From Feed to Conversation:How AI Brand Persona Continuity Shapes Social Presence and Consumer Trust

Imagine you are scrolling through your favorite social media app. Suddenly, a bright, friendly digital character pops up. Maybe it's a cool robot or a stylish animated avatar that tells jokes, shares fashion tips, or explains how a new sneaker is made. You start to like this character. You feel like you "know" them, even though you've never met them in real life. In the world of science, this feeling is called a parasocial relationship. It's like having a one-sided friendship with a TV host or a YouTuber; you feel close to them, but they don't actually know you back.

Now, imagine that same friendly character pops up again, but this time, it's inside a customer service chat window. You have a question about your order, and you start typing. The big question researchers are asking is: Does that feeling of friendship you built while scrolling your feed carry over to the chat? Does it make you trust the robot faster? Or, if the robot acts weird in the chat, does it make you feel more betrayed than if you had never met them before? This paper dives into that exact moment where a "feed" (what you see) turns into a "conversation" (what you say), exploring how our brains handle these digital friends when they switch roles.

---

From Feed to Chat: The Digital Friend Switch

This paper introduces a new idea called the PACT Model (which stands for Persona-Avatar Continuity-Trust). Think of it as a rulebook for how brands can use the same AI character in two different ways without messing things up.

The authors suggest that when a brand uses an AI character (let's call it "Rex") to post fun videos on Instagram, you start to get familiar with Rex's voice, style, and personality. This is the Feed stage. Later, when you need help with a product, Rex shows up as a Chatbot. The paper argues that if Rex looks and sounds exactly the same in the chat as he did in the videos, your brain gets a head start. It's like walking into a coffee shop and seeing your favorite barista; you immediately feel more comfortable and trust them to make your drink right.

The researchers call this the "Persona Carryover Effect." It suggests that because you already "know" Rex from the videos, you feel a sense of warmth and social presence right away when you start chatting. You don't have to spend time figuring out if the robot is nice or mean; you already think it is. This helps build Affective Trust (trust based on feelings and warmth) much faster than if you were talking to a brand-new, faceless robot.

The "Uncanny Valley" of Expectations

However, there is a catch. The paper warns that this only works if the robot is actually good at its job.

Imagine you meet a super-smart, friendly robot in a video. You think, "Wow, this guy knows everything!" Then, you chat with him, and he gives you the wrong answer or forgets what you just said. Because you had such high expectations from the video, you feel more disappointed than if you had just talked to a boring, generic robot that you didn't expect much from.

The authors call this the "Continuity Violation Effect." If the AI character drifts away from the personality you loved in the videos (maybe it starts acting rude or stupid), that prior friendship actually hurts you. It makes the bad experience feel worse. The paper suggests that for this "carryover" to work, the robot must be competent first. It has to answer your questions correctly before your feelings of friendship can kick in. If it's not smart, the "friendship" doesn't help; it just makes the failure feel like a betrayal.

The "Who Are You?" Question

Another big part of the story is Disclosure. This is about when and how the brand tells you, "Hey, I'm actually a robot."

The paper suggests that telling you the truth matters, but how you tell it changes everything. If a brand hides the fact that it's a robot until the very end, you might feel tricked. But if they tell you clearly at the start, "Hi, I'm Rex the AI," and you still feel a connection, that's a good sign. The paper notes that this depends on what you are buying. If you are buying something simple, like a pen, you might not care as much. But if you are buying something risky or expensive, like a financial service, you might want to know immediately that you are talking to a machine, not a human.

Choosing Your Own Adventure

The paper also talks about Choice. Imagine a brand has three different AI characters: a serious one, a funny one, and a sporty one. If you get to choose which one you want to chat with, you feel more in control. The authors suggest that when you pick the character yourself, you trust them more because it feels like your choice, not something forced on you. It's like picking your own avatar in a video game; you feel more connected to the one you picked.

What the Paper Doesn't Say (Yet)

It's important to remember that this paper is a proposal, not a final answer. The authors haven't run the big experiments yet to prove these ideas are 100% true. They are saying, "Here is a map of how we think this works, and here is how we should test it."

They explicitly rule out the idea that "familiarity is always good." They argue that if the AI is inconsistent, being familiar can actually make things worse. They also suggest that not everyone wants a "friendly" robot; some people prefer a cold, efficient machine because they think it's more honest and less likely to be biased.

The Big Picture

So, what's the takeaway? The paper suggests that brands can't just slap a cute robot on a video and then expect it to magically become a helpful customer service agent. They have to make sure the robot is consistent (same voice, same values), competent (actually good at the job), and honest (tells you it's a robot).

If they do all that, the "friendship" you built on social media can help you trust the brand faster. But if they mess up the consistency or the competence, that same friendship can turn into a bigger disappointment. The authors are calling for more research to test these ideas, especially in different cultures, to see if this "digital friendship" works the same way everywhere.

In short, the paper is a guide for how to be a good digital friend: be consistent, be helpful, be honest, and let people choose their own path. If you do that, the trust follows. If you don't, the "friend" might just become a stranger you never want to talk to again.