Will Digital Agencies be Relevant in the Age of AI?

The release of OpenAI's ChatGPT was heralded as a great leap forward in bringing artificial intelligence (AI) into our professional and everyday lives. Its remarkable ability to generate content about, well... anything, based on simple conversational prompts, has generated wide-eyed fascination with where this technology will take us as a society and who may be left behind. While the title of this post references digital agencies, it really encompasses anyone working in the digital space, including in-house corporate digital marketing teams, freelancers, and website or application creators of all types. With AI becoming such an accessible and easy-to-use technology, not only with ChatGPT, but with a variety of new products that are able to create original content based on normal language prompts, the question arises: are the humans currently carrying most of this workload going to need to find a new line of work? Moreover, will the idea of hiring a digital agency to handle the creative and technical tasks needed for building a successful digital presence become obsolete? It's easy to make that logical leap, but the reality is no one's job is in immediate jeopardy as long as they are willing to embrace the incredible new tools that are changing the industry for the better. On the contrary, these technologies will enable digital professionals to provide more services for clients and increase their creative and technical output. And if they've been in the industry for any length of time, they have already been doing this all along.

AI Isn’t New

While ChatGPT certainly made a big splash, it is far from being the first significant application of AI in the online business world. Predictive analytics, content optimization, and content personalization are just a few of the areas that are already largely defined by the application of AI. The many AI-based tools adopted by digital marketers over the past decade have set a new bar on the level of business value they are able to provide.

Google has been at the forefront of researching and applying AI, from its ability to deliver the most relevant results in its core search engine to enabling advertisers to target the right ad to the right user. By analyzing data about customers' demographics, interests, and online behavior, AI can identify which ads are most likely to be of interest to them and display them in the most effective context. 

There are a host of other products using AI that have become standards in the industry in areas such as search engine optimization, predictive customer service, chatbots, speech recognition, spelling and grammar checking, ad targeting, dynamic pricing, and more. These products are the legacy of a long evolution of advancements that have reduced the time spent on mundane, repetitive, albeit important tasks and have exponentially increased productivity. From the early days of spell-checking to modern content management systems that generate the majority of a website’s code automatically, the use of technology to assist with digital content creation has been on a steady march of improvement.

Today, Generative AI is capable of not only creating original content and code but also analyzing existing content and performing a wide array of tasks such as search engine content optimization, tone adjustments, highlighting hidden biases, voice transcription, summarization... the list goes on and on. Similarly, digital image creation powered by the same sort of language prompts used by ChatGPT produces professional-quality images and videos that can be used in a wide variety of applications. But rather than replacing the need for humans to be involved, these existing and emerging technologies serve to increase the productivity, quality, and value of the teams that employ them.

How We’re Using AI at Diagram

At Diagram, one of our primary goals is to deliver business value to our clients through the application of design and technology. While hard work, creativity, and technical know-how are important elements of this, we use tools that give us the ability to generate far more and better insights and functionality than we would otherwise be able to.

HubSpot

HubSpot's suite of marketing products makes significant use of AI to streamline marketing campaign management, create content, and provide reporting on lead capture and engagement metrics. HubSpot's CMS has tools to suggest content updates for search engine optimization (SEO), as well as an adaptive testing functionality that uses machine learning to automatically optimize traffic distribution when testing different variations of a page.

Optimizely

Similar to HubSpot, Optimizely uses AI to provide real-time recommendations for content improvements within the CMS. Using experimentation and personalization powered by machine learning, Optimizely is able to deliver targeted messaging, personalized offers, and recommend the most relevant content for users.

Umbraco

While it does not natively provide the AI-driven functionality of the previous two platforms, Umbraco is specifically designed as a composable Digital Experience Platform that allows for seamless integration of personalization and automation tools that best suit the needs of the site. Rather than taking the monolithic "do everything" platform approach, Umbraco recognizes that most organizations have specific needs when it comes to advanced technology and seeks to provide a low barrier to being able to pick and choose the best tools for the job.

The Human Element

These tools, along with others we utilize, allow us to take advantage of the transformative power of AI to increase the velocity and quality of our services and help make critical business strategy recommendations for our clients. However, one of our core values as a company states that great digital experience start with people. When it comes to AI and the tools we employ, the human element is still an essential part of the equation. AI provides a level of intelligence that goes beyond the ability of humans to produce in any comparable timeframe, but human judgment is still a critical part of making business decisions.

