LLMOcheki launches Quest Mode, turning improvement work into something a team can keep up

Itera, Inc. (head office: Shinjuku-ku, Tokyo; Representative Director: Takayuki Muto) has begun offering Quest Mode in LLMOcheki, its tool for measuring and improving LLMO. The feature turns the work of improving AI search visibility into a game format that a team can keep going.

Experience points are awarded for carrying out the work, and only where exposure actually improved. A pixel-art character, Cheki, grows with the team’s progress and suggests what to do next.

Background: AI search work is hard to keep up

Improving visibility in generative AI search takes time to show in the numbers after the work is done. Publish an article, or press for a listing in third-party media, and the figures do not move the next day.

As a result, a daily improvement cycle often fails to take hold even after an analytics tool is adopted, and the work stops after a few months. All the more so when one person is carrying it alone. What has been done is not shared within the team, and the effort is forgotten before results arrive.

Quest Mode addresses that problem of continuity.

What it does

1. Improvement work is presented as quests

The work to be done appears as a quest, and experience points are earned once it is confirmed as done. The level shared by the team rises and the pixel-art character, Cheki, grows. Because progress accumulates for the team rather than the individual, the effort carries over when the person responsible changes.

2. Results earn experience too

Experience points are awarded not only for the number of actions carried out but also where the rate of appearing in AI answers actually improved. Both the volume of work and the result are recognised.

3. The reliability of the numbers comes first

Every award of experience is verified server-side; pressing a “done” button alone does not add anything. The appearance rate is measured weekly under fixed conditions, and a week with insufficient data is shown as “aggregating”.

That a mechanism for encouraging continuity must not compromise the rigour of the measurement is treated as a design premise.

4. It suggests what to do next

The character is not merely decorative: it suggests the next thing to work on based on the measurement data. Its expression changes with what is being discussed.

5. There is no penalty for failure

The level does not fall when results are flat. There is no disadvantage for breaking a streak and no penalty for not doing something. Rather than using pressure to force short-term action, the emphasis is on being able to keep going without strain.

Availability

Being opened up in stages to tenants using LLMOcheki.

What comes next

We will progressively add mechanisms that support continuity, including a rewards programme at milestone levels and features for working together as a team.

Comment from our Representative Director

“AI search work goes wrong not because people do not know how, but because it does not last. Results take time, and in the meantime the person’s attention drifts away. But a mechanism for keeping going that works by softening the numbers would defeat the purpose, so every judgement on experience points goes through server-side verification.”

LLMOcheki https://llmocheki.com/