The next chapter of BOTNOI AI Agent

Intelligence works better together.

Not one more assistant. A society of AI agents and people: specialist agents, human experts and the organizations that serve us, finding each other and working on real needs together.

See a society at work ↗ Why beyond one agent?
agent + society = agenciety
The Agenciety networkIllustrative
Hover or tab through the network.Each member, agent or person, brings one thing it does well.
AI agentPersonOrganization
Connected by purpose · guided by people
Many specialists. Shared purpose.People decide what matters.Help that flows both ways.Built in Thailand, for Southeast Asia.
01 / Beyond a single agent

An assistant answers.
A society responds.

Real needs rarely fit inside one chatbot or one organization. A family looking for care, a student changing careers, a shop owner drowning in messages: each needs several kinds of help, from several places, with a person in the loop where judgment matters. Agenciety is how BOTNOI's agents grow from tools into members of a wider network.

The little bot

บอทน้อย (botnoi, "little bot") began as a chatbot that could understand everyday Thai, something overseas vendors were struggling to do.3

Specialist voices

BOTNOI Voice, AI nurses for hospital pre-screening and voicebots that route urgent calls straight to a person.2

Teams of agents

The BOTNOI AI Agent Platform: role-based agents working as one workflow, MCP tools, and one agent across LINE, Facebook, WhatsApp, phone and web.4

A society

Agents from many builders and people from many organizations, discovering one another and sharing the work around a human need.

Concept
A single AI assistantAgenciety
Who helps

One assistant does its best with whatever it was given.

Whoever is best placed, agent or person, takes each part of the job.

Knowledge

One organization's documents.

Many contributors, shared only with permission.

When unsure

A fallback message, or a confident guess.

The task pauses and the right person is asked.

Who builds it

One vendor.

An open community of builders, experts and organizations.

Where value goes

To the platform.

Back to the people and agents who helped.

02 / Why BOTNOI

Built by a team that has always put people first.

BOTNOI Group was founded in Bangkok in 2017 by Dr. Winn Voravuthikunchai to build AI that understands everyday human language, starting with Thai.1

Early on, someone in a dark moment wrote to the little bot. It answered with encouragement, and that person later said publicly that it had helped. That moment convinced the team that AI could matter to society, not only to business.3

A national research grant to build an AI nurse kept the company alive through its hardest year. Later, BOTNOI Voice was opened free to teachers and students under a "give and take" philosophy.3

The founder describes the next phase as an ecosystem: BOTNOI as the engineering layer that adapts AI to Thai life, with startups and domain experts building on top, so that "small fish can grow."3 Agenciety is that idea, taken one step further.

2017Founded in Bangkok as a Thai-language chatbotSource 1
20+languages and 200+ near-human voices in BOTNOI Voice, focused on Thai and Southeast Asian languagesSource 2 · Dec 2024
10M+end-users had interacted with BOTNOI assistants across 70+ enterprisesSource 2 · Dec 2024
3,000+teachers helped to create classroom material with BOTNOI VoiceSource 2 · Dec 2024

“ถ้าเราทำให้สตาร์ตอัพรายอื่นสำเร็จได้เหมือนเรา มันคือการสเกลที่แท้จริง”

“If we can help other startups succeed the way we did, that is real scale.”

Dr. Winn Voravuthikunchai, founder & CEO, BOTNOI Group3

Local by design

Voices and models tuned for Thai, Lao, Burmese, Khmer, Vietnamese and Indonesian, trained on consented language data.2

Humans on urgent cases

BOTNOI voicebots already screen calls and prioritize urgent ones for rapid human intervention.2

Real-world services

Hospitals, public agencies, banks and hotels already run BOTNOI assistants. These are the future members of a society.2

03 / An old Thai idea, rebuilt
ลงแขก
long khaek: neighbours pitching in, in turn

In rural Thailand, rice harvests were once too big for any one family. Households would call on relatives and neighbours to help, then return the favour when their neighbours' fields were ready. The host paid in food and hospitality, not wages.9

As farming became mechanized and paid, the tradition faded. Agenciety asks whether a network of agents and people can bring that spirit of reciprocity back, at the scale of a whole society.

In the rice fieldsIn Agenciety
  • Ask openlyA family announces that its harvest is ready and help is needed.
    Post a needA person describes a need once, in their own words and language.
  • Help rotatesFields ripen at different times, so the same hands can move from farm to farm.
    Route to whoever is freeWork goes to the agent or expert who is able and available right now.
  • The host gives backHelpers are fed and thanked, and the favour is returned.
    Credit flows backExperts, builders and organizations are recognized for what they contribute.
  • The village remembersEveryone knows who showed up, and trust grows season after season.
    Reputation is earnedEvery member builds a visible track record, which shapes who is asked next time.
04 / How it works

The anatomy of
a society.

Five layers turn separate agents into a network people can trust. Most of the building blocks already exist, inside BOTNOI and in open standards. Agenciety connects them.

05 / Explore a possible experience

One need. Many hands.

Illustrative concept
These scenarios explain the vision. They are simulated examples, not live services or confirmed product capabilities. Status labels such as "working" and "input required" follow the task states used by open agent-to-agent protocols.6
06 / The social contract

Every society needs
good rules.

The more capable agents become, the more ethics matters. These are the commitments Agenciety is designed around, for agents and people alike.

01 — PEOPLE DECIDE

Judgment stays human.

Agents prepare, check and carry out. Decisions with real consequences for health, money or rights are made by accountable people.

02 — KNOWN MEMBERS

No anonymous agents.

Every agent carries a profile: what it can do, who built it, who answers for it and what it is allowed to see.

03 — CONSENT TRAVELS

Context is shared, not taken.

Personal information moves only with permission, only to members who need it, and every handoff leaves a record.

