![]() |
| Michael O'Brien |
At a United Nations Security meeting in September, OpenAI CEO Sam Altman argued that “the right tools make people more capable.” Many CEOs introducing AI to their teams are making some version of that promise. Less routine work. Better decisions. More time fo meaningful, business-driving work.
Employees are the first to tell you whether your AI promise holds up in practice. They experience what changes in the work before customers or investors see the results. They’ll know whether a tool alleviates a task or just adds another review step. They’ll sniff out whether or not leadership’s assurances still hold up when speed and cost become the priority.
When employees can point to real improvements in their work, your AI promise gains credibility. Their experience gives customers evidence of what the technology can deliver, and employees evidence that leadership is keeping its commitments.
Honoring Your AI Commitment
An AI rollout includes multiple owners, ranging from Executive Management and CTOs to team managers and communications teams. Only the CEO, however, can own what AI means for the business and its people. It’s what makes this AI commitment so important. To build confidence in an AI promise, CEOs must confidently answer—in their own voice—four questions prior to any internal AI deployment:
- What will change? Name the tasks AI will support and the improvements you expect
- Which tasks remain with people? Outline who will review outputs, make final decisions and own the outcome.
- What does this mean for roles? Address potential changes in responsibilities and performance expectations.
- What support will people receive? Specify the training and IT support available to help troubleshoot during the rollout.
It’s important to explain what you expect AI to improve while being forthcoming in where your team is still working to find answers. Acknowledging uncertainty gives employees an honest account of where things stand. Making assurances you cannot keep gives them a reason to doubt the next one.
Say What Changes in the Workday
Even after internal AI deployment begins, employees will continue to require clarification about question three: What does this mean for my role? As employees begin using the tools, leaders should explain how responsibilities and expectations are changing, then update that guidance as they learn what works.
Let’s look at how this affects communications and marketing. When creating content, AI might help organize research or develop an initial draft. Someone still needs to validate sources, protect confidential information, understand the client's business, and ultimately decide if a recommendation makes any sense. These responsibilities should be explicit to keep a human in the lead.
This is something my team and I had to consider as we rolled out our AI platform during a period of rapid growth. We knew we wanted to offer our clients a suite of AI agents that could streamline the development and execution of campaigns, help draft relevant content, and ultimately drive customers to their business. We couldn’t do this, however, if our own teams didn’t approve of the technology first. Even the best AI solutions can’t provide perspective or thoughtful analysis, and they don’t have taste. In our business, that comes from thoughtful counselors drawing from reservoirs of judgment and experience.
We’ve supported adoption with detailed training, office hours and use-cases drawn from employees’ daily work. We’ll continue to improve workflows and refine our use of AI tools, with direct input from our teams.
In the midst of this period of change, I don’t want to leave my people guessing about our standards for AI or how they apply to their work. Yes, speed matters in the age of AI. But so do accuracy, judgment and accountability.
How we as CEOs measure success for internal AI deployment should reflect all four of these tenets. A faster first draft only saves time if the work required to check and improve it doesn’t erase the gain. The quality of AI-assisted work is the ultimate barometer by which your employees—and eventually your customers—will measure how much to trust you with AI.
Maintain an Honest AI Narrative
When it’s time to go public with your AI narrative, employees should recognize their experience in that story. Not a sensationalized version with buzzwords. Their actual experience.
Customers and investors may need different framing as you tell your AI story. They don’t need different facts. A limited pilot should not become a “companywide transformation” during an investor presentation. A process that required substantial human review shouldn’t graduate to “autonomous workflows” by the end of a presentation.
Each export of the AI narrative, whether it’s a press release, a bylined article, a paid ad, a case study, or a social media post, must always reflect a predetermined set of facts. There should be no room for an employee to question your version of their experience.
When Adoption Falls Short
How leaders respond when a rollout struggles will shape whether employees trust the next step. For example, your IT team may find low adoption rates in the first month of the program. This isn’t the time to chastise employees.
Employees may need additional training. The tools they have may not fit the task. Initial outputs may be unreliable, or output verification may be taking away from potential time saved. Find out which problem you have before deciding how to address your teams.
It’s time to ask questions. Conducting company-wide surveys is a good first step, but go further. Ask employees to show you specifically where workflow has broken down. Give managers room to report disappointing results without treating them as if they aren’t getting behind company strategy.
At Highwire, we have an entire Slack channel called Acro Hacks where we encourage open, honest dialogue about what’s working and what isn’t with our internal AI platform. Our team members can share best practices we can repeat, as well as elements they are struggling with. This allows our innovation team to address issues in real time. It’s people learning from people. When building internal AI trust, team members who find success with the tooling will be your greatest internal AI evangelists.
Earn the Confidence You Want
When your AI rollout succeeds, those internal evangelists will become an extension of your external AI narrative. They’re credible witnesses to how your company uses AI. They explain the work to clients, discuss it with peers, and answer questions from prospective colleagues. Their stories give your public claims additional substance.
Give employees an experience they can describe honestly and confidently. Their accounts of what improved, what took effort, and what still needs work give substance to your public claims.
Before you tell the market what AI is doing for your company, ask one question. Would the people using it every day stand behind that claim?
***
Michael O'Brien is CEO at Highwire.


Creators now have two audiences, the people who follow them and the AI agents those people increasingly ask for advice.
AI will make our counsel sharper and as I see it, AI can tell us what happens but PR experts will tell us what to do about it.
The hype promotes doom, but we can use it to promote PR's human element AI can never replace.
As AI reshapes how people discover companies and experts, PR is becoming increasingly important to building credibility, authority and visibility.
It’s no longer only about helping clients rank on search engine results pages. It’s about helping clients become the sources AI engines reference when generating answers.



