The pings won’t stop. Chats. Town halls. Quarterly updates. Your team is overwhelmed—and disengaged.
You used AI to adapt one announcement for multiple channels—Slack, email, intranet. Technically sound, strategically efficient. But engagement metrics started slipping. Employees could sense something was different, even when they couldn't pinpoint what.
Now you’re facing a harder truth: your people need more thoughtful, targeted communication than ever—especially with AI creating uncertainty about the future of work—yet templated messages are failing to engage.
The smartest internal comms teams aren’t using AI to say more. They’re using it to analyze engagement patterns, tailor message delivery to each team’s workflow, and identify recurring themes in employee feedback that point to deeper cultural or operational issues.
We’ve been in these conversations with internal comms leaders at organizations of all sizes. Their insights are shaping how we approach AI at Polly, and what’s next for the future of internal communication.
The early promise was exciting: AI would help internal comms teams create content, faster than ever before. Instead, it created a problem that many in internal comms saw coming.
Teams that embraced AI content generation found themselves drowning their employees in an endless stream of messages, updates, and announcements that all started to sound eerily similar.
The problems showed up in predictable ways across organizations of all sizes:
The latest research from Forbes delivers a clear warning: nearly half of employees now ignore internal messages because there’s just too much of it.
But the problem runs deeper than simple volume. According to Harvard Business Review's analysis of nearly 1,000 employees, 38% say they receive an ‘excessive’ volume of communications at work, while 27% report feeling overloaded by information sources. The average employee wastes over 3 hours per week dealing with duplicative, irrelevant, or inconsistent messaging.
For some, AI hasn’t helped. It's made things worse.
When internal comms teams first adopted AI content generation, the appeal was obvious. Write one message, generate five variations. Turn a single announcement into newsletters, Slack posts, and email updates. Create endless FAQs from basic talking points.
But distributed teams ended up flooding channels with repetitive, slightly reworded messages, and employees couldn’t tell what was essential and what could be ignored.
Carolyn Clark, former VP of Internal Communications at GoDaddy, saw this mistake unfold across teams:
“More content doesn’t equal more context. More content doesn’t equal better engagement.”
As a result, important updates got buried. Employees tuned out. Message fatigue set in.
The biggest AI failures happen when speed takes priority over authenticity.
Clark shared a particularly painful example: a CEO avatar used to deliver company updates. The intention was efficiency since executive time is limited, but employees picked up on the artificial delivery instantly. Trust dropped.
Even subtle AI tells can erode credibility. Like overused em dashes, generic phrasing, and a tone that sounds more like a press release than a person.
“There’s one phrase my CEO cannot stand,” Clark recalls. “He caught it immediately and said, ‘They will know. They will feel it. They know I don’t talk like that.’”
Clark found out just how sharp our instincts for authenticity can be—outside the workplace, too. One night, too tired to invent a new story, she used AI to generate a bedtime tale for her 11-year-old daughter.
"She loves to hear stories, and there have been times where I've been tired and used AI to develop a story about garden fairies or whatever," Clark explains. "Not only did she know—she was critical about it. She said, 'Wait a minute, the last time you told a story about a tomato fairy was different.'"
Even a child could sense the disconnect.
It reinforced something she already suspected: people—of any age—have developed a sharp ear for AI-written content.
Whether it’s a CEO’s message or a parent’s bedtime story, when something feels off (even slightly), people disengage.
Remote and hybrid workforces create a paradox for AI-assisted communications.
Most employees aren’t fully present during meetings—our research shows 98% of knowledge workers multitask during virtual meetings. That kind of divided attention makes authentic, well-timed messaging more important than ever.
On one hand, teams desperately need AI's contextual intelligence. Different time zones require different messaging approaches. An announcement that works for your San Francisco office might completely miss the mark with your Berlin team. AI can help simulate those cultural and temporal contexts before you hit send.
But distributed work also makes authenticity detection more acute. When your primary connection to leadership is through digital messages, employees become experts at reading between the lines. They notice when communication feels manufactured rather than genuine.
The stakes get higher when you can't rely on hallway conversations or coffee shop clarifications to rebuild trust. One inauthentic message can ripple across global teams within hours, with no easy way to course-correct in person.
That's why Polly meets teams where they already work across platforms and time zones with tools that match the moment, whether that's through Zoom integrations for team meetings or Slack workflows for async collaboration.
AI works best when it helps you understand your people rather than when it tries to take your place in the conversation.
The teams that recognized these problems first made a crucial pivot.
Instead of using AI to create more messages, they started using it to understand their people better and treated it as a thinking partner rather than a content factory.
Steve Jobs once described computers as a ‘bicycle for the mind.’ Powered by your own energy, extending your natural capabilities rather than replacing them.
This reframe is reshaping how internal comms professionals approach AI adoption.
Instead of using AI to write messages, use it as a thinking partner.
Maybe you could feed it context about their audience and ask: ‘How would an engineer on my offshore team receive this announcement?’ Or simulate different personas to check for bias in their messaging before hitting send.
