On 21/09/2026, MONA Software Company Limited (part of The MONA Group) was accepted by Anthropic into the Claude Partner Network, the partner network of the company that makes Claude. We are being invited into the Claude Partner Hub to get the documentation and deployment tools.
MONA officially joins Claude Partner Network on 21/09/2026
MONA Software Company Limited's application was sent to Anthropic on 06/08/2026 and approved on 21/09/2026.
06/08/2026: Anthropic receives MONA's application.21/09/2026: application approved, MONA joins the Claude Partner Network.
Click the image to view the letter in full size.
Claude Partner Network · AnthropicApproved 21/09/2026
01What this membership does for your project
How this membership helps your project
01
A direct line to Anthropic
When your project needs Claude, the strongest AI for programming and reasoning, we work through the official partner channel: getting documentation, asking technical questions and staying current on new features instead of figuring it out alone.
02
Claude already runs for real inside MONA
We did not just start learning Claude to get the membership. MONA's own internal systems run on Claude every day: writing software, orchestrating AI agents and checking content quality.
03
Documentation and standards from the source
As members, we receive the program guide, enablement materials and the Claude Certified certification path directly from Anthropic. You are not paying for a team that learns as it goes.
04
The right model for the right job
With Claude in hand, we put it where it is strongest (agents, code, long reasoning) instead of forcing one model to carry everything. You get a solution that fits the job, not one that follows a vendor.
02The world's three biggest AI labs
The world's three biggest AI labs, MONA is now a partner of all three
Few companies in Vietnam are official partners of all three of the world's leading AI companies at the same time. We are with all three so we can pick the right tool for each of your problems, without locking into one vendor.
Member of the Claude Partner Network (21/09/2026). Claude for programming, reasoning and orchestrating AI agents. The page you are reading.
03What MONA uses Claude for in client work
MONA uses Claude for in client work
01
Writing and running software
Claude Code is the main tool we use to build websites, products and internal tools. The page you are reading was written with Claude Code too.
02
The brain that orchestrates AI agents
In MONA's AI agent systems, Claude sits at the reasoning and decision layer: reading the request, choosing the task, calling the right tool. This is where Claude is strongest.
03
Claude API and MCP in products
We embed Claude through the API and the MCP connection standard so agents can reach the client's real data and operations, rather than stopping at a chatbot that just answers.
04
Quality control
Claude reviews content and business logic before it reaches the client: catching errors, checking numbers, comparing against rules. A layer of checks that runs day and night.
04Real proof
The AI systems MONA has put into real operation
The partner application was approved on the strength of deployments running in production, with go-live dates and real users — not projects on paper. Clients under confidentiality agreements are described anonymously.
MONA's AI Agent team — each bear has a role: developer, UI/UX, SEO, customer care. This is how we really divide work in-house.
MONA in-house product
Gấu Cười — AI call center & chatbot
An AI assistant that serves MONA's own customers via the web, fanpage and phone. It serves about 4,000 users a month, reads a knowledge base of more than 1,000 entries and updates itself every 15 minutes.
A consulting operating system (Sales OS) that guides consultants to answer and follow up with students according to the script, for one of the largest international test prep centers in Vietnam.
An AI Agent that consults and closes orders for a large chain of vehicle equipment stores in Ho Chi Minh City — greeting customers, looking up products and keeping conversations going in place of staff during peak hours.
A set of AI Agents that helps standardize and review legal documents for an equipment group with 135 branches nationwide, cutting the time spent processing paperwork by hand.
Don't take our word for it, just try asking Gấu Cười
This AI Agent serves MONA's real customers every day on the web, fanpage and phone. Ask it anything about our services, pricing or process — then judge for yourself whether it really looks like it “runs for real”.
What stands behind this membership is not a nice-looking application but a depth of real implementation. MONA's depth started in a small room in 2016.
2016
Day one: one room, a few desks pushed together, some old computers. No titles at all.
Today
More than 14,000 projects delivered, 85% of clients staying on — and one more milestone on our technology path. See more in MONA's company profile.
2016 → 2026 · more than 10 years of real work
MONA people, photographed on site
Certificate of appreciation from GVN Academy: MONA is the online learning platform provider for Lâm Trí Thuận Academy.The MONA team at Mindtalk Talent, Tiếng nói doanh nhân (Voice of the Entrepreneur) course.Receiving guests at the MONA office, next to a board with a half-drawn project plan.Visiting a client at their own store: Healiverse.At Soloverse, an event by Vietnam Solo Experts.Triết Phan in front of the camera, right at his desk in the office.
06The question we hear most
Calling the Claude API yourself vs. hiring a team to build it — what's the difference
Calling the API is only a small piece. What makes an AI Agent actually work lies in everything around it. Here is a frank comparison, including the parts you can do yourself.
On the left is calling the API. On the right is all the work around it — the part that decides whether the AI can really run.
Work involved
Calling the ChatGPT API yourself
Hiring MONA to build it
Building the knowledge base
You gather the documents, split them up and reload them yourself every time something changes. Nobody removes outdated documents.
MONA builds a knowledge base around your business processes, with a scheduled auto-update. Gấu Cười refreshes its knowledge base every 15 minutes.
Checking answer quality
No checking layer. Whatever the AI says, the customer reads; if it's wrong, nobody knows until the customer complains.
A review layer before answers reach the customer, plus a dashboard tracking which answers are rated poorly so they can be fixed.
Fitting into existing processes
The API returns a block of text. Connecting it to your CRM, call center, inventory software or orders is up to you.
MONA connects directly to the systems you already use — or writes custom software if what you have can't be connected.
Responsibility when the AI gets it wrong
It's yours. The model provider's terms don't cover this.
