Tencent WorkBuddy is an AI agent built for everyday office work. Make a request. Guide your AI expert team. Bring in a second opinion. Get sharpened, ready-to-use results.
Replies
Best
Poked around with it for a bit and the "second opinion" angle genuinely caught me off guard in a good way, it actually feels like having a teammate double-check your work instead of just another chatbot agreeing with everything you say. Results came back clean enough to drop straight into a doc.
It's great to hear the "second opinion" experience stood out. We built our experts to contribute different perspectives and help users arrive at more professional, well-rounded results—not just generate another similar response.
Really appreciate you taking the time to share your experience!
Report
Tencent's name on this raises an immediate data question for anyone considering it for actual office work. Where is the processing happening, what data residency options exist, and is there an enterprise tier with clear answers to those questions before someone puts sensitive work documents through it?
@ansari_adin Fair question. Tasks run in an isolated sandbox and you control what it can touch, so you keep full ownership of your data. On the enterprise side it runs on Tencent Cloud's infrastructure, which covers permission management, runtime auditing, and compliance support. For data residency specifics, happy to connect you with the team directly since it does vary by region and use case.
Report
@sherina_chen "Tencent Cloud's infrastructure with compliance support" covers a wide range depending on which certifications are actually in scope. SOC 2, ISO 27001, and GDPR are very different compliance postures and not all of them are covered by Tencent Cloud's existing certs depending on the region. Is there a trust page or compliance overview that lists which certifications WorkBuddy itself holds, separate from what Tencent Cloud offers underneath?
Report
The part I'd poke at: asking in the brief is prompt-level, and in my runs a plain-language 'keep the dissent' held up fine until the context got long, then the minority take quietly vanished maybe 1 in 5. Is preservation actually enforced in the synthesis step, or is it the model choosing to honor the instruction each time? That gap is what decides whether I'd trust it on anything with a real decision attached.
@dipankar_sarkar Fair poke. Honestly it's the model honoring the intent, not a hard rule, so what you saw (minority take thinning out on long context) is a failure mode we've hit too. What helps is the debate stays visible in the thinking process and each take is kept separate instead of pre-blended, so it survives the merge better. But I won't oversell it as guaranteed. For a real decision, a tighter brief still holds dissent better than one long run.
Congrats on the launch! 🎉 I really like the "team of AI experts" concept. How do you decide which specialists get assigned to a task? Is it automatic or can users customize the team?
@luki_notlowkey Right now it's user-driven. In the same chat you switch between experts on the fly, and if you summon an Expert Team the specialists collaborate directly so you don't hand-assign each piece. You can also build your own if none fit. Smart recommendations are on the roadmap, but for now you're the one calling the shots.
Report
Congrats on the launch. The part I'm most curious about is data boundaries rather than the synthesis logic - if I spin up a team with a research expert and a legal expert on the same brief, do they see each other's full working context and any files I uploaded, or is each expert scoped to only what it needs for its part? Multi-agent setups are great until one expert quietly has more access than the task calling it actually required.
@galdayan Access is something you grant, not something that's open by default. Tasks run sandboxed, and reaching your files needs your explicit permission, so nothing gets touched unless you've said so.
Report
Used it for a quick report draft and the multi-agent angle actually helped, gave me three different angles in like a minute instead of one generic answer. Slightly clunky UI but the results felt noticeably more useful than my usual single-prompt tools.
@kumsalk9229 Yeah, one prompt shouldn't lock you into one answer, so three angles in a minute is kind of the whole point. UI's still getting polished, but glad the results are landing more useful for you.
Report
How does the pricing work for the AI expert team — is it a flat subscription or do you pay per agent or per task?
@nazargkbaleawm It's a flat subscription, not per agent or per task. Running the full expert team on a task doesn't cost extra on top, you're just drawing from your Credits.
Report
Have been kicking the tires on this one today and the multi-agent setup is honestly pretty clever. Got a messy brief turned into a clean summary and follow-up email in one go. Curious how it handles more technical docs.
