● AI Agent Turns Meeting Notes Into Enterprise Power
The Real Reason Tiro, a Startup with KRW 4 Billion in Annual Revenue, Built an “Ontology,” and How AI Agents Change Organizations
Today’s Key Takeaways at a Glance
Tiro is not simply an AI meeting notes tool, but a company that turns conversations and work context into organizational assets.The key takeaways are ontology, AI agents, turning meeting notes into assets, security certifications, and B2B sales.On the surface, it may look like “a product that records meetings,” but in reality, it is building a system that organizes a company’s language so people and AI can work by the same standards.In today’s article, we will summarize in a news format why Tiro chose ontology, how personal agents reduce team bottlenecks, and the most important practical point that other news coverage often misses.
What Is Tiro Changing?
It does not stop at the stage of recording conversations.Tiro aims to safely store conversations, retrieve them when needed, and create a structure that allows agents to understand that context.Simply put, it turns meeting notes from “documents to read” into “data that supports the next task.”This flow aligns precisely with the intersection of generative AI, AI agents, workflow automation, enterprise SaaS, and digital transformation.
News Point 1. The Ontology Tiro Talks About Is Not a “Dictionary of Correct Answers”
Tiro’s CTO, Yeoul, does not see ontology as an overly grand theory.Instead, he sees it as a mechanism for organizing the language and rules inside a company that must never be misunderstood.For example, even with the same MRR, one person may interpret it as revenue from Japan, while another may interpret it as revenue from Korea.If the definition of a metric wavers like this, people may be using the same word in a meeting while looking at completely different numbers.
That is why Tiro first asks:What should be recorded?By what standards should decisions be made?Which actions should be automated?Who should have access?And what evidence should be left behind?
Only when these five elements fit together does Tiro elevate something into the ontology.In other words, it does not hard-code every task into AI rules, but instead strictly defines only the truly important work.
News Point 2. AI Agents Reduce Bottlenecks Rather Than Replace People
The Tiro team operates a personal agent for each team member.For example, Leo’s agent is Mio, and Yeoul’s agent is Winter.These agents first suggest responses that the team member might have given, and then the human reviews and revises them.
The reason this matters is simple.Work often stops not because “the person is unavailable,” but because “others do not know that person’s context.”This is especially true for context-heavy work such as B2B sales, responding to Japanese customers, contract reviews, and security reviews.
What Tiro has built is not a simple chatbot.It is a work-oriented AI agent that inherits a person’s work context.This is the next stage of workflow automation and a core keyword in enterprise productivity innovation.
News Point 3. Personal Agents Are Built from a Combination of Personality, Memory, and Company Rules
Tiro designs agents in three parts.
1) SOUL
This refers to a person’s style, such as tone of voice, tendencies, and when they intervene.It is built based on Slack messages, conversation habits, and work patterns.
2) Wiki
This is the individual’s long-term memory and project context.It organizes who did what, what relationships existed, and what reference materials are available.
3) Company-Wide Ontology
This includes organizational rules shared across the company, such as customer response standards, sales criteria, permissions, security, and audits.Content that must not differ from person to person is placed in this layer.
The advantage of this structure is clear.It can preserve both individual creativity and company-wide consistency.It can be seen as a real-world workplace application of what today’s AI trends call AI agent orchestration and enterprise knowledge graph.
News Point 4. Agents Handle Japanese B2B Emails, Morning Briefings, and Security Reviews
Hyeonje’s agent, Mio, operates based on Japanese B2B work context.Because it has accumulated Japanese email tone, honorific systems, the particularities of contracts and payments, and customer response rules, it creates drafts that can actually be sent, not just simple translations.
Yeoul’s agent, Winter, is somewhat more general-purpose.It sends company status reports early in the morning, organizes the day’s calendar and the people to be met, and briefs the user on what to prepare for the day based on previous meeting notes.It also helps with repetitive operational tasks such as security reviews, audit reports, and permission checks.
This is a significant change.It means AI is moving beyond being merely “a tool that writes well” and entering the practical layer that handles the front end of business operations.
News Point 5. Tiro’s Standard for Choosing Ontology Is Not “Automate Everything”
From Tiro’s perspective, ontology is not a container that holds all knowledge.Instead, it is applied only to “work that must be accurate.”
For example, email writing requires flexibility, so there is no need to rigidly define it through ontology.However, tasks where mistakes are unacceptable, such as purchase orders, contract amounts, permission changes, and refund processing, must be managed through ontology.
