6 Top Questions to Ask MarTech Vendors on AI

The AI hype gallops on. Having worked on intelligent systems in energy, finance, and marketing, I’m thrilled we are struggling with using smarter marketing solutions to help us be better engaged with our customers. That said, we can imagine the confusion and hype that has been created in the industry. Marketers are wondering exactly where and how to ‘fit’ AI into their existing stack and strategy since every martech vendor claim to be offering AI-powered solutions that can take business outcomes to the next level. It is also important to note that a lot of what passes as ‘AI’ today is not really AI, and it’s not even machine learning. It is often a souped-up version of analytics and modeling which nevertheless looks and feels like magic to many non-technical marketers. In fact, true AI professionals wouldn’t even deem Alexa Skills as ‘AI’, whereas you or I may, in layman’s parlance, would probably call it just that.

Adding ‘AI’ to platform or tool capabilities has become a convenient way to add a premium to the pricing –  so, focusing on the right fit is essential for marketers with limited resources. To avoid the shiny new object syndrome, it is important to consider the right questions when evaluating marketing vendors offering you AI-powered solutions :

6 questions to ask your AI-powered solutions provider

1. Why will we need AI to solve this problem?

While we can understand the need for Google or Facebook to leverage AI to solve global business challenges or complex medical procedures that can use some AI chops, does your email marketing system really need AI? Marketers need clarity on why exactly they need AI to address a business problem that can’t be done as effectively with predictive analytics or even some clever data modeling. Putting the scale of your operations and challenges in perspective, and working with vendors to understand why AI is needed to solve a particular problem better than existing approaches is an important first step. In addition, the margin of improvement in effectiveness and the business impact of that margin needs to be evaluated.

2. Do I have sufficient data to use AI?

Even the smartest machine cannot tell you anything unless it has some data to make inferences. Unfortunately, data- or more specifically- high-quality, real-time, actionable data – is not the strongest suite for most marketers. Without smart data, you cannot expect smart solutions or smart insights. It is important to discuss this aspect with your vendor. Understand their assessment of whether you have enough volume and quality of data to really make their solution work and deliver exponential results. Also, check what data and data formats the AI solution can ingest and how far your existing solutions meet those specifications.

It may be worthwhile to start work on a data strategy keeping in mind that one day, you will want to use it for AI applications. Your vendor can help with building a framework for a smart data strategy that would drive the best results when driving AI-powered solutions.

3. Human vs. machine: who is going to do what?

In all the talk about smart machines, we cannot undermine what human analysis is capable of. Again, keeping the scale, context, and application of your industry in mind, it’s important to understand if the AI will only do what humans can do – but more efficiently and at scale- or can it go beyond what human analysis can even fathom, and deliver insights that are truly a competitive advantage? It is also important to set the boundaries between what humans will continue to do and what machines will do when it comes to data analysis and drawing of inferences.

4. What are the skills, training, and retraining are needed?

The kind of data that AI needs means you need data scientists on the team. You need new skills and new people who can help you get the most out of your data and AI technology. Talk to the vendor about the kind of talent needed in-house to really understand the technology, how often its set to refresh and what regular marketers on the team need to know in order to make the best use of the outcomes.

5. Show me ROI?

Like with any new technology or solution, it’s important to define the outcomes and the metrics you will use to measure success and calculate the ROI. Your vendor will need to work through this process with you as they know the solution, its nuances, and the best ways to optimize outcomes in the shortest possible time. An honest evaluation of the gestation period as well as returns before an investment is crucial to ensure continued buy-in from internal stakeholders. Sharing and learning from experiences of other early adopters is also necessary, so don’t hesitate to ask your vendor for references!

6. How do you protect the brand from machine-made errors (of judgment)?

Even the biggest of brands have fallen prey to the travails of technology lacking human judgment, or worse, ‘tech gone rogue’. Every one of the big names – from Microsoft to Google have experienced this – and done the rest of us a service by demonstrating what could go wrong with customer-facing AI solutions. Perhaps they could get away with it without much brand damage – and even come off looking like pioneers – but for the rest of us brands, a scandal like that could wipe out years of goodwill and credibility. Ask the vendor about all the situations in which the technology would touch customers and what possible negative outcomes could look like. Work with them to address the eventualities, based on their knowledge of the technology and your knowledge of your customers. Either way, avoid the Ostrich approach entirely.

AI can certainly help automate millions of tasks, leveraging intelligent systems should be less about the quantity of marketing and more about the quality of marketing. While the scale is imperative, lines should be drawn where improving relationships leads to resilient growth.

Originally posted: https://www.martechadvisor.com/articles/machine-learning-amp-ai/6-top-questions-to-ask-martech-vendors-on-ai/