Tanka is an AI business chat tool for teams that has "team memory" capabilities. Its core idea is: AI used by individuals can make people faster, but work often stalls between people; Tanka, by enabling teams to share memory, drives the work of the entire team forward.
[Purpose] Tanka is designed to solve the problems in team work of "missing things, forgetting things, and not following up." For example, group communications that get scrolled past and not seen, a proposal that was agreed to be sent that day but was forgotten, a quote sent out that gets no response for a long time, or an unclear background on a case handled by a predecessor. Tanka collects information from chats, emails, files, and meeting minutes, remembers agreements and deadlines, discovers stalled work and proactively reminds, thereby supporting scenarios such as customer follow-up, work handover, and new employee onboarding.
[Main capabilities]
- Accumulation and use of team memory: Accumulate conversations, emails, and agreements as team memory, with no need to repeat the same explanations.
- Automatic information collection: Automatically collect scattered information from chats, emails, files, and meeting minutes.
- Context and memory recognition: It can recognize who "Mr. Sato" specifically refers to based on context and memory, for example determining from among three Satos that it is Mr. Sato from the memory factory, and advance the work.
- Proactive reminders: Remember agreements and deadlines, discover stalled work, for example prompting, "Mr. Sato has not replied yet; would you like to send a confirmation email?"
- Prepare the next step: Not only can it draft a reply, it can also prepare what needs to be done next, for example preparing the content of an email to Mr. Sato so it can be sent after confirmation.
- Email and file handling: Through instructions, Tanka can complete the process from drafting an email to attaching files.
- Answers with sources: Answers are given with sources, and next actions are prepared.
- Permission linkage: It links with access permissions shared by the team, and protects the scope of information sharing according to role.
- Cross-owner judgment: It reflects the latest rules and case history shared by the team in judgments spanning across owners.
- Handover support: Even if the owner changes, team memory and case history can be inherited as-is by the next owner's AI.
- Collaboration among members: Each member's AI passes work context to one another, supporting collaboration with the necessary people.
[How to use] According to the official explanation, Tanka can be started without changing existing tools, and it will gradually understand people, cases, and decisions. The materials mention tools such as Microsoft Teams and Slack, indicating that it can be used in connection with existing chat and email tools, but the specific connection methods and operating steps are not yet provided. The official offering allows starting for free, with no initial setup required, and supports individualuse, and consultation or requesting materials for team adoption is also possible.
[Applicable scenarios] The materials list multiple scenarios: if you want to know what happened in the past, what was agreed, and where things currently stand, you can ask Tanka and know immediately; without tracing back through Slack, email, or meeting minutes, you can understand "why the situation is the way it is now"; new employees do not need someone to explain the background from scratch, because team memory can become their first "senior colleague." In addition, it is also suitable for customer follow-up, work handover, and new employee onboarding.
[Limitations and security] The official statement says that features vary by service, plan, and settings. Tanka has obtained ISO/IEC 27001:2022 certification, the international standard for information security management systems, and has completed an independent third-party SOC 2 Type II audit, demonstrating data protection and operational transparency. Customer data will not be used for AI model training.
Regarding specific pricing, feature differences between plans, supported platforms, and the integration list, the materials do not yet provide this information.