@yiitwcuf That summary-plus-email-in-one-go flow is the sweet spot. Technical docs hold up better than you'd think since you can pull in an expert tuned for it, though I'd still skim the output on really dense specs. Let me know what you throw at it.
Report
@sherina_chen Congrats to the entire team on this success. Looking forward to see how impactful the product will be as time goes on. how does your product deal with agents working cross function? in a situation where 2 different agents could possibly do different task but have 2 identical task they can both do, which of the agents will take up the task?
@adebowale_adeseye Thanks! Good question, I get what you mean. It doesn't work like a dispatcher auto assigning the task to one agent behind your back. You're the one who picks which expert to bring in, and when you summon the expert team for something, they actually work together rather than compete for it, each one contributing where it's strongest and handing off to the next. So it's less "which agent grabs the task" and more "the right ones collaborate on it." You stay in the driver's seat the whole way.
I’d like to have Workboddy take over my browser to perform repetitive tasks, but I’m very concerned about the potential leakage of my account details, payment information, and personal data. How do you handle the uploading and protection of personal information?
@bob_bo Hi, thanks for you questions. WorkBuddy is not meant to freely crawl everything in your browser or silently collect account/payment details. It works within the scope the user authorizes: the task workspace, connected tools, installed skills, or the specific browser/task context needed for the job.
By default, higher-risk actions are guarded by permission controls. For example, actions involving sensitive paths, deletion, scripts/external programs, or sensitive network capabilities can require user confirmation rather than running silently.
For personal data, our principle is to process only what is needed to complete the task. Users should avoid putting unnecessary passwords, card numbers, or private personal data into prompts. Memory is also user-manageable — you can view/edit entries and tell WorkBuddy what to remember or forget.
The goal is to make repetitive browser work faster while keeping access explicit, scoped, and reviewable.
@bob_bo Thanks for asking,it's definitely something we'd ask too~
WorkBuddy only works with the data you allow it to access, and tasks run in an isolated sandbox so you stay in control throughout the process.
For enterprise users, it runs on Tencent Cloud's infrastructure with support for permission management, runtime auditing, and compliance. If you're evaluating it for a specific use case, feel free to reach out—we'd be happy to talk through the details.
Replies
Poked around with it for a bit and the "second opinion" angle genuinely caught me off guard in a good way, it actually feels like having a teammate double-check your work instead of just another chatbot agreeing with everything you say. Results came back clean enough to drop straight into a doc.
Miora
@berkay139616 Thanks so much for giving it a try~
It's great to hear the "second opinion" experience stood out. We built our experts to contribute different perspectives and help users arrive at more professional, well-rounded results—not just generate another similar response.
Really appreciate you taking the time to share your experience!
Tencent's name on this raises an immediate data question for anyone considering it for actual office work. Where is the processing happening, what data residency options exist, and is there an enterprise tier with clear answers to those questions before someone puts sensitive work documents through it?
Miora
@ansari_adin Fair question. Tasks run in an isolated sandbox and you control what it can touch, so you keep full ownership of your data. On the enterprise side it runs on Tencent Cloud's infrastructure, which covers permission management, runtime auditing, and compliance support. For data residency specifics, happy to connect you with the team directly since it does vary by region and use case.
@sherina_chen "Tencent Cloud's infrastructure with compliance support" covers a wide range depending on which certifications are actually in scope. SOC 2, ISO 27001, and GDPR are very different compliance postures and not all of them are covered by Tencent Cloud's existing certs depending on the region. Is there a trust page or compliance overview that lists which certifications WorkBuddy itself holds, separate from what Tencent Cloud offers underneath?
The part I'd poke at: asking in the brief is prompt-level, and in my runs a plain-language 'keep the dissent' held up fine until the context got long, then the minority take quietly vanished maybe 1 in 5. Is preservation actually enforced in the synthesis step, or is it the model choosing to honor the instruction each time? That gap is what decides whether I'd trust it on anything with a real decision attached.
Miora
@dipankar_sarkar Fair poke. Honestly it's the model honoring the intent, not a hard rule, so what you saw (minority take thinning out on long context) is a failure mode we've hit too. What helps is the debate stays visible in the thinking process and each take is kept separate instead of pre-blended, so it survives the merge better. But I won't oversell it as guaranteed. For a real decision, a tighter brief still holds dissent better than one long run.
