Table of Contents
AI has been shaking up nearly every industry over the last several years. Thus far, it’s been used primarily for ingesting data to identify patterns and make predictions that inform business decisions. But a new category has recently entered the market: generative AI.
Generative AI uses a learning algorithm called a Large Language Model (LLM) to create new outputs in the form of visual imagery or text. For sales and marketing professionals, it shows a lot of promise in easily creating and tailoring messages for prospects and customers.
But as the technology races forward, it’s important to be aware of its current limitations and the best practices for incorporating generative AI into your workflows.
The Latest Advancements in AI Tools
The introduction of ChatGPT by startup OpenAI in late 2022 made a huge splash, boasting the ability to produce all kinds of content — including song lyrics, essays, and stories — from user prompts. It even passed exams from both the University of Minnesota Law School and University of Pennsylvania’s Wharton School of Business.
With all the excitement surrounding ChatGPT, other big names in tech have quickly followed suit with their own generative AI solutions. Most recently, Microsoft, which has invested in OpenAI, incorporated ChatGPT-like technology into its search engine Bing and web browser Edge, and announced plans to help businesses develop custom versions of AI-powered chat tools. And Google will soon be releasing its own generative AI tool called Bard, rolling it out to a small test group before making it available for public use.
The Benefits and Drawbacks of Generative AI
Generative AI has many benefits for marketers and sellers. Rather than starting to write sales or marketing materials from a blank page, these tools can help to organize information and provide a first draft as a starting point. They also save time in crafting customized copy to nurture buyer relationships at scale.
Viral Bajaria, 6sense CTO and co-founder, recently noted to AdWeek that AI “can now accomplish the equivalent of 400 to 500 hours of manual work in a day,” sending marketing and sales emails and fielding responses.
But this potential comes with downsides. For B2B users, most LLMs are not fine-tuned for business use cases. These bots don’t “know” brand voice or tone, so the language used within AI-generated copy can feel generic and robotic.
There are also serious legal concerns presented by relying on AI to craft marketing and sales communication. Various tests and demonstrations have revealed factual inaccuracies, instances of plagiarism, and copyright infringement.
Why AI Content Can Be Dangerous
Generative AI tools like ChatGPT are trained on vast amounts of text data, including many diverse sources and genres of writing.
One of the main challenges is that tools like ChatGPT are not inherently aware of the context or intent behind the input they receive, and they do not understand specific brand guidelines or messaging. As a result, they may generate responses that are generic or off-topic, or that contain information that is incorrect or inappropriate for a specific brand.
Additionally, these models are trained on data that may contain biases, stereotypes, or other harmful content. This can result in the generation of inappropriate or offensive responses.
To mitigate these risks, it’s important for brands to carefully review and filter the output, and to provide additional training data that is specific to their brand and tone.
Editor’s note: This section was 90% written by ChatGPT. The human editing fixed a factual error about the nature of LLMs, as well as some sloppy grammar.
Optimizing AI Output
One key to improving generative AI output is to submit detailed prompts. Think of the AI like a young freelance writer in need of detailed guidance. The more background information you can provide, the better the results.
In addition to the topic, including things like:
- Audience
- How the content will be used (email, social post, blog intro, etc.)
- Format preferences (bulleted lists, word counts, etc.)
- Resources (URLs to cite)
Training the AI to your own data will further improve output. Over the next couple of years, expect companies to integrate tools like ChatGPT into their existing data systems to generate smarter content.
6sense’s Conversational Email offering is a great example. It uses the same LLM as ChatGPT but also taps into 6sense intel to help produce highly personalized emails for potential B2B buyers based on intent signals and buying stages.
Conclusion
We’re just getting started with generative AI tools. Expect a rapid evolution over the next several years. While the tools can save time and boost productivity, AI cannot fully replicate the skills and experience of a professional writer. There needs to be a human in the driver’s seat to ensure the content generated is factual, original, and represents brand voice and personality.
Conversational Email from 6sense produces strong content by tapping into intent data and business intelligence. Engage buyers and improve conversion rates with automated and timely AI-driven messages.
