AI Characters in Corporate Communication: Strategic Advantages, Risks, and Avatar Design Choices

Nikki Callahan

Imagine you open your laptop at work and instead of a faceless slide deck, there’s a consistent, friendly guide who walks you through every new process, every quarterly update, every safety module. Same person, same energy, in any language you need… but they’re not actually a person. That’s what we mean today by “AI characters.” Digital presenters with a body, a face, a voice, and a style that can show up across your company on demand.

John Harvey

Yeah, and they’re not just floating heads for marketing anymore. An AI character in 2026 is basically a programmable communicator. Under the hood you’ve got text, video, voice synthesis, sometimes a live engine streaming them in real time. On top of that, you layer brand, tone, and behavior so they feel like a recognizable “someone” whenever they appear in training or internal comms.

Nikki Callahan

If we roll the clock back a few years, these were mostly gimmicks. Cute landing-page greeters, little one-off campaign mascots, maybe a one-minute announcement video someone generated in Synthesia and dropped into Slack. They were like digital party tricks. People would say, “Oh wow, it moved its lips in Spanish too,” and then go right back to sixteen PowerPoint decks.

John Harvey

And now they’ve quietly become plumbing. You see it in the enterprise stack. Zoom, for instance, has been pushing further into intelligent meeting tools—automated summaries, AI assistants in the call. That’s infrastructure. Synthesia takes a script and turns it into a presenter-led video in dozens of languages, using either stock avatars or a custom-recorded one that looks like your CMO or your head of safety.

Nikki Callahan

And then you’ve got players like AKOOL stepping in with this more photoreal, avatar-centric stack. So instead of just “upload script, download video,” you can build a specific character: maybe it’s your CEO’s digital twin, maybe it’s an archetypal trainer, maybe it’s a host that only exists in your brand universe. That avatar can be reused in hundreds of clips or even streamed more like a digital human, rather than frozen into one canned video.

John Harvey

Exactly. The pattern looks something like this: first layer is script-to-video—fast, templated training content. Then custom or personal avatars—your people, your brand, showing up consistently. Layer three is multilingual—same character, different languages, no need to fly an executive to six studios. And the frontier layer is streaming digital humans, where that character can respond, at least somewhat interactively, inside a Zoom call or a browser.

Nikki Callahan

The reason all of this is catching on is very… unromantic. Cost, speed, and consistency. Most companies don’t wake up saying, “Let’s explore the nature of identity in a digital age.” They wake up and say, “We need ergonomic training, in four countries, by the end of the quarter, and our one charismatic trainer is fully booked.”

John Harvey

Right. Human presenters are scarce. The good ones burn out, get promoted, or just… don’t scale. But the enterprise still needs a reliable face for compliance, onboarding, product updates. AI characters give you something close to a “persistent presenter” who never loses their voice, never has jet lag, and doesn’t need three days to reshoot a line.

Nikki Callahan

And that persistence is huge for learning. In the coaching world I see this a lot: if people trust the guide, they’re more willing to lean into discomfort. There’s some early ergonomics training work where digital human presenters performed about as credibly as traditional online formats. Not magic, not better-than-human in some sci-fi way, but solid… “good enough to carry the message.”

John Harvey

Yeah, the signal there is: we don’t have to wait for perfect holograms for these to be useful. Even relatively simple digital humans can deliver structured, repeatable content just fine. When you combine that with rapid iteration—you can tweak a script and push a new video globally in hours—you start to see why this becomes infrastructure, not novelty.

Nikki Callahan

So if you imagine a company nervous about yet another tool: this isn’t replacing every manager, it’s becoming the connective tissue. A communications layer that can carry policy, training, brand stories… at scale. And the question for leaders isn’t “Will this arrive?” It’s “Where do we plug these characters in without losing our humanity?”

John Harvey

And how do we do it strategically instead of just letting a hundred random avatars bloom in every department. Which is where we’re headed next.