At the core of AI is the ability to reference huge volumes of data at speeds that a human cannot possibly match. However, AI cannot go beyond the data that it has access to. AI struggles with nuance and context, flexibility in thinking, and other non-digital inputs such as verbal communication and observational data (i.e., what our ears and eyes tell us). At the end of the day, a well-informed human is still the most vital aspect of any digital strategy. The AI-powered tools we use are the ultimate helper in this regard, but not the controlling factor.

There is also the need for a talented and informed human operator of any AI interface. Today and for the foreseeable future, AI tools are not going to write, draw or analyze anything without being prompted by someone made of flesh and bone. The operator is key in ensuring the output delivered is what is actually wanted. For all of its speed and power, AI will only deliver what is specifically asked of it. If a prompt is vague or lacks specificity, the AI algorithm will return the closest results it can based on the information it is given. The less information given, the less chance of a useful result.

Take the following examples of prompts entered into the DALL.E image generator:

“Picture of a squirrel”

It’s a very nice, realistic squirrel. But with added context and description, a much different result can be returned with minimal effort.

“Portrait of squirrel, eating a nut, in the style of Monet + pastel colors + forest background”

As you can see, the output will be very different based on the inputs given. Furthermore, the operator has the ability to further refine and improve the content with follow-up prompts and direction until they get the results wanted. A good operator should have a solid understanding not only of AI but of the subject matter of the content being produced and the context in which it will be used. In the right (human) hands, AI can produce very advanced, useful content in mere seconds, freeing up time and resources for other more important tasks.

Where is this Leading?

As with most technological advancements, it is incumbent on individuals to decide whether they will embrace the new technologies or hang onto their old, comfortable tools and processes... and get left behind. The speed, flexibility, and sheer volume of production delivered by AI will create an insurmountable advantage for those who choose to utilize it. The only jobs lost will be for those who choose not to adapt.

The sheer breadth of capabilities being introduced by AI is also presenting a major challenge to any organization hoping to take advantage of it. It's likely there will be a need for new roles such as Director of AI or AI Prompt Engineers devoted to understanding how AI can be applied to a business and developing strategies for its integration. Much of what we're seeing now through ChatGPT and similar products is the trial period of getting to know what is possible. It is by no means perfect yet and can produce some questionable results. But in the near future, the most successful businesses will build a deep understanding of how to put this technology to work for them.

In the end, AI will not replace digital professionals but will increasingly become an essential and evolving part of the tools we use to achieve the business goals of our clients. Clearly, there are some roles that will be streamlined or redefined (copywriters, editors, customer service agents, etc.) but the digital industry stands much more to gain than it will lose through AI. But don't just take my word for it. I asked ChatGPT what job roles it will replace. Here is its response:

“As an AI language model, ChatGPT can assist with tasks that involve natural language processing and generation, such as answering questions, summarizing text, and generating content. While ChatGPT can help with many tasks, it is unlikely to fully replace any job roles.

However, ChatGPT and other AI language models may impact certain job roles that involve repetitive or routine tasks, such as data entry or basic customer service. In some cases, ChatGPT may be used to automate some of these tasks, allowing human workers to focus on more complex or creative work.

Overall, it is more likely that AI language models like ChatGPT will complement and enhance human work, rather than completely replacing it.”

Couldn't have said it better myself.

Sources:

https://www.hubspot.com/artificial-intelligence

https://www.marketingaiinstitute.com/blog/the-future-of-business-is-ai-or-obsolete

https://hbr.org/2022/11/how-generative-ai-is-changing-creative-work

https://www.optimizely.com/insights/blog/why-ai-will-supercharge-your-marketing-and-how-to-do-it/

https://www.optimizely.com/insights/blog/5-ways-ai-helps-you-get-market-research-right/

https://www.optimizely.com/optimization-glossary/content-intelligence/

https://umbraco.com/explore-umbraco-benefits/composable-dxp-for-umbraco/

https://www.livemint.com/opinion/columns/ai-is-yet-to-learn-the-contextual-nuances-of-what-we-say-11661187006053.html

https://www.datatobiz.com/blog/artificial-intelligence-in-advertising