Let’s get something out of the way: I’m a big fan of AI.
I use it strategically every day, and I believe marketing and communications practitioners should be learning how to put it to work.
So, if you’re looking for an argument about why we should all retreat to our legal pads and pretend this isn’t happening, you’ve come to the wrong place.
AI gives us opportunities to explore ideas, examine information, challenge assumptions, and improve how we work. There’s tremendous potential here, and we should be making room to understand it.
But enthusiasm comes with expectations. AI marketing value depends on what these tools help our teams accomplish and how that work benefits the organization.
I use AI every day. I also expect it to help me accomplish something.
That matters when the conversation turns to whether AI will take our jobs. Some tasks and roles will change. Learning to use AI won’t make anyone immune to those changes, but avoiding it leaves us with fewer ways to contribute as expectations evolve.
Our opportunity is to combine these capabilities with the experience, audience understanding, and judgment we bring to the work.
I’ve talked before about how we remain responsible for checking the evidence, understanding the people involved, and defending our recommendations.
Knowing how to use the tool matters. Knowing what to do with its output is where our expertise comes in. And that expertise should help us explain what the investment makes possible.
That expectation is getting plenty of attention at the companies building these tools, too.
Ahead of Meta’s quarterly earnings, investors are looking for evidence that its AI spending and adoption of its Muse assistant can support sustained growth. The technology may be impressive, but there’s still a business question to answer.
While your marketing or communications budget probably has a few fewer zeros than Meta’s, the question still applies.
Eventually, “look what this can do” becomes “what is this doing for us?”
Early experimentation deserves resources. People need time to learn, test applications, and discover where AI can help. Expecting every experiment to produce an immediate financial return would discourage the learning we need.
But eventually, someone will ask what justifies continuing or expanding that investment. “Everyone else is doing it” is going to have a rough afternoon in that meeting.
For communications leaders, this should sound familiar. We’ve spent decades pushing for communications to be recognized as a strategic business function while too often reporting our value through things like coverage secured, content published, or impressions generated.
Those measures tell leadership what we’ve done, but too often we leave them to connect that activity to something the organization actually needs.
As we invest in AI, we have an opportunity to make that connection clearer from the beginning. Before adding tools, training, and new workflows to the budget, marketing and communications leaders should be asking: What problem are we addressing? What will our team be able to do better? And how will that improvement support our organizational goals?
Answering those questions gives us a basis for evaluating the work and explaining why it deserves continued investment. It also helps leadership understand the value our teams bring to making that investment worthwhile.
I remember when “we need a Facebook page” passed for a strategy.
Yes, I was around for that. Let’s keep moving.
Social media arrived with plenty of excitement, concern, and questions about who was supposed to handle it. Marketing? PR? Customer service? Whoever was youngest in the office?
We suddenly had places to publish, conversations to manage, and expectations to meet. All of that required resources, often before anyone had clearly explained what the organization would get in return.
The value became easier to explain when we connected the work to a purpose: answering customer questions, building relationships, resolving concerns, or helping people take a meaningful next step.
Strategic practitioners had an opportunity to guide those decisions. Understanding the audience, setting priorities, and connecting the work to organizational goals became even more valuable as the responsibilities expanded.
There’s a familiar lesson here for AI adoption.
But AI reaches further into our work. Alongside changing how we research, plan, analyze, and create, it’s changing how customers discover and evaluate our brands.
That calls for investment in our teams’ capabilities and in the information audiences encounter through these tools.
We’ve learned how much work can follow “let’s create an account.” With AI, we should bring that experience to the decisions we make now.
Adoption gets us started. What we build into our daily work determines whether it becomes useful.
A collection of AI subscriptions doesn’t tell you how much your organization has progressed. It tells you the credit card works.
Progress happens when something useful becomes part of how the team operates.
That might begin with one practitioner testing one AI tool to organize customer feedback. They check the results against the original comments, refine the instructions, and identify where it helps, and where it misses context.
Once that approach proves useful, the next step is making it repeatable. What information goes in? What does someone need to verify? Who interprets the findings, and how do those findings inform a decision?
Then the team can document the process, teach others to use it, expand to other platforms, and evaluate whether it improves the work consistently.
A small business might apply that approach to one recurring task, giving an already stretched team a better way to understand customer questions. An enterprise might need coordination across departments, approved systems, and more formal review responsibilities.
Both need a clear purpose, someone responsible for the process, and standards for what good work looks like.
They also won’t move at the same pace across every application. A team may have an established process for analyzing feedback while still experimenting with campaign planning.
That’s reasonable. The goal is to develop useful practices that hold up beyond one person’s successful experiment.
