing customers that made the deployment successful.
Who’s The Boss?
In these early stages of evolution for artificial intelligence there is some overlap between the way AI works and what is already possible using automation tools available in many pieces of the martech stack.
Steve puts it this way: “If you’re a B2B marketer, the question that you need to ask yourself is, ‘Am I going to get more value out of using this AI driven marketing automation system … or just go with the straight, sophisticated yet powerful marketing automation tool?’”
The answer to that question is actually fairly complicated in the rapidly changing landscape of marketing tools. And the only way to figure out the best solution for your context is testing, Doug says. “I think you need to go find an AI that makes sense for your problem and you need to go try it. And if it works, embrace it, if it doesn’t work, tweak, and if it still doesn’t work, wait.”
At the end of the day it comes back to understanding your marketing goals, the objective you’re trying to achieve with any of these tools.
It’s important to step back and determine whether AI tools are adding value for your organization. Does AI make sense for your workflow? Is it providing you with actionable insights?
Whether it’s email, paid social, or events, every company uses marketing tools in slightly different ways, Pam says. “You need to think through in terms of what that is, and then determine … How would that apply to you? How can you benefit from it?”
Taking AI To School
Artificial intelligence is only as good as its training, and that training relies on quality data. “The data training set is so critical,” Pam says. “The quality of data, I think that’s why there are a lot of companies, they are kind of hesitant to actually take the AI to the next level, they don’t trust their own data.”
Bias, whether related to faulty data or bad assumptions on the part of AI trainers, is a complex issue facing the full integration of artificial intelligence into marketing practice. Headline-grabbing instances of biased algorithms remind us of how high the stakes can be when it comes to training AI to make decisions.
These types of problems aren’t endemic to artificial intelligence, but the complexity involved in creating algorithmic models can quickly amplify problems that already exist. “There’s, there’s inherent bias in data,” Steve says. “There is intentional bias in data. And then there is bias that’s introduced by the ‘dirtiness’ of the data. And then there’s bias that can be introduced by the individuals who are actually doing the training.”
“Those are issues that the company had independent of the AI,” Doug adds. “All the systems and processes and algorithms that they have today may not be AI, but they certainly have issues.”
A marketing organization that wants to incorporate AI will need to begin with a long, honest look at the data it is collecting. The old programming adage applies here: Garbage in, garbage out.
AI In The Wild
If you’re itching to try out AI for yourself, there’s no need to invest in a custom-programmed robot, Pam says. Plenty of third party tools are already beginning to offer features that are powered by artificial intelligence.
“My recommendation is: understand the third party tools that you are using right now, what features actually have the algorithms and AI built into it. You don’t have to run a big big AI initiative to say that, you know AI.”
Both Pam and Doug agree on the importance of a “trust but verify” relationship with artificial intelligence. This technology is in its infancy, and we have a long road ahead as the technology and our practice with it evolves.
And no matter what level of adoption AI might achieve, there is simply no substitute for human judgement and intuition.
You should be aware of your data quality, and trust your team to validate the information the AI is giving you. If the output doesn’t intuitively make sense to you and your team, it’s time to reassess the data and the processes behind it.
“We’re at the front end of a five-, ten-, twenty-year journey here,” Doug says.
Three Things to Consider
As your marketing organization begins to explore AI, keep these three things in mind to help set expectations and refine best practices.
- AI is here, and it’s likely to stick around. This genie isn’t going back in the bottle. Now is the time to get familiar with artificial intelligence tools, and a great place to start is learning about AI-powered features in third party tools and platforms you’re already using.
- Clearly understand your objective. Just because AI exists doesn’t make it the best tool for every job. Just as you would with any marketing initiative or project, clearly understand what you intend to do before the work begins.
- AI should enhance human intuition and human judgement, not replace it. As you begin to use artificial intelligence, keep in mind the potential for biases, data corruption, and other factors that might invalidate the output. Trust your team’s judgement and intuition as you move forward.
You can watch the entire discussion between Pam, Doug, and Steve on Leadtail TV.