How We Stay People-First While Embracing Technology

Every business wants to have an edge on their competitors, and sometimes the perception of being different is all it takes. There’s a scene from Mad Men that illustrates it well; Don Draper comes up with the “It’s Toasted” slogan for a client, and they’re confused because their competitors all toast their product too, but Don’s point is that merely sounding different is enough.

There are many businesses currently labelling themselves as being “AI-powered” or "AI-centered", and marketing agencies are certainly among them. It sounds progressive and techie, and they could very well be centralizing their business processes around AI tools.

It begs the question, though: how does that benefit you as a client?

Maybe they use an LLM to save time compiling reports, or generative AI to create ad graphics without having to pay for photo licensing, or a chatbot to avoid hiring a customer service rep.

Whatever they use AI tools for, is the experience for you as a client made any better?

We at IBA are pro-technology and not anti-AI, but we do observe that many businesses that are vocal about being “AI-powered” are largely doing so in an attempt to capitalize on hype and to appear more advanced than their competitors, while sacrificing quality for speed. Fast food restaurants do the same thing and there’s a market for that, but at least they charge less and nobody expects a Michelin-star experience.

In describing ourselves at IBA, we often say that we are tech-forward, but people first. The distinction is important there. The reason it’s not tech first despite being in a tech-dominated space is that, although we use technology to enhance our capabilities and offerings, we prioritize the human aspect of our services.

Reasons Why We Prioritize Humans in How We Operate as an Agency

We believe that marketing is about people, both in terms of the message reaching the end users and the communication with our clients. While we do use technology such as AI in our processes, we don’t blindly implement it wherever possible at the expense of the client experience.

Here are three concrete reasons why the human side of the business should be prioritized.

1) Humans Can Be Held Accountable

There is a famous quote attributed to an IBM training document:

“A computer can never be held accountable, therefore a computer must never make a management decision.”

They even wrote it in all caps, so you know it’s important!

In digital advertising, we are playing with live ammo - your budget. If someone were to build an AI agent to manage their own budget, there would be no accountability if it made a mistake, and the percentage of LLM responses giving incorrect information or “hallucinating” confidently wrong responses is shockingly high. Depending on which study you cite, they generate incorrect information >30% of the time.

If/when an AI agent managing your ad campaigns makes a decision based on incorrect information and spends on something you don’t want or goes over budget, you won’t get your money back. The ad platform would say your agent is an external system to which you irresponsibly gave admin privileges, and the company providing the tool would say your inputs and directives were apparently faulty, so don’t expect compensation from either party.

Even if your inputs/prompts are tightly defined, thoroughly anticipating all hypothetical edge cases, you would have to constantly review and adjust your parameters in light of new information. That alone would be an ongoing game of whack-a-mole, but it would still be foolish to give control over financial decision-making to a robot that routinely gets facts wrong and can’t spell “strawberry”. (Seriously: example 1, example 2)

We do not defer our decision-making to computers, nor allow them to act independently with money.

We can use AI for help with things like collecting and processing information, but humans should review and make key judgments on how their budget is spent.

2) Humans Deal With Nuance

Computers understand data, not nuance, and decision-making is not always black and white based on objective data.

Lead-generation campaigns are a great example of this. A particular ad or keyword could be generating a large quantity of leads, but if those leads don’t convert to sales, it becomes an issue of quality. An automated system instructed to get the highest volume of leads on a given budget wouldn’t pivot from what is apparently working.

There are ways to address this, of course, such as uploading sales data to the ad platform to use the list of customers as a basis for targeting, essentially telling the system “get more of these”. However, anyone who has ever worked in sales will tell you that no two customers are identical. Perhaps some were frustrating to deal with and took longer to get the sale, or had issues after the transaction leading to negative reviews, making them not ideal in the big picture. That is where human input is needed because the computer only sees a list of customers and doesn’t understand the nuanced interactions surrounding the sale.

Inventory constraints also require pivots that an AI agent may not make based on the data it observes as it would run counter to its directive. We’ve had clients who say “product A is selling well and product B isn’t, so let’s capitalize and put more of our advertising budget towards product A than product B” and others who say “product A is selling well and product B isn’t, so let’s put more of our advertising budget towards product B, since product A doesn’t need the help”.

Which approach is correct? We would argue there isn’t one right answer, it’s up to the client. Depending on what they need, the right strategy might seem illogical to a computer tasked with making the operative decision.

We do not override client preference if their real-world experience is not reflected in the data that we observe.

We can use AI for things like calculating bids or flagging KPIs, but humans should intervene and give new rules when machines fixate in the wrong direction.

3) Humans Want to Connect With Humans

Computers don’t buy things, people do, and public reaction to AI-generated media is often harshly negative.

Human reaction to AI-generated images and video ranges from amusement to indifference to staunch opposition, with the latter being the most vocal. People accuse brands using AI for their ads of looking cheap, taking shortcuts, and even theft. For those unfamiliar with how they were trained, generative AI tools use pattern-matching based on massive exposure to existing media, and the ethics around what constitutes intellectual property theft is subject to ongoing debate.

Beyond the legal/ethical discussion, machine pattern-matching at scale means that content starts to all look & sound the same, and originality is lost. Marketing involves creative experimentation, and machines don’t understand emotion, meaning they have to guess. If an AI tool is testing iterations of the messaging in ads but says something counter to what your customers want to hear, you risk losing your customers’ trust and buy-in.

The AI proponents tell you to let the system eventually figure out what to say and who to show it to, but there are a couple of problems with that: one, the testing is on the fly using your money, and two, there’s a difference between taking creative risks and guessing. It’s entirely possible that they test something tone deaf on people who may have turned into customers but now won’t because they saw something off-putting (which you paid for).

Instances of AI-generated media making mistakes is well-documented, with even Google themselves running an AI produced ad that got crucial information wrong. Plus, due to the nature of AI tools being trained on previously existing media, they wouldn’t know what is considered a sensitive topic based on what’s in the news today. Humans can take our understanding of present culture, and create messages that resonate with people based on knowing which emotions to address - and do it without hurting your credibility.

We do not use robots as a substitute for human connection, and this is reflected in our creative materials as well as how we communicate with our clients.

We can use AI for help with deployment and iteration, but humans should finalize all customer-facing materials to ensure they aren’t tone-deaf, inaccurate, or just plain unappealing.

So What?

People like using machines for speed and accessibility, not as a substitute for quality, transparency, or customer service. They should give you more control and capability without changing the essence of what you’re trying to accomplish.

Again, it’s not accurate to say we at IBA are “anti-AI”, rather we are selective with the tools we use to ensure actual benefit to us, you, and your customers.

It's the same reason why we don’t make you wade through the steps of convincing a chatbot that your issue is worth talking about before you can speak to the person who can get you what you need. Our clients get access to real, experienced professionals, always happy to talk shop or brainstorm ideas to drive results.

Even if you are someone who doesn’t mind dealing with robots (and hey, sometimes we use the self-checkout at the grocery store too), we believe technology should be treated as a tool and not a replacement for humans in your advertising operations.

That is why we are tech-forward, but people first.