ChatGPT has been all the rage lately as companies rush to not only explore how they can leverage this new technology but also to capitalize on it as a business model. There are ChatGPT-based services for legal, medical, and of course, content development. As writers and marketers often asked to develop content such as whitepapers and blogs, we thought it would be prudent to dig into what ChatGPT can do for B2B-technology marketing. If you are a B2B or B2C technology company wondering how ChatGPT might help your marketing efforts, such as by developing content, what follows may help you make a more educated decision.
What Exactly Is ChatGPT?
Whether or not you have already played with this new technology or not (through OpenAI or another service), it will help to understand how it works as you begin to evaluate if it will become part of the way you market your product or service.
In short, ChatGPT is an AI system based on a Large Language Model (LLM). This model is developed by providing the AI system input, often in the form of example content. So to develop a LLM about video encoding, you would need to give it examples of content about video encoding: explanations of codecs, explanations of algorithms, etc. This content can be highly technical in nature or very high-level. For automated AI systems though, like ChatGPT, training the LLM often involves scraping available content sources (such as website URLs). Regardless of how the LLM training happens, what makes ChatGPT and other LLM-based AI systems interesting is that they look not just at the meaning of individual words, but the words within the context of a larger sentence. This allows the system to respond to queries more naturally. For a good in-depth look at how these systems operate, and how LLMs are constructed, you can read an excellent piece in the New York Times (not written by ChatGPT, but we will get to why that’s relevant later in this post).
It’s important to understand, though, that a specific ChatGPT-based system (whether generalized or specific) is not “intelligent” per say. Rather, it is a very clever system which works within the confines of its LLM. For human beings, we often have multiple LLMs in our heads (i.e., areas of knowledge, that correspond to a specific lexicon, around which we have developed connections, analysis, and feelings; these LLMs can be colloquial or specific). This allows humans to create connections and surface insights across LLMs: our brains are highly developed pattern-recognition engines. So, for example, if someone asks us something, our response may not only based on connections across our domains of expertise or experience, but also edited in real-time: “oh, you can solve that this way…wait, that won’t work, let me think about it some more, oh, here try this.” This is important to recognize when developing content for marketing technology products which should be focused on solutions.
Use Cases For ChatGPT Technical Marketing
Recognizing that ChatGPT will probably not have native expertise for the specific application of your technology solution to your users, how can you put it to use? It is a very powerful natural-language system which can be used for a variety of specific use cases:
- Answering basic product questions. Having a ChatGPT enabled bot on a website, for example, can be a great way for prospects to get basic questions answered about the product, the technology, etc.
- Customer support. Because ChatGPT responds to inquiries in a very natural way, it can be used to provide a Tier-1 support: answering basic questions (related to support content against which its LLM can be trained) and providing direction to more technical resources if needed.
- Generate basic content. Without a deep expertise in your product, how it’s applied (what problems it solves; and this is generated through experience), and the industry/market in which it operates, any content developed through ChatGPT will notably be high-level.
The Trade-off: Horizontal Versus Vertical Expertise When Using ChatGPT for Technical Marketing
For many marketers experimenting with ChatGPT, the third use case, generating content, is the most exciting. Marketers have often had to rely on internal resources to help create webpage, blog, social posts, and other technical content. And, when the demands for content are too great, they have turned to external writers which can pinch already tight marketing budgets. Most marketers know that content is critical to search engine placement and domain authority but delays and additional costs can keep them from generating the amount of content they need.
Unfortunately, though, ChatGPT is only a part of the solution to the challenge of generating needed content. That’s because using ChatGPT to create content falls into the problem of expertise. As mentioned already, unless ChatGPT is trained with very specific (and related) content, the LLM will always be very shallow. Using a shallow model to develop technically-deep content won’t be valuable to a reader looking for content created by an “expert.” As such, the key to effectively utilizing ChatGPT ito develop content is to align it with LLMs that you can specifically train. So if you can develop an LLM around your product and then supplement it with more generalized information about the technology underlying your product, the content which is produced through ChatGPT will most likely feel more “expert” than using an AI system without a relevant LLM.
Bolstering Organic Search Engine Reach Using ChatGPT In Technical Marketing
Search Engine Optimization (SEO) is abroad set of marketing activities which help companies rank higher in search engines. The ranking is a combination of authority of the content as well as relevance (i.e., through backlinks). So a blog post, for example, would appear in a search engine ranking against specific keywords. Whether or not the blog post is specifically relevant to the search query has to do with both the post meta content (description and keywords) as well as the content (the saturation of the keywords within the search query) and any backlinks to that content (so other sites or content linking to it).
Given that, there are generally two types of content that any marketer can create for SEO purposes: organic and targeted.
Organic content is created around very general keywords. So if someone is searching for “automotive parts” a blog post could appear there if it generally talked about automotive parts. But, as you can guess, that’s a very generic search term. Which bring us to the second type of content: targeted. Most marketers work with SEO experts to develop a targeted keyword list based usually on analysis with SEMrush or another SEO tool. This keyword list more than likely includes search terms for which the company wants to improve domain authority, brand recognition, and generate clicks. This can, for startups, be a term that they are trying to establish. It may have low search volume at first but, as it grows, consistent content produced against it will almost ensure improved search performance.
The key here is that ChatGPT is a perfect tool for short-form, organic content because that content is non-specific. It’s general content and, as we’ve pointed out, ChatGPT is great horizontally. But targeted keyword content, because it is so specific, often demonstrates expertise in the topic. ChatGPT is not good at demonstrating expertise. The kind of deep vertical knowledge that’s required to produce targeted keyword needs to come from an industry or technology expert.
What Can ChatGPT Do For Your Technical Marketing?
There are obvious use cases where ChatGPT can help with technical marketing, but none so interesting and exciting as content development. For marketers looking to use the AI systems as part of their marketing toolkits, it’s best to utilize it for short-form, organic content. You need to produce a blog post a week, this approach can definitely be cost effective. But when it comes to those targeted keyword content pieces, whether long-form blog or whitepaper, it’s best to only use ChatGPT for the generic parts (say like an introduction or overview of the market) and rely on technical or industry expertise to craft the bulk of the piece.
Note: the header image was generated through Dall-E (the Open AI image generator)
