It’s been a whirlwind nine months of AI hype — so much potential and fear. AI is here to stay, poised to transform how your company operates and engages with customers.
As AI-based applications make their way into enterprise organizations, here’s how to ensure you are in a position to use them properly.
A look at AI’s pros and cons
Every day, I hear from colleagues about ways they’ve used generative AI to either speed up a process or create content. One colleague forwarded a LinkedIn post from Eric Partaker with a graphic (created by generative AI!) showing 120 generative AI tools. What a treasure trove to mine!
Companies have seemingly appeared out of nowhere, offering myriad ways to leverage AI in creative endeavors. But as much as generative AI has captured the imagination, AI’s potential for transforming the customer experience across almost every industry and customer touchpoint is what will ultimately ignite excitement.
Yesterday, I spent 45 minutes booking what should have been a simple short-haul flight. I lost count of the number of screens I clicked through. I envision a day when I can use a natural language interface to tell an AI-driven assistant where I want to go and when and then let it do the hard work of getting it done.
Yet, there’s still fear, uncertainty and doubt — ranging from concerns about accuracy and hallucinations all the way to AI destroying humanity. A professor friend summed it up with a tongue-in-cheek Facebook post:
“So … I’ve been working a lot with chatGPT in developing a new course. I am always very polite in my interactions with the program on the assumption that (1) everything that I write to ChatGPT will be saved forever and (2) if I’m nice to our soon-to-be AI Overlords … maybe they’ll be nice to me and not send me to a mining colony somewhere in the boonies…”
Dig deeper: Artificial Intelligence: A beginner’s guide
Where are we headed?
Regardless of your stance, it’s time to accept that AI is here and will impact your business. It’s unlikely that there will be a sudden “flip a switch” moment. AI-based technology applications will roll out like most technology, one product at a time.
In many cases, AI technology will initially yield incremental changes to operations and functions. Over time, as AI technology evolves, we’ll see a bigger impact as workflows are optimized and customer touchpoints become more intelligent.
From a marketing perspective, it’s essential to have a strategy that encompasses where and how you will use AI, establishing the guardrails to ensure that everyone is working to a common set of rules and parameters and a plan to make sure your data is ready to be leveraged by various AI applications.
Defining your role is the starting point:
- Are you and your company an AI user, an AI provider or both?
- Are you leveraging AI tools or are you delivering AI-enabled tools?
Most of us will be AI users (the proliferation of ChatGPT points to that), and some of us will also be AI providers.
What to consider as AI users
As you consider your strategy for using AI as part of your marketing efforts, it’s important to look at generative AI applications as discrete from other AI-enabled applications.
Using generative AI applications requires engaging with an external interface. There is no internal data dependency and very few security risks. There are now many generative AI tools to choose from, as evidenced by the graphic above.
Generative AI capabilities are popping up in established apps and products. HubSpot, for example, has an integrated AI assistant to help create and polish content.
Chances are that your team members are already using these tools to create and edit content. If you haven’t already done so, frame the parameters around the usage of generative AI in your organization and address the following:
- Which tools are approved for use and how to obtain approval to use new tools.
- When and how these tools should be used.
- Content creation vs. content editing
- Content propagation. One of the real risks in using generative AI is that it will be used to boost the amount of content being distributed. That could be a good thing if it assists with personalization but a bad thing if it is used to create multiple versions of essentially the same content, which in turn is used to bombard prospects and customers. I often use Bard and ChatGPT to polish content that I’ve written and use the command “improve this” or “rewrite this” — I could easily do that again and again to get multiple decent variants of the same content and then use those variants as posts, emails etc. creating a content tsunami that would dilute, not enhance, my marketing efforts.
- How content created by these tools should be attributed. How are copyrights handled (time to engage legal)? If an employee has used generative AI to create content, should that be disclosed? If so, how and when?
- How to fact-check a reference generated by one of these tools if the reference is to be cited. We know that generative AI can create false references. How are you going to address this?
- Content ethics. You can now use generative AI to create audio files using another person’s voice — what is your position on this?
We are in new territory here, and I’d recommend engaging your entire team in establishing parameters and guidelines. These should be regularly reviewed to address new applications and any issues that surface.
Though we currently tend to use AI and generative AI interchangeably when it comes to marketing, many other AI-enabled applications can add value to your marketing stack. Most of these applications require the use of your data to deliver value in prospect targeting, creating and refining the customer experience, analyzing performance and other marketing functions.
Remember the old phrase “garbage in/garbage out”? It holds true here. If your data is not clean, whole and free of artifacts, you will not derive any benefits from these applications.
I continue to be amazed at how many marketing departments do not have a clear view and understanding of what data they have, where it resides, how it passes from application to application, how clean the data is and what data is missing.
If your department is one of these, now is the time to take action to document the details of your stack and the associated data collection, storage, sharing and distribution mechanisms and then work to ensure that everything is being collected, stored, shared and distributed as required.
AI will find its way into just about every marketing technology application over time, and it will only perform well if it is working with a clean, complete and accurate data set.
What to consider as AI providers
If your company is a provider of an AI-enabled application, you as a marketing leader, will naturally gravitate to communicating how and why your application leverages AI and the benefit that using AI delivers to your customers.
Communicating how your company thinks about AI ethics may not be top of mind. Similar to how you publish data privacy compliance information, you should be thinking about publishing an AI ethics statement.
This is very quickly going to be something that any prospective customer, and in particular, their infosec department, will want to review as part of their qualification process. Now’s the time to act on this.
Salesforce has been talking about AI ethics for a long time, and companies are increasingly publishing AI ethics statements, including Adobe, WellSaid and Resemble AI.
Externally, several groups have formed to address AI safety. They include:
The U.S. government is trying to address AI safety and has proposed an AI Bill of Rights. They also convened a group of tech leaders (Amazon, Anthropic, Google, Meta, Microsoft, Inflection and OpenAI) to establish a commitment to testing and securing AI developments and watermarking AI-generated content.
These external efforts are laudable, but they do not negate the need for each AI developer to establish their own AI boundaries and ethics. The truth is that we don’t know how AI and AI applications will evolve.
It has been suggested that the pace of development will exceed our ability to conceptualize what’s possible, which almost guarantees that capabilities and new applications will emerge that we’ve never imagined.
As AI-based applications make their way into enterprise organizations, our objective should be to ensure that we are in a position to utilize them properly. That means:
- Establishing rules for the use of AI internally.
- Ensuring that our tech stacks are well documented, including data flows, and that data is of high quality.
- If you are a provider of an AI-enabled product, establish your position on AI ethics and then publish an AI ethics statement.
Get MarTech! Daily. Free. In your inbox.
Opinions expressed in this article are those of the guest author and not necessarily MarTech. Staff authors are listed here.