Automation and AI have the ability to make a significant impact — for better or worse — in public relations. The market is ripe with new tools and opportunities to streamline processes and increase productivity.
Deciding which guidelines, guardrails and general philosophy to implement in your agency to govern this technology is more important than ever, and navigating its rapidly evolving landscape can feel overwhelming.
At Media Minefield, we’ve used automation and AI for years. Here’s a quick guide on what to automate in your organization.
A “Human in the Middle” philosophy
For years, we’ve used the philosophy of automating tasks and processes where, at a key point, a human is given control and determines the next step. With the introduction of AI, this philosophy now has a name: Human in the Middle (HIM).
Real people are placed at inflection points in a process to ensure key decisions are made correctly. This may seem obvious, but pause and think for a minute about something you have automated, or used AI for recently: Did you vet the results, or did you assume the output was accurate/all-encompassing? A recent study by F-Secure indicates that most of us don’t check, citing only 30 percent of workers believe there is a need to vet AI outputs for accuracy. It doesn’t take much imagination to think of the worst-case scenario here: an AI output runs unchecked, is inaccurate and is delivered to a client or outside organization. At that point, forget about efficiency gains; you’ve got a reputation issue/crisis on your hands.
Bottom line: whatever you automate or use AI for, have a HIM philosophy in place as a safeguard, especially when working with outside clientele.
What to automate? The manual and mundane
I love alliteration, and in this case, the phrase makes a ton of sense when working with automation and AI. Think of the things that you routinely do as part of your job that require little to no critical thinking (data entry, organization, process execution). That’s where automation and AI can increase productivity for you and your organization.
Here are some examples to get your mind going:
Automation within Operations:
- Client onboarding (creating files and folders)
- Accounting (report creation, reconciliation)
- Internal company communications
Automation within Services:
- Researching (clients, markets, contacts)
- Ideation (pitch ideas, industry trends)
- Client tracking (deliverables and other KPIs)
AI-assisted work vs. fully AI-generated work
Once you’ve determined your philosophy and where you want AI to play in your sandbox, the next decision is: How far will you allow it to go?
To answer this question, you have to understand the two camps: AI-assisted work vs. AI-generated work. AI-assisted work is work that is super helpful to your employee and generally lands in the low-risk but high-impact category — the productivity sweet spot. Undoubtedly, you are already using these types of AI in your business. Look no further than Gmail or Grammarly offering writing tips when you draft an email or when Zoom, Fathom, or any other video conferencing software offers to take notes during your next call. These are tasks that can immediately increase productivity without running a high risk of failure.
AI-generated work, on the other hand, offers high impact along with high risk. Completely allowing an agent to create something or complete a process — like writing a client pitch or creating a graphic — might seem appealing. “Think of the time we will save” is the general train of thought. But in reality, the amount of time spent reviewing and editing such tasks can far outweigh the efficiency gains in having AI create it in the first place.
Partly, this is due to the current state of generative AI. But also, the nature of AI-generated work — the level of complexity that you are asking of it — means it goes against the earlier principle of “manual and mundane”, requiring far more review and a discerning eye.
Using an agent to take notes during a video call vs. having it create a graphic intended for client use could not fall on further ends of the spectrum. At Media Minefield, we do not delve into AI-generated work. Instead, we focus on AI-assisted work, which frees up our minds for more critical thinking.
TLDR
Follow these main guidelines to best determine what to automate:
- Use a HIM (Human in the Middle) philosophy
- Automate the manual and mundane
- Lean into AI-assisted work to keep risk levels low
With this mindset, you will see productivity gains across your organization.
Nate Carey is the Executive Director of Operations at Media Minefield.