Trending Now

Four things that prevent marketing teams from getting the most from AI

Why is it that with all the marketing technology vendors claiming their products are fortified with AI pixie dust, business results aren’t better? The same goes for internal IT projects. Everyone wants their process to be better–and artificial intelligence, machine learning, and text analytics can provide that something special. And yet, it doesn’t feel like something special is occurring. At least not in the way that lots of consumer products are getting smarter and better.

AI Promise vs AI Reality

This is a problem as old as the first promise made by an enterprise sales rep. I imagine that the promise, scratched onto papyrus, informed the pyramid builders that his lever was twice as powerful as current models because it was using an iFulcrum. 

Would the iFulcrum really have transformed the way pyramids were built? It’s hard to know. Could the iFulcrum scale? Was it easy to deploy? Did builders trust it? It could have been a solution that, while amazing, delivered little differentiation in the end result. That’s what enterprise AI feels like for many today.

Four Factors in AI Marketing Technology Success

AI can be magical in its impact on business processes. Like many things technical, the key to that magic is largely dependent on things that have nothing to do with technology. The key to AI progress is making sure that the business is ready. That largely depends on four factors.

1. The Business Problem

One of the biggest errors is to use AI to solve a business problem that no one cares about. I know that sounds like something that would never happen in your company. However, we see this. A vendor pitches something that can easily integrate with your CRM stack and “Voila!” you’ve got AI-powered something. So what?

Like any business problem, you need to set measurable objectives that matter. Sure you’re only going to do a proof of concept. But what are you going to prove? With our B2B clients, we’re focusing on initial progress on bounce rates and exits without journey progression. But the real goal is conversions. What are the measures of your success?

2. Domain Expertise

Nobody likes trusting their success in a black box. Yes, we want better performance. But we want to understand that performance especially when it comes to explaining it to our boss. That’s why the AI that has the greatest odds of transforming your business if AI applied to places where you and your team understand the domain intimately. 

For example, there are a lot of search engine vendors, primarily in the Enterprise Search space, that are using text analytics to improve indexing and relevancy scoring for search. While that’s one area where those technologies can make a great impact, most companies have very little expertise in knowledge management. So when pressed to understand or explain its impact, they can’t. Worse, they can’t be partners with the vendor in translating that impact into greater improvements in the future. How well do you understand the domain? If it’s important to you, what experts can you involve?

3. Data Science Skills

With vendor sourced technologies, this isn’t always critical. That said, you do want to understand what data is driving the predictions, how that data is being sourced, and how that data is being used to predict outcomes. This goes back to the importance of domain expertise and understanding your business context. Will the manner in which these particular vendor solutions work to get you what you need to grow?

4. Expectations

This is one of the biggest risks in every project. Have you set the right expectations with all the stakeholders about what business processes are going to be affected, what resources will need to be committed both during and after deployments, and what measurable results can be expected over a timeframe?

That last point is often the most critical expectation to set. Everyone expects miracles on day one. You have to make sure you understand (back to that domain expertise point) how the impact will affect your business over time and how those changes will manifest themselves in results. When you’re talking about results you should also be talking about how you’ll react if results aren’t playing out. What’s Plan B? Plan C? 

In Summary

There’s no reason any of this should intimidate you from getting started. The goal here is to help create a roadmap for avoiding the pitfalls on any new business endeavor including those where AI promises to help you make a leap forward.

  1. Have a measurable business problem that matters to your business. You don’t have to bet the ranch. Just make sure the results are meaningful.
  2. Start with a domain you understand so you can evaluate the impact on processes, resources, and outcomes.
  3. Understand the data. This is the critical component to AI models making great predictions.
  4. Set expectations correctly about what the outcomes will be, when they are coming, and what’s going to change for all the internal stakeholders.

AI is going to transform how our businesses operate in much the way the internet changed business models starting in the mid-90s. It’s early days. Get started. Set yourself up for success. Iterate and win.

Steve Zakur

With over 20 years of experience in marketing and digital technology, Steve is now CEO of SoloSegment. SoloSegment is a marketing technology company that uses machine learning and natural language processing to improve engagement and conversion for large enterprise, B2B companies.

Join the Discussion

Your email address will not be published. Required fields are marked *

Back to top Back to top