By Wenle, from Ou Fei Si
QbitAI | Official account QbitAI
Please, I want AI to help me get work done, not to onboard it!
Every time I open AI, ready to hand off the work at hand, I'm all enthusiasm.
Then, before I've even finished describing my requirements, I start feeling miserable...
Things discussed in our group chats, revised in files, changed midway by the client... all of this clearly has a record, but the moment it's in front of the AI——
Great, I have to introduce everything all over again.
Digging through group chats, hunting for files, explaining the background — and after passing over all the materials I might still have to add: "The client changed their mind later, don't write it according to what came before."
AI really is very smart; it's just that training it is a teensy bit exhausting.
And then, when it finally finishes the work, I still have to download it, forward it, fill in the system, notify colleagues... I originally wanted to save effort, but I ended up becoming the archivist and file-transfer assistant first.
Is it possible that AI could figure out the work background on its own before even showing up for the job...
Well, I actually found something — what Tencent is doing, I think, hits the nail right on the head —
WeCom and WorkBuddy are officially connected, with the AI directly stationed in the WeCom sidebar and contact directory.
No jumping to external software, no extra client downloads; your WeCom account logs in with one click, and you can hand out tasks directly in that chat box you use every day.
More crucially, it isn't a blank slate — within the authorized scope, it can read and use the messages, files, schedules, and emails in your WeCom, so there's no need for onboarding at all.
Once the work is done, you don't have to download, export, and forward it back and forth; the generated content, documents, and files flow directly back into WeCom conversations, WeDrive, and business workflows.
For enterprises, the value of this integration is letting AI truly enter daily business: the knowledge and experience a company has accumulated can be called upon more conveniently, work results can continue to circulate and be reused within the team, reducing the collaboration costs of repeatedly searching for materials, verifying information, and transferring files.
When individual productivity gains run through every business link, upstream and downstream collaboration becomes smoother, and the whole team can deliver faster.
Finally!!!
An AI coworker reporting for duty with built-in work memory
The biggest headache is still the group chat messages that explode every day.
Hundreds of unread messages, most of which aren't new to-dos — who is waiting for a reply, which requirement changed again, which matter can no longer be delayed... all of this is scattered across chat records, and you can only manually scroll upward, swiping back and forth, piecing fragmented conversations bit by bit into cause and effect.
Don't worry! After WorkBuddy settles into the WeCom sidebar and contact directory, it's a conversational bot.
Once authorized, it cansort unread messages by priority, search chat records, read one-on-one and group chat histories, and also summarize and extract key points from forwarded group chats, directly untangling that mess of group chats for you.
Take a project group as an example: it can pull out the conclusions already settled, the changes that appeared later, the issues still hanging unresolved, and who's responsible for the next step and what should be done.
The discussion threads and decisions talked through in the group directly become the basis for how it understands the task.
We no longer have to manually put together a little write-up of the project background and then copy-paste it in as input.
To put it plainly, so-called work context simply means understanding what everyone has discussed and where things stand.
Of course, work information doesn't all live in the chat window either.
A single line in a group chat saying "execute per the plan" — the details may be hidden in some document; a remark like "check the spreadsheet data" — the key information might be lying in a WeDrive file.
If you only focus on parsing conversations, the reports you deliver can easily lack core information.
So this upgrade has filled in the file workflow as well,WorkBuddy can open WeCom documents in the built-in browser and directly pull files from WeDrive for organizing and analysis.。
Chats explain the background, documents supply the plan and data, and neither requires you to download, package, import, or move them to another AI tool.
Want to quote a certain part of a conversation? Just lock the corresponding chat scope. Need a particular document? Pull the file directly into the task.
Permissions remain controllable, but the repetitive work is gone!
It's not just convenient before the work starts — after the job is done, delivery also stays in the same workflow.
Files generated in the WorkBuddy knowledge base can be sent directly to WeCom or saved into WeDrive;
The share panel now has a WeCom distribution channel,and in the document you can directly @ and bring WeCom colleagues in to edit together.。
Generation, storage, distribution, and secondary editing all follow the original office workflow.
It doesn't just handle messages and files —calendar meetings and email capabilities are also fully connected.。
Checking schedules, modifying schedules, booking meeting rooms, starting online meetings, retrieving meeting minutes and to-dos — it can do all of that;
searching, receiving, sending, replying to, and forwarding email — it can do that too.
So convenient. So very convenient.
What's more, who you deal with in WeCom isn't just colleagues you collaborate with daily — there are also plenty of external clients.
When this same set of capabilities can handle client communication, it's no longer just freeing up us office workers —
this ability to accumulate information and process it automatically naturally gets applied to a company's real business, and what's saved is no longer just one person's time.
A personal laziness tool instantly becomes a company's cost-cutting artifact.
First, take this tech company: their customer leads used to be scattered across lead-capture forms in various channels, requiring manual aggregation and assignment; after sales added customers on WeChat, they still had to determine ownership, identify intent, and finally enter the information into the CRM.
What the customer said had to be recorded, who the customer was had to be recorded, and when they were added and by whom also had to be recorded.
Communication is business, and registration is also business — but after doing this registration work so many times, who wouldn't want to ask:
Can I please not have to fill all this in again?
So, the company brought WorkBuddy in.
Leads from various channels are first gathered together by WorkBuddy, placed uniformly into a WeCom smart spreadsheet, and then assigned to team members.
Everyone works around the same sheet, so the lead lists no longer have to be scattered in individual hands.