AgentKey
Miora
@luki_notlowkey Right now it's user-driven. In the same chat you switch between experts on the fly, and if you summon an Expert Team the specialists collaborate directly so you don't hand-assign each piece. You can also build your own if none fit. Smart recommendations are on the roadmap, but for now you're the one calling the shots.
Congrats on the launch. The part I'm most curious about is data boundaries rather than the synthesis logic - if I spin up a team with a research expert and a legal expert on the same brief, do they see each other's full working context and any files I uploaded, or is each expert scoped to only what it needs for its part? Multi-agent setups are great until one expert quietly has more access than the task calling it actually required.
Miora
@galdayan Access is something you grant, not something that's open by default. Tasks run sandboxed, and reaching your files needs your explicit permission, so nothing gets touched unless you've said so.
Used it for a quick report draft and the multi-agent angle actually helped, gave me three different angles in like a minute instead of one generic answer. Slightly clunky UI but the results felt noticeably more useful than my usual single-prompt tools.
Miora
@kumsalk9229 Yeah, one prompt shouldn't lock you into one answer, so three angles in a minute is kind of the whole point. UI's still getting polished, but glad the results are landing more useful for you.
How does the pricing work for the AI expert team — is it a flat subscription or do you pay per agent or per task?
Miora
@nazargkbaleawm It's a flat subscription, not per agent or per task. Running the full expert team on a task doesn't cost extra on top, you're just drawing from your Credits.
Have been kicking the tires on this one today and the multi-agent setup is honestly pretty clever. Got a messy brief turned into a clean summary and follow-up email in one go. Curious how it handles more technical docs.
Miora
@yiitwcuf That summary-plus-email-in-one-go flow is the sweet spot. Technical docs hold up better than you'd think since you can pull in an expert tuned for it, though I'd still skim the output on really dense specs. Let me know what you throw at it.
@sherina_chen Congrats to the entire team on this success. Looking forward to see how impactful the product will be as time goes on. how does your product deal with agents working cross function? in a situation where 2 different agents could possibly do different task but have 2 identical task they can both do, which of the agents will take up the task?
I hope you do understand my question?
Miora
@adebowale_adeseye Thanks! Good question, I get what you mean. It doesn't work like a dispatcher auto assigning the task to one agent behind your back. You're the one who picks which expert to bring in, and when you summon the expert team for something, they actually work together rather than compete for it, each one contributing where it's strongest and handing off to the next. So it's less "which agent grabs the task" and more "the right ones collaborate on it." You stay in the driver's seat the whole way.
@sherina_chen cool!
I’d like to have Workboddy take over my browser to perform repetitive tasks, but I’m very concerned about the potential leakage of my account details, payment information, and personal data. How do you handle the uploading and protection of personal information?
WorkBuddy
@bob_bo Hi, thanks for you questions. WorkBuddy is not meant to freely crawl everything in your browser or silently collect account/payment details. It works within the scope the user authorizes: the task workspace, connected tools, installed skills, or the specific browser/task context needed for the job.
By default, higher-risk actions are guarded by permission controls. For example, actions involving sensitive paths, deletion, scripts/external programs, or sensitive network capabilities can require user confirmation rather than running silently.
For personal data, our principle is to process only what is needed to complete the task. Users should avoid putting unnecessary passwords, card numbers, or private personal data into prompts. Memory is also user-manageable — you can view/edit entries and tell WorkBuddy what to remember or forget.
The goal is to make repetitive browser work faster while keeping access explicit, scoped, and reviewable.
Miora
@bob_bo Thanks for asking,it's definitely something we'd ask too~
WorkBuddy only works with the data you allow it to access, and tasks run in an isolated sandbox so you stay in control throughout the process.
For enterprise users, it runs on Tencent Cloud's infrastructure with support for permission management, runtime auditing, and compliance. If you're evaluating it for a specific use case, feel free to reach out—we'd be happy to talk through the details.