ChatGPT: Embracing AI-Generated Content in Agencies
With the recent launch of OpenAI’s ChatGPT in late 2022 (and GPT-4 as of this week) AI-generated content has been a hot and somewhat contentious topic in the marketing world and beyond, sparking both feelings of excitement and apprehension. Now, months after its release, ChatGPT continues to be a buzzing topic, with more and more individuals and businesses experimenting with the tool and recognising some of the instantaneous benefits it can bring, even with a few pitfalls.
As an agency, we have already embraced some of the benefits of AI writing tools, effectively using them to create time-saving opportunities for internal processes and clients. However, with the problematic perception that ChatGPT and other AI writing tools have the ability to replace genuine content writing and, quite frankly, make marketers a bit lazy, we wanted to chime in and clear the air for business owners, clients, and fellow marketers who have these types of concerns.
Below, we outline the pros and cons of tools like ChatGPT, highlighting the many benefits of an AI writing assistant for marketing teams, along with tips on how to create efficiencies in your business and what mistakes to avoid when using these powerful language models.
The Basics: What Is ChatGPT/the GPT model?
Straight from the horse's mouth, ChatGPT defined itself as below:
I am ChatGPT, a large language model developed by OpenAI, an artificial intelligence research laboratory consisting of the for-profit corporation OpenAI LP and its parent company, the non-profit OpenAI Inc.
As a language model, I am capable of generating human-like text, answering questions, completing sentences, and even carrying on conversations on a wide range of topics. I was trained on a massive corpus of text data using deep learning techniques, which allows me to understand natural language and generate coherent responses.
While GPT-3 has been around since 2020, it wasn't until November 2022 that ChatGPT was released and available for free online to anyone who signs up for an OpenAI account. On 15 March 2023, OpenAI also announced the release of GPT-4, further advancing the capabilities and intelligence of ChatGPT:
We’ve created GPT-4, the latest milestone in OpenAI’s effort in scaling up deep learning. GPT-4 is a large multimodal model (accepting image and text inputs, emitting text outputs) that, while less capable than humans in many real-world scenarios, exhibits human-level performance on various professional and academic benchmarks.
The release of ChatGPT on GPT-3.5 has allowed OpenAI to thoroughly test its foundations and technology stack to iron out some of the bugs and pave the way for GPT-4. For many, this is an exciting advancement in AI that speaks to the fast-developing nature of the technology.
Why has it been such a talking point?
ChatGPT's ability to produce informative, conversational-style short and long-form copy on virtually any topic is undoubtedly impressive. 'Conversations' with ChatGPT flow naturally enough to make it feel like a human being is on the other side of the screen; on top of this, it can answer even the obscurest questions faster than you could Google them, be it something like "explain the space-time continuum" to complex coding or maths problems. With these capabilities, ChatGPT has generated a lot of interest and excitement across the board, with hundreds of millions of users flocking to the site to try it out since its launch.
Of course, as more and more users experiment with ChatGPT and recognise the breadth of its capabilities, concerns have been raised about how the tool can potentially be abused in professional and educational settings, or lead to job losses if a business decides to rely more heavily on AI rather than actual employees. While there are certainly valid points to be made on the ethics of use, it's important to keep in mind that ChatGPT (and similar AI writing tools) are not without flaws.
For example, a widely known pitfall of AI tools like ChatGPT is that there are issues with the reliability and objectivity of the content itself. When asked whether the information and answers it produced were reliable, ChatGPT responded with:
The reliability of ChatGPT's information depends on several factors, including the quality of the training data and the accuracy of the model's responses.
...The accuracy of its responses can depend on the quality and relevance of the training data. If the training data is biased or incomplete, the model may not be able to generate accurate or reliable responses.
Additionally, ChatGPT is a machine-learning model and does not have the ability to verify the accuracy or reliability of the information it generates. It generates responses based on patterns it has learned from the training data, but it cannot independently verify the accuracy of the information it generates.
With this in mind, it’s important to get a strong grasp on how to properly use these tools, ensuring the input is thoroughly considered, and the output is properly vetted. No matter how powerful or advanced AI like this becomes, it’s important to remember that these tools are not a replacement for the creativity, critical thinking, knowledge, and experience of human writers.
The Use of AI in Marketing Agencies
Like it or not, AI is not going anywhere, and it only stands to become more sophisticated and integrated into our lives over time. As more and more businesses embrace the technology, there will naturally be some curiosity from existing and prospective clients about how it will be used (and to what degree). Concerns may also arise around the ethics of use, and whether or not an agency may potentially become over-reliant on AI or produce different / lesser quality work.
In response to these types of concerns, it’s important to stress that one of the most critical elements of any marketing strategy is the bigger picture; the overarching strategy that requires a strong understanding of a client’s brand, agency-client discussions, creative and critical thinking, experience, and a host of other nuanced factors that AI is incapable of orchestrating. Even when using AI tools, the input or prompt that is fed to the AI has to be refined and considered - and specifically in line with the strategy - in order to produce output that is workable and usable.