Nikki Callahan

Let’s lay out the upside first, because otherwise this can sound like a Black Mirror episode. When you and I were sketching this, John, we landed on five big advantages leaders keep coming back to.

John Harvey

Yeah. Number one: scalability. Once you’ve designed an AI character and a workflow around it, adding ten more videos isn’t ten times the effort. It’s almost a marginal cost problem—one more script, click render, you’re done.

Nikki Callahan

Second is message consistency. You’re not relying on twenty different trainers all improvising slightly different metaphors for the same safety rule. The AI character says it the same way, every time, in every region. There’s a kind of… mantra quality to it.

John Harvey

Third, rapid revision. If the policy changes, you don’t rebook a studio and coordinate calendars. You change the script and regenerate. Tools like Synthesia and newer stacks like AKOOL make it routine to version content quickly, which matters a lot in fast-moving environments—regulatory change, product launches, crisis comms.

Nikki Callahan

Fourth is scalable expert presence. This one feels almost mystical to me. You can capture a subject-matter expert or an executive once, and then their digital presence can show up everywhere—factory floors, home offices, different time zones—without that human having to physically be there.

John Harvey

And fifth, potential instructional effectiveness in repeatable learning contexts. We’re seeing that when the content is structured—ergonomics, compliance, step-by-step procedures—a well-designed digital human can be just as effective as a typical online course. The value isn’t that it’s “cooler,” it’s that it’s repeatable and always available.

Nikki Callahan

Now, that’s the bright side of the mountain. The shadow side is just as important. First big risk: authenticity degradation. If your workforce starts feeling like the company only speaks through synthetic faces, trust erodes. People start to wonder, “Do real leaders still stand behind this message?”

John Harvey

Yeah, and sitting right next to that is the uncanny valley problem. Photoreal but not quite right can trigger this subtle discomfort. Eyes that don’t blink quite naturally, smiles that are just a bit off. Over time, that can become distracting or even creepy, and suddenly no one’s listening to the actual content.

Nikki Callahan

Then there’s governance of likeness and voice. Who owns the executive’s digital twin? What happens when that person leaves the company, or their role changes, or, you know, life happens? Without clear agreements, you can end up in messy ethical and legal territory, replaying someone’s synthetic face in contexts they’d never consent to now.

John Harvey

And the fourth piece here is disclosure. Employees and customers deserve to know when they’re hearing from a synthetic presenter. If you blur that line—even unintentionally—you undermine long-term trust in every other channel. So you need simple, clear cues: labels, intros, design choices that say, “Hey, this is an AI character guiding you.”

Nikki Callahan

A-KOOL is an interesting example because it kind of turns up the volume on both sides of that equation. On the one hand, photorealistic, streaming avatars open up new use cases: live training with a digital facilitator, branded support agents that feel more present, a consistent digital host for your internal town halls.

John Harvey

On the other hand, the more real it looks and the more it behaves like a person—in real time—the heavier the governance burden. You can’t treat that as a throwaway asset. It’s close enough to “a real colleague on screen” that you have to ask: what guardrails do we set? How do we log what that avatar is allowed to say? Who signs off on its scripts?

Nikki Callahan

It’s a bit like martial arts, actually. You start with simple forms—slow, contained moves. Then as you gain power, you’re expected to take on more responsibility. Enterprises playing with streaming digital humans are stepping into black-belt territory. They need the discipline to match the capability.

John Harvey

And that’s the mindset shift. Don’t treat AI characters as toys in the corner. Treat them as a new, powerful communications layer that can either deepen trust and clarity… or quietly corrode both if you rush in without a plan.

Nikki Callahan

So let’s say you’re convinced these characters matter. The next question is: what kind of character? Human, human-like, or cartoonish and stylized? Because those choices aren’t just aesthetic, they’re psychological.

