Hakko AI User Case | Monthly Collaborators Grow from 10 to Nearly 100

Why a $100 Micro-Creator Outperformed a $20,000 Top Influencer？

This is not a simple story about “low cost and high efficiency.” It reveals a growth pattern emerging from the AI gaming sector: the core of creator marketing is not buying exposure, but buying content — more precisely, buying the probability that a piece of content becomes a breakout hit.

01｜Meet Hakko AI: Your AI Gaming Companion

Hakko AI is an AI-powered gaming companion product. In the company’s own words, it is “your AI gaming buddy.”

The product takes the form of a floating “desktop pet” that sits alongside the gameplay screen. By understanding what is happening in the game in real time, it proactively interacts with players through voice or text. Users do not need to devote 100% of their attention to the AI, but it is always present. The inspiration comes from Bilibili-style bullet comments: they do not dominate the main experience, yet they create a sense of companionship through shared resonance.

Wang Bihao, also known as Oratis, is the co-founder and COO of Hakko AI, where he leads product design and go-to-market. In his view, AI will not only reshape productivity, but also transform the entire entertainment experience.

“In the internet era, companies competed for users’ time. In the AI era, that logic no longer applies.” Hakko AI has chosen a direction that does not compete for attention: companionship runs in parallel; it does not replace the original experience.

To date, the overseas version of Hakko AI has accumulated more than 2 million registered users, with around 300,000 monthly active users. Its China version, “Doudou Game Partner,” has accumulated approximately 13 million users.

02｜From Coaching to Emotional Value

In its early days, Hakko AI focused on a “coaching” feature — teaching players how to play games better. But the team quickly realized that this approach did not work.

“If someone truly wants to improve at a game, the issue is usually not that they don’t know what to do. What they need is extensive practice and repeated conditioning,” Oratis said. “Even when an esports coach teaches you, what they are really doing is telling you to go practice. It is not teaching in the conventional sense.”

The team then shifted its focus toward emotional value: amplifying players’ sense of achievement and providing companionship throughout the gaming experience.

“What players want ultimately comes down to two things: an amplified sense of achievement, and a feeling of companionship during the experience.”

The product direction was right, but a new growth challenge emerged: the value users received was highly personalized. Everyone interacted with the AI differently, making it difficult to distill the experience into a single, unified marketing asset.

03｜Growth in Gaming Requires Leverage

Unlike most AI tools, gaming targets an extremely broad audience.

“If you are building AI design tools or AI video generation products, your target users are relatively vertical — video editors, designers, and so on. But in China alone, there are 700 to 800 million gamers, and the global audience is even larger.”

For such a massive audience, performance advertising is not necessarily the optimal solution. Oratis identified a mismatch:

“Users come in because a selling point is directly exposed to them, not because they already have that need. And performance advertising does not create lasting assets.”

Creator marketing works differently.

“First, creator marketing has leverage. Once you work with one strong creator, other creators see it and think, ‘I can make something like this too.’ Second, it creates accumulated assets. Creator content can be reused and further amplified.”

More importantly, Hakko AI separates creator collaborations into two distinct models.

“One model is buying traffic: you look at CPM, country distribution, audience distribution, and ultimately you are buying views. The other model, which I think is better, is buying the content itself — buying the probability that it becomes a hit.”

“For example, suppose you buy 100 pieces of content at $1,000 each. You spend $100,000. If you calculate CPM alone, it may look much higher than the industry average. But if one or two of those videos reach 10 million views, the actual outcome far exceeds directly buying exposure. If you work with creators only to buy exposure, you might as well buy ads. The logic of buying content is what truly creates both leverage and compounding value.”

But this logic has one hard prerequisite: sufficient volume. If a brand can only secure a few creator collaborations per month, then “buying 100 pieces of content to find breakout hits” is just an empty slogan.

04｜Before AhaCreator: Efficiency and Communication Bottlenecks

Before adopting AhaCreator, Hakko AI mainly relied on its in-house team to manually manage creator collaborations. The company had also worked with several domestic creator agencies.

In terms of efficiency, the traditional approach was extremely labor-intensive.

“The old model was to first compile a large creator list, evaluate it, then go to YouTube to find contact information, send emails, and wait. Most creators either did not respond or rejected the offer. Ninety percent of the time was wasted.”

At that time, the team could complete fewer than 10 collaborations per month.

Oratis was also candid about the agency model:

“Their approach is to prepare a warm pool and a cold pool of creators, then select creators from those pools based on your proposal and provide quotes. A common industry problem is lack of price transparency. The markup can be quite high, and everyone knows this.”

In terms of tools, the team had used third-party creator data platforms such as NoxInfluencer, Newrank, and HuoShaoyun to review creator metrics.

“But none of these provided a one-stop model that could truly deliver against its promise. Brands still had to build relationships themselves and negotiate with creators one by one. Running KOL marketing in-house requires a significant investment in manpower.”

Hakko AI also operates in global markets, which made communication even more difficult.

“It is not just a language issue. Even if you can speak Spanish, the communication style may still be wrong. You need a Spaniard to communicate with another Spaniard. In the past, the solution was to hire local teams in each country, then recruit Chinese employees from foreign language universities who spoke those languages, and combine the two teams. The cost was extremely high.”

05｜What AhaCreator Changed

Hakko AI discovered AhaCreator through a friend’s recommendation.