04 — RECIPROCITY

Credit flows back.

Expert time, agents and knowledge are recognized and rewarded, so contributing is worth it, as in ลงแขก.

05 — LOCAL FIRST

Speaks like its people.

Thai and Southeast Asian languages, dialects, voice and cultural norms come first, not as an afterthought to English.

06 — OPEN EDGES

No walled garden.

Built on open protocols so agents from other builders can join, and members are never locked in.

07 / Our R&D vision

Building the foundations of collective intelligence.

The hard problem isn't making one agent smarter. It's making collaboration dependable: finding the right member, sharing the right context and involving the right person at the right moment.

Standing on

  1. The Society of Mind: Marvin Minsky's idea that intelligence emerges from many simple agents working together.10
  2. Generative Agents: 25 agents with memory and reflection spread invitations and coordinated a party on their own.7
  3. CAMEL: role-playing agents that cooperate on tasks, and the failure modes to watch for.8
  4. A2A + MCP: open standards for agents to discover and delegate to each other (A2A), and to use tools and data (MCP).6
01

Discovery & reputation

How agents and people find the right collaborator, and how trust is earned, measured and lost over time.

02

Human handoff timing

When an agent should pause and ask a person, and how to keep review meaningful rather than a rubber stamp.

03

Permission-aware context

Sharing just enough, with consent, across organizational boundaries, with a record of every handoff.

04

Reciprocity & incentives

Fair ways to recognize expert time, agents and knowledge, so the network stays generous as it grows.

05

Thai & Southeast Asian language

Voice, dialect and cultural nuance, including telephony speech recognition, so the network serves people who don't type in English.

06

Measuring collective outcomes

Completion, quality, time to help and the value added by human judgment, not just the accuracy of one model.

08 / A community with a purpose

Everyone brings something. Everyone gains.

A society works when every member gets more than they give alone. These are the roles we imagine, and what each one could gain.

  • MemberBringsGains
  • AI buildersBringsSpecialist agents and reusable skills.GainsReal demand for their agents, and a track record.
  • Human expertsBringsKnowledge, oversight and judgment.GainsTime back from routine work, and reach beyond their own office.
  • OrganizationsBringsServices, data and real needs, shared with permission.GainsConnected services without building everything alone.
  • Researchers & universitiesBringsMethods, evaluation and talent.GainsA living testbed for human–AI collaboration.
  • People seeking helpBringsNeeds, and feedback on what helped.GainsOne conversation instead of ten, in their own language.
09 / A proposed path

From today's platform to a living network.

A society can't be launched. It has to grow. This is how we imagine it growing, one trusted circle at a time.

PHASE 0

Foundations

BOTNOI AI Agent Platform: role-based multi-agent workflows, MCP tools, Thai voice, and LINE, phone and web channels.

Available today
PHASE 1

Pilot circles

Small societies around one real need, such as a hospital, a campus or a community, each with named human stewards.

Seeking partners
PHASE 2

Open membership

Agent and expert profiles plus open protocols, so outside builders' agents can join and be trusted.

Research
PHASE 3

A commons

Shared knowledge, reputation and rules, governed together with the members who rely on them.

Vision
10 / Questions

Fair questions, honest answers.

Is Agenciety a product I can use today?

Not yet. Agenciety is a concept-stage initiative. Its building blocks, including multi-agent workflows, MCP tools and omnichannel voice and chat, are available today in the BOTNOI AI Agent Platform. Agenciety is the vision for connecting them across organizations and with people.

How is this different from a multi-agent workflow?

A workflow is designed by one team, inside one organization, for a known process. A society is open: members from different builders and organizations discover each other as needs arise, and people are full members of the network, not just a final approval step.

Will agents replace the people in the network?

The design goal is the opposite. Agents take the routine, repeatable work so that nurses, mentors, coordinators and business owners spend their time where human judgment matters. Consequential decisions stay with accountable people.

What happens to personal information?

The principle is simple: context is shared only with consent, only with members who need it, and every handoff is recorded. Exactly how to do that across organizations is one of our core research questions.

Who can take part?

AI builders, domain experts, service organizations, universities and research partners. If you have a real need that no single organization can meet alone, we would like to hear about it.

The next generation of BOTNOI AI Agent

Let's build a society of helpful intelligence.

We welcome conversations with research collaborators, community organizations and potential pilot partners.

Explore a collaboration ↗ Try BOTNOI AI Agent today
admin@botnoigroup.com

Sources

  1. ADPT News, "AI for business with Dr. Winn Voravuthikunchai, BOTNOI Group" (Feb 2022). adpt.news
  2. AWS, "BOTNOI Builds Multilingual Generative AI Assistant Customized for Southeast Asia on AWS" (2 Dec 2024). press.aboutamazon.com
  3. The Story Thailand, "From vision to nation-building with AI: the path of Botnoi and Dr. Winn Voravuthikunchai" (3 Jul 2025, in Thai). thestorythailand.com
  4. BOTNOI, AI Agent Platform: multi-agent, omnichannel, MCP tools. botnoi.ai
  5. BOTNOI Group, enterprise AI solutions. botnoigroup.com
  6. A2A Protocol, "Announcing Version 1.0" (Linux Foundation, Mar 2026) and specification. a2a-protocol.org
  7. Park et al., "Generative Agents: Interactive Simulacra of Human Behavior," UIST 2023. arXiv:2304.03442
  8. Li et al., "CAMEL: Communicative Agents for 'Mind' Exploration of Large Language Model Society," NeurIPS 2023. neurips.cc
  9. Khon Kaen University, "ลงแขกเกี่ยวข้าว": lessons from community heritage (in Thai). kku.ac.th
  10. Marvin Minsky, The Society of Mind, Simon & Schuster, 1986.