This standard is extremely important.That is because many companies begin AI adoption by thinking, “Wouldn’t it be good to automate everything?”In reality, that is not the case.What is truly needed is not automation itself, but a design that separates rules according to the importance of each task.
News Point 6. Pre-Ontology Is “Material Just Before a Decision”
What Tiro calls pre-ontology is not the ontology itself.It is closer to the material before rules are defined, such as conversation records, meeting notes, context, and relationship information.
The reason this matters is that before building an ontology, a company must examine what was actually said inside the organization and what hidden definitions exist.In other words, conversations do not replace agreements; they provide the evidence and context needed to create agreements.
This perspective is also very important when understanding AI search, enterprise knowledge management, internal wikis, knowledge graphs, and RAG systems.More important than simple storage is a structure that can retrieve context.
News Point 7. The Core Point That Makes Agents Trustworthy Is Not the “Answer,” but the “Evidence”
Tiro shows which tools the agent used, which wiki it referenced, and which rules led to the conclusion directly on the screen.In other words, it does not merely present the result; it exposes the evidence and path together.
There are two reasons this approach matters.
First, it can reduce AI hallucinations.Second, when a human reviews the output, they can immediately see where to look.
Especially in contracts, security, refunds, and legal reviews, “what evidence led to this answer?” is more important than “is the answer correct?”That is why Tiro does not hide logs in a separate admin page, but instead makes review happen directly inside the actual work screen.
This is a highly practical approach from the perspectives of AI governance, audit trails, and explainable AI.
News Point 8. Security Is Not a Later Issue, but the Entry Ticket for B2B Growth
The reason Tiro invested in security early is clear.B2B is closer to a market where customers adopt a product only when “there is no reason not to use it,” rather than simply saying, “If it is good, we will try it.”In other words, even if the product is strong, adoption itself can be blocked without security certifications.
In fact, after obtaining SOC 2 and ISO certifications, the atmosphere of initial conversations with customers reportedly changed.Previously, Tiro had to explain security at length from the beginning, but now it can begin conversations on a foundation of basic trust.
There is one point that must not be missed here.Security certification is not the end.Rather, it is a trust mechanism that helps a company cross the first threshold of sales.This is especially true in global B2B, the Japanese market, and enterprise SaaS.
The Most Important Point Other News Often Misses
The most important point is not that “Tiro sells AI meeting notes.”The real core point is that it is building a structure where a company’s decision-making standards remain even when people leave.
This matters because, even after adopting AI, many organizations still experience bottlenecks due to the know-how stored inside individuals’ heads.In other words, if a company remains slow even after adding AI, the reason may not be model performance but the fact that the organization’s standards have not been documented.
Tiro is solving this problem by connecting conversation records, wikis, ontology, and agents.This is not just a productivity tool, but closer to an attempt to build an operating system for organizational intelligence.
Why This Matters from the Perspective of the Global Economy and AI Trends
The global economy is currently in a phase where economic slowdown, cost reduction, and workforce efficiency are all applying pressure at the same time.In this environment, companies tend to focus less on simply hiring more people and more on how much they can increase the productivity of their existing organizations.
That is why AI agents, workflow automation, enterprise software, knowledge management systems, and security compliance are all attracting attention at the same time.In particular, B2B SaaS, enterprise AI, workplace LLMs, AI meeting notes, knowledge graphs, and digital transformation are likely to remain strong trends going forward.
Tiro’s case is a classic example of this shift.It signals that generative AI is moving from “a tool that writes well” to “a tool that supports organizational judgment and execution.”
In Summary
Tiro is not a company that simply creates meeting notes, but a company that turns conversations and work rules into organizational assets.Ontology is not a rule system that controls everything, but a mechanism that accurately defines only the work that must be aligned.Personal agents reduce bottlenecks rather than replace people, while a shared ontology protects company-wide consistency.And security certifications are a practical trust mechanism that enables B2B adoption.
Summary
Tiro is an AI company that turns conversation records into assets and unifies organizational standards through personal agents and a shared ontology.The key is not meeting notes, but a structure that preserves context, rules, and evidence.Security certifications and explainable agent operations are central to B2B expansion.
[Related Articles…]
How AI Agents Are Reshaping Workflow Automation
Why Enterprise Ontology Matters for Business AI
*Source: https://maily.so/josh/posts/d5rywxx1z1w?from=email&mid=pzlq3932mrk
