Unlocking the Potential of Chat-GPT for PR and Marketing

With Chat-GPT dominating conversations in the communications world, it's important to understand what implications AI can have on our industry. This article is based on a recent presentation to my FINN Partners colleagues on the powerful tool, looking at it through the lens of marketing and PR professionals - outlining how we can leverage its capabilities today and speculating on where modern advancements will take us tomorrow. By exploring potential use cases and understanding the limitations, discover just why the marketing and PR world is excited about the possibilities ahead.

What is Chat-GPT?

Chat-GPT shook up the Artificial Intelligence landscape in late 2022, becoming one of the most searched topics on Google - even outshining the entire AI topic category. Although GPT had been integrated into various SaaS offerings for over a year, Chat-GPT added a new dimension to the already popular model by making it more conversational.

So what's the difference between the two?

Language Model: Machine learning model that learns the statistical structure of a language to predict the likelihood of a sequence of words.

GPT: Language model developed by OpenAI that can generate human-like text in a variety of languages. (Other models: LaMDA - Google, GSLM - Meta)

Chat-GPT: Variation of GPT specifically designed for conversation by training w/ real chat logs to generate responses that are more similar to those of a real person.

Adding such a user friendly interface to a powerful language model, which builds on requests in a conversational manner, made this an instant success - if you can write an email, you can now build with AI.

What can Chat-GPT do?

A blank canvas can be daunting in the communications world, and that's exactly what Chat-GPT is. I compare it to the classic text-based computer games of yesteryear: you have to understand where you are to know where you can go. To get started, there are three main areas for professionals to focus on: "Creation", Virtual Assistant & Planning, & Analysis. Let's dive into what these mean and then get into more specifics for our industry.

Creation

• For marketers: Create high-volume content for marketing and branding efforts: social media posts, blog posts, brainstorming, headlines
• For writers: Generate ideas, outlines and full-form content tailored to specific genres, topics, or styles
• For developers: Build syntactically correct and best practice code for a wide range of programming languages and frameworks

Analysis

• Sentiment analysis: Understand the sentiment of text data, such as customer feedback or social media posts
• Text classification: Classify text data into different provided categories or labels, such as topics, themes
• Topic summary: Identify the main topics or themes present in a given text dataset

Virtual Assistant

• Provide accurate and relevant answers based on specific data sets, such as product information and tech specs
• Generate automated responses to frequently asked questions in a conversational manner
• Sort, organize, parse large amounts of data that usually requires a degree in Excel

How do I start using Chat-GPT for Marketing, PR or Creative Services?

To get you started using the platform, test out a few of the below prompts and build off of the results. Tell it what you like, don't like, what to adjust and continue to give the platform feedback on the results.

1. Write a pitch in email format that’s three paragraphs or less in a slightly conversational tone about a client named [INSERT NAME HERE] who specializes in [INSERT VALUE PROPOSITION HERE]
2. What’s the best formula for Google Sheets that I would use to [INSERT WHAT YOU’RE TRYING TO SOLVE FOR HERE]
3. What are four headlines for the following blog post optimized for [INSERT KEYWORDS]: [COPY/PASTE BLOG POST]
4. Provide six creative titles or taglines for the following [INSERT COMPANY HERE] marketing campaign: [INSERT CONCEPT BRIEF]

Tips for results improvement:

1. Be as specific as possible with your requests
2. Include tone, voice paragraph number, word count, copy/pasted inspiration text, etc.
3. Continue to follow up with feedback on the results to tweak updated versions?
4. Input as much background information as possible – can be copy/pasted from client websites, LinkedIn profiles, previously written content, etc.?
5. Ask it questions if you’re unsure how to get where you’re trying to get with the content it’s outputting

How will Chat-GPT - and AI generated content - change the Marketing & PR industries?

With my career beginning in the early days of social media, AI stands out as yet another milestone. A decade ago it was revolutionary to tell a client we can reach audiences by income level, job title, and who their friends are. Now it's another "WOW" moment as we add AI augmentation to our communications tool belt, in the following ways:

Maximum Efficiencies?

