The future of marketing: How to make the most of AI for social selling
Everybody is talking about AI tools such as Chat GPT. While the latest AI tools hold numerous opportunities, they also carry many risks. We look at the fundamentals and give a strategic point of view regarding AI in connection with the world of marketing in our interview with industry expert Apoorv Gehlot.
Apoorv Gehlot is the founder and CEO of Material Inc., a software engineering services company focusing on product design and development. Material Inc. operates in AI, ML, and IoT and offers embedded cloud computing as well as web- and mobile-driven solutions. Apoorv is a technology enthusiast with enterprise solution design and development expertise and loves helping businesses across the globe thrive in digital transformation, from Fortune 500 to SMBs. He is especially interested in AI's impact on B2B marketing and the practical implementations and benefits it can bring to businesses. Outside of leading Material Inc, he loves to explore innovative technologies and cares about the environment.
In this article, Apoorv takes a deep dive into AI and how it relates to the future of social selling and marketing. Read our key takeaways below.
What is AI in general and how is it classified?
Artificial Intelligence is not a new topic in general, but it is new in the marketing world. AI mimics our own human intelligence. While that sounds scary at first, this means it can eventually create drafts for ads or even social media posts for you. A bonus is that AI can help you analyze data to understand customer behavior and preferences and predict their likelihood of buying your product or service. However, you must use the correct level of AI for this to work.
"When you look at AI, there are three high-level classifications," Gehlot explains. The categories are artificial narrow intelligence, general intelligence, and super intelligence. It's crucial to understand that these three categories are reactive, limited memory-driven, or theory of mind and self-awareness.
When you look at those three categories, narrow is in an environment with little to no memory and only responds to specific questions, like chatbots, which are popping up everywhere. It will do exactly what you asked but nothing beyond that. General memory is where it starts to remember a bit of context. It has a limited knowledge and data set about what I might be trying to do. The most evolved version is the self-aware, human-like intelligence everybody tries to get. This level is super intelligence. While it's not yet fully developed, progress is being made toward it. Organizations are certainly using them internally to analyze data, even if they are not yet customer-facing.
Where is AI in marketing currently on its lifecycle? Is there still hype? Are there practical applications?
The answers to these questions vary depending on whether you're asking on a technical scale or a marketing one. In the tech world, we see AI implementations where both enterprises and consumers benefit. However, when it comes to AI in marketing, especially in B2B marketing, it is split in terms of where it is in its lifecycle.
"I don't believe that it's a hype," Gehlot claims. "I think it will evolve into something that every marketing professional will end up relying on down the road." Today, in B2B marketing especially, AI is limited to large enterprises that can engineer these solutions. It has not become accessible to a level where a smaller business's marketing executives can truly benefit from it just yet, as it is very fragmented. However, that is changing very quickly.
Can the AI we have today help us understand and improve marketing data? Or are we still a little further away from it?
"I think it's already starting to do it," Gehlot answers. "Sometimes the best technology is that which you almost don't know exists. It disappears in the background." Many e-commerce experiences are being powered like that today, which helps recommend products to consumers efficiently.
For example, let's look at buying clothes online. Say a person is trying to buy a red shirt, but companies want to show generic colors for mass appeal—white, blue, or black, not red. If an AI determines that the person is looking for a red shirt, then they figure that the likelihood of conversion is much higher for a red shirt picture, so you change the product matrix.
Companies are doing this all over. These applications are already very much used, both at large and small companies. And the primary reason that they have come to fruition and are coming down the market is that they are available on e-commerce platforms that people have used for over a decade.
What AI should companies consider first, and how can they go about it?
The million-dollar question is, where do you start with AI? In tech, it is easy to subscribe to the engineering R & D mindset, which has its place. However, when contemplating bringing technology to the consumer world or scaling that technology, you must first consider the use case. Rather than having a cool piece of technology with a model built around it and then figuring out how to use it in business, it must be the other way around. Think of the use case, and then the technology will follow.
"To me, when I think about AI in B2B, it has to go back to basics," Gehlot claims. "We want to help understand what the customer needs, what they're looking for, then match the customer with the right product or service that we might have, and then persuade them to buy from us."
So how should I learn about AI's progress and capabilities if there are no solid best practices yet? How do I know these foundations or basics will work?
Unfortunately, we don't know what we don't know. "The way I tend to approach it right now is it's not so much about driving a high ROI from a specific tool," Gehlot explains, "it's about trying to figure out how can I improve my workflow or improve on the metric that I am currently measuring against because the industry is not very evolved." Some CRM platforms are starting to integrate some AI tools. A good example is Salesforce's Einstein AI program, which tries to track every lead that comes in based on historical data that it has from its own platform.
However, there is so much that people are used to, and people keep trying to find that one magical tool that can replace all of this and solve all the problems, but that, unfortunately, does not exist yet. And so, you must find a specific, very niche product or service within your needs and then go after that rather than looking for broad solutions.
What is the connection between AI in a social media and social selling environment? And what is the link to content creation, analytics, and targeting?
There is a ton AI can do for content creation. AI is being used to optimize ad copies now, although AI writing long content that feels natural is still hard. But building an ad copy that speaks to your target audience, especially with a specific lead, is easily customizable. And you use AI tools to customize that ad copy. "We've seen that to be very effective and generate engagement a bit," Gehlot claims. "But it is limited in the scope of how much content it can generate and the type of content that it can generate." For example, relying on it to generate an entire blog is not advisable. However, depending on it to create an ad copy is reasonable. That ad copy could be based on content that is being tracked, including what your audience likes on different social media platforms, what comments they are making, and other statistics. This procedure will help produce content that a person might be inclined to engage with.
Analytics and targeting feed in data points used to generate that content and copy. For example, suppose somebody saved some content or read a blog with certain languages. In that case, these analytics might give you demographic data, whether they interact more with videos, images, or infographics, and what content speaks to them. This input serves to enable AI to create a well-performing copy for you.
AI becomes powerful when you learn about the intent of your target audience and then tailor your own delivery of your content to what people are searching for. Sometimes AI is used to help digest technical articles and summarize them in a short social media-ready text. Of course, it always needs to be checked by humans to make sure that it really reads and flows, but it can save time and work on the human side to process a lot of material and make the content more appealing. Little words can make a difference in what users engage with.
Getting subject matter depth becomes an interesting value proposition. "As an example," Gehlot claims, "if you're trying to market to a cardiologist, people in sales and marketing are unlikely to know terms that a cardiologist would engage with." Right now, marketing teams are tasked with trying to produce that content, and they rely on technical writers to be able to do that. But there is a gap between the two, where technical writers are more methodical and focused on the technical aspect of the product. But marketing people want a particular flair for the language. Often, marketing solely seeks key terminology, and AI can do that. Therefore, you could go out and have AI read a ton of published research in your product area, catalog technical terms, and give that to marketing people to use in the marketing copy.
Do today's marketing departments need AI to be successful tomorrow?
"I do believe that," Gehlot answers. "If you don't start on the journey now, you'll be playing catch up later." Even though AI is still developing and evolving, it is essential to outdo the competition and do what everyone else is not doing. Staying ahead of the curve will pay out to whoever invests in these tools because they can achieve their objectives and be more successful. That will allow them to scale up some of these technologies.