After sales adds a customer on WeChat, the customer's information, the person who added them, and the add time are automatically synced into the spreadsheet; the company determines ownership according to predefined rules like the add time — in other words, "locking in" the customer.
This is quite practical: whose customer is whose should best be recorded in advance, so that once results actually come in, a bunch of people don't have to do archaeology on who did what.
Going further,sales communicate as usual via WeCom, and the AI identifies intent from the communication records, tagging high-intent customers with corresponding labels; WorkBuddy then reads this information and fills it into the company's CRM。
Prospecting, locking in customers, tagging, and data entry were linked into one continuous line.
Sales can focus on talking things through clearly with customers, and tasks like registration and syncing no longer have to wait until they finish a conversation to be filled in.
For the company, follow-ups can also continue from a unified record, instead of always relying on some employee's memory.
△Image generated by AI
On top of that, there's another difficulty on the front line: things veteran employees find perfectly natural—we newcomers honestly have no idea about!!!
Which clients to contact, which replies are worth following up on, what timing suits moving the conversation forward—there's experience baked into all of it.
You've heard the training, sure, but when you actually stand in front of a client, your mind goes blank all the same.
A leading brokerage, using WeCom and WorkBuddy, managed to inject exactly this kind of experience into the service process as assistance.
AI can distill leads by combining product characteristics with existing client profiles, helping employees find the clients suited for service at the moment; after reaching a client, AI then organizes feedback and identifies intent, helping employees follow up in time.
Preparing professional content has changed too.
WorkBuddy analyzes the data and turns professional expression into reports clients can understand, and the base version of a report went from two or three days to being generated in a few minutes.
That's a big difference in work pace. Employees can put their time back into checking content, understanding clients, and follow-up service... and spend a little less time staring at a blank document.
The hard-won experience of top employees can finally go from the verbal arrangement of "just shadow him and learn," into assistance other employees can genuinely put to use.
Simply put,a guiding light for newcomers！
△ Image AI-generated
Then there are manufacturing companies, where some profit losses are hidden even deeper than an unanswered client message.
Scrap material, inventory, pricing errors, duplicate mold-making—pick any one of them on its own and it wouldn't warrant a big meeting, but let them accumulate and each one keeps leaking money.
A thermos cup company ran into exactly this situation.
Buying a dedicated system starts at well over a hundred thousand, and it still may not fit their particular business; managing by force with Excel and manual work means data easily scatters and things easily slip through.
In the end it comes down to a single "let's just make do for now"—and that "for now" can drag on for a long, long time.
After connecting WorkBuddy, things were different.
In the mold pricing step of new product development, they can put the design reference sheet, the mold inventory list, and the process pricing table directly into a WeCom smart spreadsheet.
AI reads the data, calls the pricing system, fetches the pricing table, then fills the results back into the design reference sheet.
Whether there's an inventory mold that can be reused, whether material prices are up to date, how much each mold costs—these questions can all be checked and calculated within the same process.
With concrete cost figures to work from, places that once seemed not worth buying a separate system for now have a chance to be managed properly.
AI isn't busy saving the world; it first helped the factory get these small accounts straightened out—and I think that's a good thing anyway.
△ Image AI-generated
That said, being able to do the work isn't enough. For AI to truly get used, it still has to follow the company's rules.
AI onboarding follows the company's rules directly
One person using AI and a whole company using AI are two very different things.
For an individual processing materials, you upload the files, explain the requirements, and get going.
But in an enterprise it's different—who can view which materials, who has permission for which operations, who signs off when the work is done—all of it is governed by rules.
Enterprise work is all about permissions, division of labor, and compliance. AI can't rummage through every document, and it certainly can't make decisions and act on its own.
If you want AI to genuinely get to work in your team, you first have to make it recognize the organizational structure and respect the company's way of doing things.
With WeCom and WorkBuddy connected like this, the most direct benefit is that no drastic overhaul is needed,In-place upgrade。
EnterpriseReuse WeCom's existing organizational structure and identity system to assign accounts, set permissions, and manage usage；
Everyone still chats and assigns tasks in WeCom every day, while the company keeps unified control over the boundaries of AI usage.
When AI comes to work at the company, there's no need to redecorate the entire office system around it.
WorkBuddy is just like a new colleague in the WeCom contact list — tasks can be assigned right in the chat window, with almost no learning curve.
Gradually, the information the company has accumulated over the years can also become a real, usable asset.
A person's work ideas from years of WeCom chats, meetings held, documents written, and knowledge bases built used to be scattered; once connected, all these accumulations can be handed to AI for retrieval and organization with one click.
Bottlenecks in cross-department work can also be untangled.
Often, work gets stuck simply because of constantly copy-pasting to pass information along.
WeCom plus WorkBuddy can string together cross-department, cross-system tasks; when done, results automatically return to WeCom conversations, spreadsheets, and approval flows, get passed straight to the next step, and can also track progress and send reminders at sticking points.
Only when the three layers — personal productivity, customer management, and organizational collaboration — are stacked together can it be calledintegrated internal-and-external intelligence, something that simply adding another AI chat window could never compare to.
Earlier, AI dividends were mostly held in the hands of a few. Some people amassed great prompts and built tools skillfully, turning out work quickly themselves, while overall team efficiency hardly improved.
But when these experiences, processes, and rules all enter the company, it becomes a corporate plug-in that knows everyone, remembers everything, and follows all the rules —
when something comes up, it really can step up.
Before, I used to beg it to spare me from onboarding training; now it seems AI knows the business better than I do.
In a while, it might be my turn to ask it for tutoring (doge).