• Clients are going to be looking for quality and an increase in value - you can't just go through the motions anymore, but you need to be strategic and think big picture to hit their goals.
• Complex technical clients usually take weeks to onboard and truly understand their business. The ability to parse data and summarize information allow you to get through the technical jargon and simplify the complex in a digestible manner for you and your team.
• The ability to do more with less - and focus on what matters. Automating and using AI to handle the busy work is ideal and has been possible for a while, but not in a way that is so user friendly and convenient.

Quantity vs. Quantity

• Now that this is an accessible tool to all, volume is no longer a competitive edge - and you're pitting high volume quantity against other high volume quantity in media, SEO, social and beyond.
• Increased volume across the board means more noise and the need to spend our time (and client dollars) on true uniqueness to breakthrough.
• Being that AI work is all derivative in nature, our purpose and value is creating something that has never been done.

Everything is Personal

• Marketing and PR campaigns can quickly be tailored and highly personalized to the target audience. Gone are the days of blanket marketing campaigns or mail-merge pitching.
• Customer service and community engagement becomes (mostly) automated, allowing teams to handle more of "response monitoring" and less being in the weeds. For social media community managers, this would be very much welcomed.

SaaS Revolution

• Chat-GPT created a hunger for AI, and we're already seeing existing SaaS companies looking to integrate where and how it makes sense for their customers. This is going to accelerate in the coming months and years, enabling even more tool benefits when partnering with an agency.
• Imagine Meltwater enabling AI to scan reporters coverage history and drafting pitches tailored to their previous coverage and Sprout Social providing social content recommendations after analyzing sentiment of your post history and competitors public engagements.

And it's already happening, here are a few examples of AI software companies building platforms from the ground up through language and visual models:

Runway: Generate and edit content in seconds.

Synthesia: Create videos from plain text in minutes.

PRophet: Predictive AI-driven PR platform built by and for PR professionals.

Jasper: AI Content Generator that helps break through creative blocks. Used in developing this article to edit for clarity and inspiration.

What are Chat-GPT's limitations?

As with any tool or service, especially the "freemium" model (OpenAI begins piloting ChatGPT Professional), there are limitations of how it can be used. Given how new the tool is, it's important to look at all of the outputs with a critical eye and not push heavy reliance on the platform just yet.

Intellectual

• All work produced by the model is derivative in nature - it's not really creating anything new, but using statistical probability of what has already been created for the outputs.
• Licensing and usage rights are still a bit fuzzy, especially around where/when you can use the outputs.
• Reliance can hinder our ability to learn from our work and truly understand the reasoning/strategy behind the value we bring as professionals

Financial

• We don't know what the cost is going to be, but OpenAI having a recent valuation over $29 billion dollars, they're going to capitalize on the popularity in a premium way soon.
• As AI becomes integrated within other SaaS offerings, there's the possibility to pass that implementation cost on to the customer - leading to a price increase or limited feature access.

Information Access

• It's currently built in a sandbox, so it cannot access the internet or other external sources other than your input.
• The dataset is from a finite moment in time, thus limiting the ability to provide accurate or up-to-date answers
• OpenAI is still scaling, thus frequent outages or rate limits have been occurring - and will continue to occur as usage increases.

Bias/Inaccuracies

• Conscious and unconscious biases passed along the dataset GPT was trained on creeps into the outputs. This is part of a bigger conversation around AI as a whole and one I highly recommend diving into.
• It produces confidently incorrect answers - and if you don't look at the outputs with a critical eye, you'll miss it.
• Unlike conversation and collaboration with a human, Chat-GPT will make a guess on the output, rather than asking further clarifying questions.

AI is not the future; it's being adopted and utilized as you're reading this. For marketing and PR pros, it should be viewed like Meltwater, Sprout Social and all of the other tools we have available: a way to augment the work we do to create efficiencies that allow us to do more with less, and not something that is going to replace us.