On October 29, 2026, at the Tencent Global Digital Ecosystem Conference, Tencent did something quietly ambitious: it took five office apps that used to live in separate lanes -- WorkBuddy, Tencent Docs, Tencent Meeting, Tencent Netdisk, and ima -- and tied them together with AI into a single strand. Let's be clear about scope first. This article is the "frontier hotspot" piece in this batch that focuses on personal cross-app collaboration: what it means for an ordinary user -- one entry point, content aggregated across apps, fewer window switches. The enterprise edition's industry rollout and Tencent Netdisk's standalone role as a storage base are not unpacked here (separate pieces cover those), and this piece is deliberately distinct from other pieces in the same batch that only break down pricing or the enterprise side. The backdrop numbers from Dowson Tong, Tencent's CSIG CEO, show why this matters: WorkBuddy has handled 9 billion task requests in total, the enterprise edition has landed in more than 50 industries, covering 50% of China's top 500 companies and 60% of its top 100. The personal edition is that enterprise-proven capability flowing down to everyone's desktop for the first time.
One question, and the material comes to you
The most annoying workflow used to look like this: you wrote a proposal in Tencent Docs last week, ran a review meeting in Tencent Meeting the day before, and saved some references in ima this morning. Now WorkBuddy is supposed to help you write the weekly report, so you dig out all three, copy them, paste them into the chat, and tell the model to "use these." Buddy One is built to cut exactly that step.
The mechanism is called "full-app search." You ask once in WorkBuddy, and it retrieves your authorized documents, meeting records, and knowledge bases, integrates the hits, and produces an answer with sources attached. The crucial part is source labeling: the answer shows where each claim comes from -- ima, an authorized Tencent Doc, or a meeting transcript -- instead of a black-box mash. And after you link the accounts once, there is no repeated re-authorization and no manual context copying.
Under the hood, all five apps share WorkBuddy's Agent foundation. Tencent Docs, ima, and Tencent Netdisk all plug into the same task-execution harness, so you get the same Agent experience across them, calling the same WorkBuddy execution layer underneath. In other words, Tencent did not build yet another app to "summarize everything"; it lets the material already scattered across places be pulled together by AI only when you need it. For an ordinary user the perceived change is simple: the place you ask questions is still WorkBuddy, but the range it can see now reaches beyond WorkBuddy's own library.
The payoff shows up most on meeting-heavy days. After three or four meetings, the conclusions sit scattered across several minutes and docs; in the past you opened each one to pull a summary together. Now a single question gathers them and tags the source, so you only verify instead of hunting. That shift is small per question and large per week, because the "switch, open, compare" loop is exactly the tax knowledge workers pay all day. Buddy One's bet is that removing that tax, one authorized query at a time, is worth more than any single flashy feature.
One entry point: account, credits, memory
The second way they become "one rope" is unification. Previously, when WorkBuddy called other products through connectors, authorizations kept expiring and had to be redone. Now you bind a Buddy One ID for free, the accounts connect, and one link works long term. This sounds small but changes the feel a lot: instead of re-consenting every time you touch a new product, you consent once and it sticks.
Credits merge too. Once you join Buddy One, your old WorkBuddy credits automatically upgrade to "Buddy One credits," usable across WorkBuddy, Tencent Docs, ima, and Tencent Netdisk -- one shared balance that deducts in sync wherever you spend. Spend some in Docs and the pool in WorkBuddy shrinks by the same amount. One detail worth remembering: Tencent Meeting does not consume credits, so meetings themselves don't burn your balance. The point of this "top up once, use everywhere" credit pool is plain -- stop forcing you to maintain a separate balance inside every single app and stare at four or five bills at month end.
Memory sharing is the farther horizon. Officially the wording is "gradually explore cross-app memory sharing." What works today is account interconnection, authorized retrieval, credit sharing, and unified storage; later versions will let memory flow between apps. On the PC side, update WorkBuddy to 5.9.0 or above and link your accounts -- you don't have to hand over all your memory now, because cross-app memory is a next step on the roadmap, not a switch that is already flipped on by default. This rhythm of "connect the material first, then slowly connect the memory" is actually friendlier to privacy-conscious users.
Alongside Buddy One, Tencent also shipped Tencent Netdisk, positioned as a unified storage base for the agent era. Its role is to give this office system a common storage home and to support connecting third-party agents so they can call your deposited materials within permission. This piece does not deep-dive Netdisk as a standalone product, but its place in "one rope" is clear: materials finally get one home, and only then can AI pull from them within authorization at any moment, instead of starting from zero every single time.
What it actually does for you: three scenes
Scene one, writing. You ask, "Summarize the high-frequency client requests mentioned in our communications this quarter," and WorkBuddy scans your authorized meeting records and documents, gathers the hits, and labels each one with its source meeting or doc -- no more jumping between five apps. In the past this kind of task meant first locating from memory "where did that meeting say it," and now that step is gone.