“It came through referral and word of mouth. That kind of recommendation is the most efficient because it comes with trust.”

For Hakko AI, the most important changes brought by AhaCreator can be summarized in several areas.

1. A Step Change in Execution Efficiency

AI agents took over the entire execution chain, from creator discovery and outreach to multi-round price negotiation, content follow-up, and data collection.

The brand team’s workload was compressed into four key decision points: creating the campaign, reviewing the creator list, reviewing creator content, and optimizing strategy based on performance data.

What the team received was a list of creators ready for collaboration, complete with real quotes, audience profiles, and estimated performance. The only decision left was simple: should we work with this creator or not?

Monthly collaboration volume increased from fewer than 10 creators to nearly 100.

2. Cross-Language Communication Was No Longer a Barrier

AI agents communicate with creators around the world in their local languages, with personalized messaging from initial outreach to multi-round negotiation. Brands no longer need native-speaking staff in every country.

“For AI now, both language and communication style have been connected. Whether it is writing briefs or communicating with KOLs in different countries, AI can express things in a way that may even be more efficient than humans.”

3. Greater Price Transparency

Each creator’s quote, audience data, and estimated CPM are clearly visible. Brands can make decisions with complete information, without having to guess how much markup may have been added in the middle.

4. An Extremely Lean Team Structure

The team setup is simple: one person in China, one person overseas, plus two overseas interns.

“Productivity per head is already at a very high level.”

06｜$100 vs. $20,000

1. Creator Selection: Prioritize an Audience-First Mindset, Not Follower Count

“If a creator’s content is going to go viral, that creator must have an audience-first mindset. The content they create is not designed only for their own traffic; it is designed for their viewers.”

Hakko AI once worked with a Chinese Genshin Impact guide creator called “Your Yingyueyue.” On average, every 50 views brought in one new registered user.

“ The reason is that he thinks entirely from the player’s perspective and considers what players actually need. This is something even we sometimes fail to do in marketing. We have KPIs, so we can easily end up looking at problems from a KPI perspective instead of a user perspective. ”

Oratis also noted that creators are, in essence, market-facing product managers.

“To grow their followers and views, they are constantly studying their audience. Sometimes, creators understand our product even better than we do.”

Creators do not necessarily need to say only positive things.

“Sometimes pointing out your problems makes the content feel more authentic. The key is that after they point out the problem, you fix it quickly. That can be even more effective than saying only good things.”

The $100 breakout video validated the same logic. The creator had only a few hundred followers. In the video, they showed a real reaction to interacting with Nika, a sharp-tongued AI character, inside GTA. The video generated hundreds of thousands of views and several hundred comments.

"You do not need to focus too much on follower count. A creator with only a few hundred followers can still generate hundreds of thousands of views in the U.S.”

2. Briefs: Explain the Product, but More Importantly, Explain the Creator’s Benefit

A strong brief has three core components:

“First, explain your product clearly. Second, explain the collaboration model. Third, think from the creator’s perspective and explain why they should work with you.”

For top-tier creators, Hakko AI even customizes briefs.

“For example, when I DM a creator, I might say, ‘I follow your content every day, and I think several directions in your content are highly aligned with what we are building.’ The creator’s goal is to serve their fans better. We need to help them serve their fans better.”

3. For Long-Form Video, Buy Content; For Short-Form Video, Buy Traffic

“Long-form video is about content quality and the value created by the content. Short-form video is about the ability to capture attention within the first three to five seconds.”

In Hakko AI’s practice, long-form video follows more of a “buying content” logic, emphasizing compounding effects and audience-first content quality. Short-form video follows more of a “buying traffic” logic, where the key question is whether the opening can instantly capture the user’s attention.

4. Do Not Operate Country by Country. Go Global Directly

Oratis said:

“In the AI era, there are only two markets in the world: mainland China, which needs to be handled separately, and everywhere else, which should be treated as one market. Many people mistakenly think that 40% to 50% of a creator’s audience comes from one specific country. That is almost impossible. For highly successful global products, even the largest single market usually accounts for only 10% to 20%. When you restrict yourself too tightly, the channel becomes narrower, and performance is very likely to suffer.”

07｜Personalized Products Require Distributed Communication

Toward the end of the interview, Oratis shared a broader view.

Hakko AI recently launched a UGC feature that allows users to create their own AI companion characters. As a result, more than half of users marked their characters as “private.”

“In this era, creating content itself is a form of consumption.”

He believes that the core logic of the internet era was the “shelf”: selecting the content best suited to you from a massive content library. The logic of the AI era is real-time generation — creating personalized experiences directly based on your current state.

When the product experience itself is highly personalized, the way it is communicated must also change. A single brand advertisement cannot replace one hundred creators showcasing their authentic experiences in one hundred different ways.

For Hakko AI, AhaCreator did not merely solve the question of “how to find creators faster.” It made it operationally feasible to cover global users with a large volume of diverse content.

Monthly creator collaborations grew from fewer than 10 to nearly 100. Behind this was a strategic shift in growth logic: from selecting a small number of top influencers to running scalable tests and using content volume to discover breakout hits.

More and more AI products are entering global markets, and they all face the same challenge: the more personalized the product, the more growth depends on authentic, diverse, and distributed user expression.

And the prerequisite for making that expression possible is simple: execution efficiency must be able to keep up with strategic ambition.