How it Can Help & Use Examples
As we established above, AI tools like ChatGPT are not (and should not be seen as) an alternative to copywriters, project managers, or SEO and PPC professionals. The knowledge, training, and experience of the individuals in these roles will be integral to the success of any strategy, even with AI assistance.
However, an AI writing tool like ChatGPT can be used to improve automation and create efficiencies, giving marketers the freedom to allocate more time, focus, and creativity to the strategic marketing activities that will drive the most impactful results within the time constraints of a given marketing retainer or campaign period.
Here are some ways agencies can use AI writing tools to speed up their processes and deliver faster results for their clients. We have also identified where the experience of marketers must come in, even with these automations:
Example Activity: Cold pitching to multiple publications to secure a guest feature and link (linkbuilding)
Example Activity: Updating webpage metadata (particuarly useful if many pages require updating)
Additional use ideas:
- Helping to beat writer's block (i.e. can generate an introductory line or paragraph)
- Generating a closing paragraph to summarise a piece
- Generating a rough blog post that can be expanded, refined, and fact-checked
PPC and Paid Social
Example activity: Creating new ad sets
As demonstrated above, AI can be used to quickly generate multiple variations of copy to speed up certain marketing processes that can be fairly time-consuming. Ultimately, it is down to the strategy that is decided before using the AI tool and what is done with the information after it is produced that matters most. In most instances, the text will require changes to add a human touch, ensuring that the content sounds natural, fits with the brand voice, and adds value for the target customers for whom the content is being produced.
Here are some additional examples of use in areas of digital marketing.
- Generating e-commerce product descriptions and category content
Campaign Organisation and Planning
- Creating checklists and templates to follow for various projects or campaigns
- Client or customer sentiment analysis from reviews and feedback
- Data analysis
- Improving on-site chatbots and digital customer support assistance
Pitching (for PR or Linkbuilding)
- Quickly summarising articles and blogs into a few lines
- Turning existing blog into a press release format
- Organising and tidying up spreadsheet data
- Summarising meeting notes
- Translation of text
- Generating copy for automated customer service emails, site pop ups, form text, etc
- Creating text/summaries for presentation slides
Here to Help, Not Replace Marketers
With the capabilities of AI tools like ChatGPT, it’s natural to wonder, “Will AI steal my job?”
No. While impressive, AI writing tools simply cannot replace a human marketer with in-depth knowledge of the client and years of real-life experience and learnings. There is so much behind marketing outside of just delivering the content and uploading it to a client's website.
In order to use AI as effectively and responsibly as possible, it is important for all marketers to know some of the weak points of these tools when using them. They should also rely on their unique client knowledge and human discretion to determine when they should be used, and to what extent.
Not only is this important for marketing professionals to bear in mind, it should also give any worried or apprehensive clients some peace of mind that their marketing budget is being spent on what matters: the expertise and skill within the agency.
Here are some of the weaker points of AI and why human marketers are so critical in driving client success:
Lack of Client Relationship and Knowledge
While ChatGPT has plenty of impressive capabilities, AI will never know your clients as you do. Naturally, many businesses are very specific about their preferred diction, tone, brand voice, and messaging when it comes to content. Knowing what, when, and how to say something (or not say something) on behalf of your client is incredibly important, and getting this spot on will be down to the client relationship, brand knowledge, and discretion.
Requires Careful Input and Thorough Review
As AI tools scrape the web to learn and produce their answers, and because the data is not always current, there will inevitably be flaws in the accuracy and objectivity of the content. This is yet another reason that human input, knowledge of the client, and lots of experience are required, even when using AI.
For the best possible results, the input must be carefully tailored at the start to get the right output and must be thoroughly checked and refined once it is produced. In this way, content marketers can start from a workable content basis and refine it to perfection.
This is particularly important for industries like legal, medical, and finance, as the ramifications of publishing unchecked and unreliable information generated by AI (which end-users rely on and trust to be accurate) can be extremely serious and even dangerous, which can harm the credibility and reputation of a business and negatively impact end users.
Lacks Human Creativity
Marketing is all about innovation, and AI lacks the ability to be truly creative. After all, they reiterate what they have learned through preexisting online content. Regardless of the output, be it a single blog post or a full-blown campaign, human creativity is at the core of any successful marketing strategy. This is a necessary ingredient in order to create meaningful communication between businesses and their customers.
AI Isn't Going Away, So Hop To It
Ultimately, as this technology continues to improve, it's inevitable that more and more businesses - including agencies - will be using the technology to create in-house efficiencies to better service clients.
At every step of the way, marketing professionals will have to rely on their unique creativity, their years of hard-earned knowledge and experience, their nurtured client relationships, and their own judgement in order to hit the mark for their clients. Rather than a replacement for human marketers, AI should be seen as a tool that marketers can use to maximise their clients’ marketing budgets and drive the best possible results.