Scene two, finding. You can't remember where a conclusion was written, so you ask, "What did last month's review meeting decide?" It pulls it from the transcript and tells you which Tencent Meeting it came from, sparing you the chore of scrolling through chat logs and meeting minutes. Source labeling here is not merely "you can see where it came from" -- it is "you can go back and verify." One click and you know the exact original, instead of having to trust the model's paraphrase.
Scene three, cross-source synthesis. Notes in ima, drafts in Tencent Docs, and files resting in Netdisk can all be pulled together in one question to produce a sourced summary, sparing you the manual labor of moving context around. The common thread across these scenes: the material stays in its original app; the AI only aggregates it temporarily within authorization. You never have to move everything into one new place, and you never have to bounce between a pile of windows. For knowledge workers who run three or four meetings a day and leave documents scattered everywhere, what gets saved is not one or two clicks, but the whole invisible tax of "switch, copy, paste" repeated all day long.
A concrete daily example makes the value concrete. Suppose you are preparing a client pitch and need last month's meeting decisions, a draft document, and three reference notes. Instead of opening four apps and stitching them by hand, you ask once and receive a single sourced brief. The time saved is not the answer itself but the searching, opening, and copying that used to precede it.
Privacy and permission boundaries: it reads, but only where you allowed
Cross-app reading is exactly what makes people uneasy. Buddy One's boundary is explicit: it can only call content you have authorized, and it stays bound by each product's existing permission rules. There are already data walls between Tencent's own products -- WeCom, for instance, has its own chat-record agent to protect privacy, and WorkBuddy does not by default read other people's chats. On the enterprise side, permissions can be set per employee identity, and external companies broadly guard their core databases, so WorkBuddy only calls within authorized scope.
For an individual user this means two things. First, an app you haven't linked or authorized is invisible to it. Second, even after linking, it only operates inside the permissions you granted -- it won't slip off to rummage through things you never opened. In other words, "one rope" presumes "you nodded"; however tight the rope, it is tied within the range you authorized. For anyone who cares about privacy, this boundary deserves more attention than "how much can it aggregate" -- because the stronger the aggregation, the more it matters exactly where the line is drawn. Tencent's repeated emphasis on "within authorized scope" echoes the line every peer in this space holds: AI can help you string your material together, but it cannot open a door you never unlocked.
The peers: everyone is tying apps into a rope
Tencent is not alone in this idea, but the approaches differ. Feishu's route is to natively embed agents inside its own closed suite, where its smart companion and multi-dimensional table AI collaborate in a closed loop across Feishu Docs, Base, meetings, and cloud drive. DingTalk uses its AI assistant as the lever, stressing one entry point plus agents plus a seamless collaboration network, completing cross-app coordination inside its own closed loop. Microsoft's 365 Copilot uses Microsoft Graph to connect Word, Excel, PowerPoint, and Teams into one context, offers cross-app universal search, and like the others only touches content you already have permission for, with clickable citations attached.
All four stress "only authorized content," and Tencent additionally highlights the data walls between its own products and enterprise-grade authorization isolation. On "answers with sources," both Tencent and Microsoft make explicit source labeling, while Feishu and DingTalk lean more on in-ecosystem loops and proactive collaboration. This is a collective race to tie multiple apps into one AI rope; the differences are mostly in the tying method, the boundaries, and the billing -- not in whether to tie at all. For ordinary users, the real comparison is not whose slogan is louder, but which rope can actually knot the few apps you live in every day.
Is it worth it: what 700 yuan a month buys
Buddy One also puts a clear price tag on the table. The VIP continuous monthly plan is 700 yuan per month, covering WorkBuddy Flagship, Tencent Meeting Pro, Tencent Docs Super Member Plus, cross-app shared AI credits, and 1TB of cross-app shared storage. Binding the Buddy One ID itself is free -- you don't have to buy a membership first. That means an ordinary user can link accounts at zero cost, enjoy the unified entry point and authorized retrieval, and only then decide whether to pay for higher credit and storage ceilings and those membership perks.
Side by side, those 700 yuan buy "five apps' entitlements bundled plus one cross-app AI workbench." Unlike Microsoft's per-seat 365 Copilot subscription or Feishu and DingTalk's team- and enterprise-seat billing, Tencent's VIP is a personal cross-app plan. Whether it pays off depends on how long you already live inside these Tencent office products and how heavily you call AI. If you shuttle daily between Docs, Meeting, Netdisk, and ima, the switching cost it saves may outweigh the fee; if you only use one of them, the rope is a bit loose for you. For most people, linking for free first and upgrading on demand is a steadier way in than paying 700 yuan up front: let the AI save you a few window switches, confirm the rope holds, and only then talk about paying to thicken it.
Related reading
- Google's New Hire: An AI Coworker With Its Own Inbox
- AI Agent Office Automation SOP: Stop Learning Tools-You Need a Digital Worker
- WorkBuddy Field SOP: Turn Off Auto-Model and Take Tencent's AI Agent from Clueless to Competent
- Tencent's Hy3 Is Free Worldwide via WorkBuddy Until Aug 31: A Window to a Frontier Open-Source Flagship
- AI Agent Memory Tools Compared: Mem0, Zep, Letta, and How to Pick