Bilal Aijazi, Polly's co-founder, sees this shift consistently: “You want to use the authenticity, the core empathy and humanity that people have, and then use AI to extend your reach.”
The bicycle analogy works because you still do the pedaling. AI just helps you go further, faster, with better balance.
Clark has developed what she calls ‘fake focus groups’—AI personas modeled after different employee segments that she consults before major announcements.
“I have these focus groups that I've set up that I return to regularly and say, 'Hey, if you're in this group, how would you feel about this?'” Clark explains. The approach automatically checks bias and surfaces blind spots that single-perspective messaging often misses.
The most sophisticated teams treat AI like that colleague who always asks the hard questions. Clark configures her AI agents to push back rather than please: “Don't tell me what I want to hear. Poke holes in what I give you.”
Running messages through multiple AI models—Claude for clarity, ChatGPT for reframing—creates a feedback loop that improves messaging before it ever reaches your employees. It's the difference between AI as a yes-person and AI as a strategic thinking partner that actually makes your communication stronger.
AI isn’t plug-and-play. To be useful, it needs to understand your team's language, your leaders’ tone, and the unspoken rules of your culture.The solution requires feeding AI your company's communication history that capture authentic company culture like:
By grounding AI in your actual communication style, it can deliver messages that sound like they came from your team.
No matter how well you train your AI, leadership comms demand a human eye. These are the messages employees read most closely, and they’re the quickest to sense when something sounds off, inauthentic, or too polished to be real.
In every conversation with internal communications leaders, one theme keeps surfacing:
Teams don’t need AI to write more. They need help managing what comes back.
One leader told us a single town hall generated thousands of questions. Another described the challenge of processing global feedback that arrives faster than any human can tag, sort, or act on.
What looks like a content challenge is actually a signal challenge
That insight is shaping how we approach AI.
We’re building features that enhance workflow intelligence—not replace human judgment. Think:
Just as important is what we won’t build.
We won’t automate sensitive communication like layoffs or crisis updates. We won’t push messages without human signoff. And we won’t remove the oversight that protects context, timing, and tone.
Instead, we’re focused on making authenticity scalable:
We design tools that help you deliver messages people read, understand, and remember while keeping your tone and intent intact as your company grows.
We’ve distilled our conversations with internal comms leaders into a practical approach: one that prioritizes learning over speed and clarity over automation.
It’s designed to support thoughtful AI adoption, giving teams space to experiment while maintaining the authenticity employees expect.
Before teams can use AI properly, they need to sharpen their judgment of what good AI output actually looks like. Bilal calls this skill “AI taste”—the ability to recognize when AI is helpful versus when it’s creating fluff or undermining credibility.
“AIs follow the taste of human beings. They don’t develop the taste of human beings,” he explains.
Strong AI taste means:
Without that distinction, teams end up publishing generic content that sounds robotic, tone-deaf, or off-brand.
Building this taste isn't a one-time training exercise. It develops through regular practice and feedback loops. Teams using systems like Polly benefit from ongoing signals about what resonates with their audience, refining their AI judgment with each message cycle rather than relying on initial instincts alone.
As AI models improve, so should your expectations. Teams that treat AI like an intern with oversight, feedback, and quality control will consistently get more value than teams chasing speed alone.
Instead of jumping straight into full automation, take a phased approach. Each stage builds confidence and skills before moving on to higher-stakes applications.
Start where the stakes are low but the insight is high. Use AI for:
This stage builds familiarity. The team stays in full control while using AI to pressure-test messaging and uncover blind spots before publishing.
Once your team is comfortable, introduce AI into repetitive, low-risk tasks:
These are efficiency gains that reduce manual overhead without touching sensitive communication. Your voice remains intact while your team regains time for strategic work.
Finally, bring AI into your distribution logic. At this layer, AI helps:
This is how you scale intimacy. Instead of blasting the same message to 10,000 people, you use AI to meet them where they are without diluting voice or clarity.
AI conversations often revolve around speed, but in your internal communications strategy, faster isn’t always better.
Publishing messages 50% faster means nothing if employee engagement drops by 20%. Research shows that when employees feel overloaded by information, understanding of company strategy drops by more than half, and only 6% report being highly likely to stay with their current company.
Instead of measuring productivity alone, track how communication lands:
Use these signals as early indicators of authenticity. Build regular pulse checks into your workflow with short, targeted asks like “Did this message feel genuine?” or “Was the intent behind it clear?”.
Regular pulse checks with different employee segments surface authenticity problems before they become widespread trust issues. Ask directly whether communications feel genuine and whether people understand the human intention behind messages.
The goal remains human connection at scale. If your AI implementation sacrifices that connection for efficiency, you've optimized for the wrong variables and created long-term engagement problems that no amount of speed can solve.
Curious how other teams are navigating AI in internal comms?
Watch our full conversation with Carolyn Clark. We break down real-world implementation strategies, mistakes to avoid, and what it really takes to build trust with AI in the mix.