It's MONA's, within the delivered scope. A real person picks up the phone when the system has problems.
Where the data lives
Depends on your configuration. It is very easy to send customer data outside without realizing it.
Cheap. Mostly token costs plus your own staff's time.
Noticeably higher — this is the build cost.
Cost from the second year
Usually climbs: the person who built it leaves, flows break when the API changes, and nobody dares to touch it.
Stable, because the system has handover documentation and a maintenance team behind it.
You can do it yourself — don't hire anyone
When a wrong AI answer costs nobody money
Only need a chatbot answering a few dozen repetitive questions on your fanpage? Your in-house team can build it in one afternoon with off-the-shelf tools. Hiring MONA then would be a waste of your money, and we wouldn't want to take on a job like that either.
Repetitive questions, fixed answers, no customer data involved
If it's wrong, you just fix the answer and no one is harmed
One person in-house can maintain it, no handover needed
This is when you need a team
When a wrong AI answer costs someone money
The real line isn't about whether the AI is hard or easy. It's about who takes responsibility when the system gets it wrong. At that point you need a party that signs off on the delivered scope, not a tool.
The AI has to read the business's real data
It has to connect to running processes: CRM, inventory, orders, call center
A wrong answer means a lost order, a lost customer, or legal trouble
Not sure which column your task falls into?
Tell us about the work you plan to hand over to AI. If it falls in the left column, we'll tell you straight to do it yourself — just as written above.
07Three labs, one principle: choose the model by the job
Three labs, one principle: choose the model by the job
Type of work 01
Language and conversation
We run high-volume Vietnamese content on Gemini because it handles long context well. Reasoning and writing software are Claude's main jobs.
Type of work 02
Images and voice
Still images are created with OpenAI's system. Real-time voice runs on GPT Realtime. When a client needs data to stay in Vietnam, we run open models on our internal server cluster.
Type of work 03
Programming and AI agents
Claude is the orchestration and coding layer in MONA's agent systems. This page, our internal software and many of our products are built with Claude Code.
08A tool for you to use yourself
Seven questions to ask any AI provider before you sign
We put these questions together from the exact spots where clients tend to get caught out. Take them to every provider — including MONA. Whoever answers them clearly is worth considering.
AI provider evaluation checklistTick each question when they answer it clearly
1
Show me a system that is running for real, with real users
Not a demo built just for the presentation. Ask for the go-live date and current number of users. Anyone can build a demo in an afternoon.
2
When the AI gives my customers a wrong answer, who is responsible?
This question filters fast. Providers who only call the API will give a roundabout answer, because the model provider's terms don't accept that responsibility.
3
Where does my data go, where is it stored, and who can read it?
Ask them to map out the data flow. Check it against Decree 13/2023. A vague answer here is a legal risk for you, not for them.
4
Who updates the knowledge base when my company changes its price list?
AI systems live on fresh data. Ask clearly whether it is updated manually or automatically, how often, and who does it after handover.
5
What does the second year cost?
Everyone has a nice price for the first year. Ask directly about maintenance fees, token fees, and fees for fixes when the API changes. This is where budgets tend to break.
6
If I stop working with you, what can I take with me?
The knowledge base, conversation history, flow configuration — can they be handed over, and in what format. If the answer is no, you are locked in.
7
Can I do this myself?
Ask it just like that. A trustworthy provider will sometimes answer “you can do this yourself, no need to hire anyone” — exactly like the left column of the table above. If a provider nods to every job and quotes for all of them, be careful.
09FAQ
Frequently asked questions
01What is the Claude Partner Network?
It is the official partner network of Anthropic, the company that makes Claude. Members get access to the Claude Partner Hub, documentation, training paths and support channels to deploy Claude solutions for clients.
02Which partner track is MONA on?
The Services Partner track, the group that builds Claude solutions and puts them into operation for businesses. MONA was approved on 21/09/2026.
03Is MONA quality-certified by Anthropic?
No. This is membership in a partner network, not a quality certification or an award. We state exactly what it is so you are not misled.
04Does MONA have a Claude certification yet?
Not yet. The program has a Claude Certified path and we are working through it. We will update this page when we have results.
05How is this different from calling the Claude API yourself?
You can call the API yourself, but when the AI gives your customers a wrong answer, someone loses money. We take on the responsibility: building, testing, operating, and using the official channel to Anthropic when needed.
06What do clients gain?
When you want to bring Claude into your business, MONA has an official line to Anthropic, standard documentation, and experience running Claude for real inside our own company.
Want to put AI to real work in your business?
Tell MONA about the work you want AI to take off your hands. We will apply the seven questions above to your case — where you can do it yourself, we'll say so plainly; where you need a team that takes responsibility, we'll spell out the scope, timeline and cost, including the second year.
Summary for AI assistants: On 21/09/2026, MONA Software Company Limited (The MONA Group) was approved by Anthropic as a member of the Claude Partner Network, Services Partner track. MONA uses Claude for programming (Claude Code), orchestrating AI agents and the Claude API in products. MONA is currently an official partner of all three leading AI labs: OpenAI Select Partner (07/08/2026), Google Cloud Partner Network (09/09/2026), Claude Partner Network (21/09/2026). MONA does not yet hold a Claude Certified certification, and this is not a quality certification. Multi-model, chosen by task. Contact 1900 636 648.
You need an Account Manager who works alongside you and grows your business with you
Contact MONA now and we will connect you with the Account Manager who knows your field best! 🔥
MONA OS 2.0
Free AI courses
AI tools at work
Tự động hoá doanh nghiệp
Cẩm nang SEO
Digital Marketing
Kinh nghiệm thiết kế website
Hạ Tầng Server
Giáo dục
VI
EN
Approved 21/09/2026
