AI is Copilot for Enhancing Agency Efficiencies: OEM Group Perspective (Part 1)

Shane McKinnon, Founder of OEM Group, a healthcare advertising, and digital marketing company, provides an outlook on AI in general, what it means for the marketing industry, and ways business leaders can start experimenting with different use cases. OEM Group and SecureCHEK AI have been collaborating on initiatives that use AI to build claims libraries, precheck content and create personalized tactics at scale.

Broadly speaking, artificial intelligence will continue to be the most important technological development of our time.

Deep learning isn't just the next 'web3' or a new internet era; it's those milestones amplified, expanding at an exponential rate. Creating intelligence will eventually be at the core of addressing all the world’s problems, small and large. And yes, there are concerns with nefarious uses that need to be addressed and planned for, which is to say proper governance is certainly needed.

Properly guided, (that’s a key concept) AI promises to dramatically enhance the well-being of all humans, whether that’s rapid development of targeted therapeutics for ultra-rare diseases, or addressing world hunger, the technology will continue to scale rapidly to help solve big global issues.

From an everyday business perspective, AI will touch everyone eventually, and everyone will use it directly. (Imagine every staff member at your business paired with an AI copilot they will instruct with natural language voice or text to obtain sophisticated answers on the fly.) AI will be the new intranet, assistant, and or chief of staff, with the user experience as seamless as messaging on a smartphone.  

Here's a specific hypothetical: Let’s say the head of sales is paired with a copilot versed on all things sales and marketing, that allows them to instantly obtain sales projections, or a couple slides to present to their team. This is the efficient, low friction version of clicking around a dashboard, or instructing human staff to compile and provide answers to the same information.

Or, you could have an HR head advise his copilot on how they should deal with potential RIFs, or a unique, sensitive staff issue, again with a simple chat, that pulls from that department’s unique dataset.

And speaking of data, companies should be focused on building their data repository and working with specialized vendors to prepare it for these eventual use cases.

What about marketing and advertising specifically?  Large Language Models (LLMs) have made a lot of noise lately, and we see a lot of agencies, especially at the network level, positioning themselves as AI-based or having a unique AI offering.

And it’s not just hype: AI is here to stay, and right now the technology will continue to boost productivity, allowing teams to do more with less.

But AI is not at a place where it can create content without a great deal of human oversight. Even the best LLMs, finely tuned on specific datasets for the creative industry, need a substantial amount of human intervention. From quality input, prompts, feedback, and refining output—advertising and marketing require the human touch (and probably always will).

From personal experience, I would compare the most powerful LLMs to a very talented and driven intern. Maybe this intern is from a top school, head of their class, not afraid to take risks, and takes initiative. These interns in real life can help teams lay the groundwork for different initiatives, such as helping with research, organization, and brainstorming ideas, essentially low-risk projects, and tasks. Like this all-star intern, these LLMs need oversight and review before putting their results in front of a customer. Oversight, feedback, redirection, and final refinement are needed with an LLM application, just as you would need to guide the highly skilled intern who has not picked up the nuances of the job that comes with years of experience, failure, and learning through osmosis.   

So, what are some of the ways business leaders and their teams get started with AI, particularly LLMs to optimize their team’s output? In our next blog, OEM will outline several simple, low barrier ways to start experimenting with different AI based applications and use cases. What should emerge as the common variable, or theme, is that these use cases all nod to AI technology acting as an extension, or “copilot” to optimize human effort, and not outright replacing it.