And once those practices are in place, we can ask a more meaningful question: what are they helping the team accomplish?
Before we answer that, we need to separate two responsibilities that can easily get lumped into the same AI conversation.
One is helping your team use AI effectively. That includes the tools themselves, training, workflow development, and time to review the work. You’re investing in your team’s ability to research, plan, analyze, and communicate better.
The other is helping people find and understand your organization when they use AI.
Buying your team a subscription doesn’t automatically put your brand in a customer’s answer. There’s no complimentary reputation upgrade at checkout.
That second investment involves understanding the questions customers ask, developing useful content, supporting claims with evidence, and keeping information consistent wherever people might encounter it.
It also means building credibility across your media ecosystem, with owned information and independent sources that help substantiate your story.
These responsibilities can support each other. Your team might use AI to identify gaps in public information, then work with subject matter experts to resolve them. But each investment needs its own purpose, resources, and measures.
Internally, you might evaluate the quality of an analysis or how effectively it supports planning. Externally, you might examine whether AI answers describe your services accurately and whether customers understand what you offer.
Neither requires chasing every platform. Both require deliberate choices about where to put your resources, and what you expect those resources to accomplish.
Let’s start with the work your team needs to improve.
Maybe campaign planning gets squeezed because everyone is buried in production. Maybe preparing an executive for an interview takes so much time gathering information that there’s little left for rehearsal. Or maybe your team keeps returning to the same ideas because nobody has time to explore alternatives.
Those are specific problems you can test AI against.
For interview preparation, AI might help organize approved background materials and develop potential questions. The practitioner checks the information, evaluates which questions matter, and uses the additional preparation time to work with the executive on difficult answers.
The value could include stronger preparation, more consistent briefing materials, and greater capacity for counsel. You’ll need to evaluate whether those improvements actually happen.
Choose a recurring task and establish a baseline. How long does it take? How accurate and useful is the finished work? How much revision does it require?
Then compare the AI-assisted process against those same standards, including time spent prompting, checking, and correcting. A faster first draft only tells part of the story. Evaluate the work that’s ready to use.
And if AI saves your team five hours, ask what those five hours make possible.
More audience conversations? Better analysis? Time to challenge a recommendation before presenting it?
That’s how we begin to explain the value to the team. The next question is how that better work benefits the organization.
Five hours back on the calendar is useful. But leadership needs to understand what happens because your team uses those hours differently.
Suppose AI helps your team organize customer feedback and identify a recurring misunderstanding about a service. Your practitioners verify the pattern, speak with customer-facing colleagues, and improve the explanation across your website and sales materials.
Now you have something specific to evaluate. Do customers understand the service better? Are inquiries a better fit? Does the sales team spend less time correcting the same misconception?
Those results connect the team’s work to customer understanding and potential revenue. In another organization, the priority might be employee alignment, donations, or awareness among a particular audience.
Start with that organizational goal, establish a baseline, and decide what evidence would demonstrate progress. The same discipline we apply to supporting our brand’s claims with proof belongs in our own business case.
Then account for the full investment: subscriptions, training, implementation, review, and ongoing maintenance. Redirecting staff time creates capacity; it becomes a financial saving only when spending actually decreases or revenue increases.
And be honest about contribution. Improved inquiries could also reflect pricing, sales activity, or market conditions. AI helped enable the work, but claiming it single-handedly delivered the result won’t strengthen your credibility.
The business case gets stronger when you can explain the connection: here’s what we invested, here’s what our team improved, and here’s the evidence that improvement helped the organization.
Every connection in that business case involves a judgment someone has to make.
Someone decides which customer feedback deserves attention, whether a pattern is meaningful, and what the organization should do about it. AI can assist throughout that process, but the practitioner still needs to explain and defend the recommendation.
“Because AI seemed pretty confident” is unlikely to satisfy the executive asking questions.
That’s why investing in AI capability should include developing the people using it. Training needs to cover how to frame a useful question, recognize missing context, verify evidence, and decide when an answer needs more work.
We’ve said it before, but editing an AI-generated answer isn’t the same as evaluating the thinking.
Leaders need to make that scrutiny possible. Give people time to review, access to colleagues who understand the subject, and clear responsibility for approving recommendations. Match the depth of that review to the consequences of getting something wrong.
A brainstorming exercise and a public statement during a crisis deserve different levels of examination.
And if someone spots a problem, they need permission to stop and resolve it, even when doing so interrupts the promised efficiency.
Human judgment needs time, training, and a defined place in the process. Those resources belong in the investment because they help determine whether the technology produces work the organization can actually trust and use.