John Harvey

Yeah. Human-like avatars—photoreal faces, realistic motion—tend to carry authority. Whether we like it or not, our brains treat them as “someone in charge.” They’re useful when you want executive continuity: your CEO can have a digital presence in every office without physically teleporting.

Nikki Callahan

But when those human-like avatars land in that uncanny valley, all that authority gets weird. People focus on the strangeness instead of the message. Research over the years keeps pointing to this: almost-real humans can be more off-putting than clearly stylized ones.

John Harvey

That’s where cartoon or more stylized characters shine. Think of classic mascots and character branding. They’re not pretending to be human; they’re owning their “otherness.” That makes them warm, approachable, and very clearly part of the brand. You can give them exaggerated expressions, a playful tone, and people tend to accept them on those terms.

Nikki Callahan

There’s also an emotional buffer. If a cartoon-ish guide walks you through a frustrating process—expense reports, for example—it can soften the blow. The character becomes a kind of mediator between the cold system and your very human annoyance. In my workshops, I see the same thing with archetypes; a symbol can absorb tension that might otherwise land on a real person.

John Harvey

So let’s turn that into a rough decision matrix. When do you use realistic avatars? I’d put them in the “authority transfer and executive continuity” bucket. Policy announcements, strategy walk-throughs, maybe critical safety messages—places where you want the weight of leadership but can’t get the actual leader into every room.

Nikki Callahan

When do you go stylized? Anywhere brand mediation and warmth matter more than authority. Onboarding tours, product education, ongoing micro-learning, help-center flows. A recurring character that’s clearly part of your brand can become ownable IP—a kind of internal guide or external mascot people genuinely look forward to seeing.

John Harvey

And when do you insist on real humans? For me, it’s emotionally sensitive or accountability-heavy messages. Layoffs, serious misconduct issues, big cultural shifts. Anything where you expect questions like “Who decided this?” or “Do you, personally, stand behind this?” That’s where a living, breathing leader should show up on camera or in person.

Nikki Callahan

I’d add healing moments to that—company grief, community crises, those raw times when people need unscripted pauses and visible emotion. No avatar, no matter how advanced, can replace eye contact and a cracked voice in those moments.

John Harvey

So if you’re rolling this out, start simple. First recommendation: pick low-risk, repeatable use cases. Think standard ergonomics training, tool tutorials, routine reminders. Let your AI character earn trust on the boring stuff before you point them at culture-defining messages.

Nikki Callahan

Second, separate character governance from content governance. One team or process decides what the avatar is allowed to be and do—likeness rules, tone, disclosure. Another governs what it actually says—policies, scripts, legal review. Mixing those up is how you end up with “funny safety mascot accidentally comments on layoffs.”

John Harvey

Third, when you experiment with realistic avatars, test for trust, not just for visual polish. Don’t only ask, “Does it look real?” Ask, “Would you follow this person’s instruction? Do you feel they represent the company honestly?” Sometimes a slightly less photoreal style actually wins on comfort and clarity.

Nikki Callahan

And finally, think hybrid architecture. Match avatar type to message type. Real humans for the high-stakes, human-like for authority and continuity where it’s safe, stylized characters for ongoing education and brand experiences. Let each form do what it’s best at instead of chasing one “perfect” digital human to do it all.

John Harvey

If leaders approach this with that kind of nuance, AI characters stop being this scary, abstract future thing and become just… another part of how we tell stories and share responsibility at scale.

Nikki Callahan

And they can free up actual humans to do the work only we can do: listening, improvising, holding space when things are hard. That’s the piece I never wanna see automated away.

John Harvey

Couldn’t agree more. Nikki, this was fun. I feel like we only scratched the surface, but hopefully it gives people a map, not just a buzzword.

Nikki Callahan

Yeah, and we’ll keep circling back as this space evolves—especially as tools like AKOOL and others push more into live, streaming characters. For now, thanks for being here with us. John, take care.

John Harvey

You too, Nikki. And thanks to everyone listening. We’ll see you in the next one.