
Lead-gen software gives you 800 leads; WorkBuddy picks out the half you can actually contact
Last Tuesday afternoon, an old client in the building materials business exported over 800 leads from his lead-gen software backend and sent them to me, with a WeChat message: Old Chen, help me take a look at how to follow up on these. I opened the spreadsheet, glanced at it, and got a bit worked up. There were duplicates—the same company entered three times across three channels; incomplete emails; and company names written as abbreviations, with just a "Co" tacked on and nothing else. The ones actually ready to send emails to—I'd estimate less than half.
That evening I came across an article on how to choose AI foreign trade lead-gen software in 2026, going on at length about five dimensions and a five-step lock-in, with a ranking at the end. One line I agree with:
When choosing AI foreign trade tools in 2026, first clarify the task your company most needs to solve—that's more reliable than picking off a Top 5 ranking.
That's true, but it doesn't go deep enough. What most small and medium foreign trade companies most need to solve is what to do after finding them. Lead-gen software handles the first half; the second half it basically ignores.
These past two years people keep telling me, just get a lead-gen software and someone will find clients for you. That's only half true. Lead-gen software gives you a list, not clients. Between the list and clients there's still cleaning, categorizing, background checks, the first email, the second follow-up, quotes, samples. That whole chain—lead-gen software doesn't touch any of it. That selection guide also mentioned a claim that today's intelligent lead-gen systems can autonomously complete the whole set of finding clients, judging value, writing emails, and human-confirmed sending. The first two steps I believe; the email-writing step I have doubts about, because it doesn't know what date your production line is booked through this month, and it doesn't know which client dragged on payment last year.
So this time I didn't recommend a new tool to the old client. I had him first run those 800 leads through WorkBuddy.
Step one is cleaning. I had him create a new folder, drop the exported CSV and the spreadsheets from the other two channels into it, and have it dedupe by company domain, standardize company names to full names, and flag in a separate column the ones where the email doesn't match the company domain. From my testing, 800-plus records—deduping plus completing takes about forty minutes to produce a version, leaving just over 500 valid leads. It merges the ones that are obviously the same company, like ABC Trading and ABC Trading Co Ltd. But it also guesses wrong sometimes—two companies with very similar names that are actually in different countries, it merged into one. So this step absolutely requires spot-checking; I usually randomly flip through twenty, and if they match I move on, if not I go back and revise the rules and rerun. I wrote a post in August saying WorkBuddy cleaning foreign trade lead-gen spreadsheets can't replace judgment, and I still hold that view—it saves hands, not brains.
Step two is follow-up emails. The old client has buyers in five or six countries; before, one follow-up email was copied and pasted around, and clients could tell at a glance it was a mass send. I had him drop in the correspondence with clients in these countries over the past two years, plus product spec sheets, and have WorkBuddy draft separately for each client based on their market and the content of the last communication. The drafts that came out were clearly more presentable than a template—at least the opening could mention the model the other party asked about last time, and the tone was different. But price and lead time it can't fill in; it doesn't know what date your production line is booked through this month, and it doesn't know whether this client can be given payment terms. So my usage is to let it write the framework and wording, I fill in price and lead time by hand, then let it smooth out the tone.
Step three is categorizing inquiries. The client had seven or eight inquiries come in at once, with model, quantity, and destination port all mixed into the email bodies. I selected all the emails and dropped them in, had it categorize by model and produce a table, marking which ones had asked for a quote and which were first-timers. I wrote about this before—about ten minutes to produce the table, whereas doing it manually used to take half an afternoon. After the table comes out I still go through it myself, because occasionally it mistakes the MOQ in an inquiry for the order quantity.
After the three steps, looking back at that selection guide, I feel it put the emphasis a bit off. The guide says to first clarify the task most in need of solving, then use the same materials for small-scale testing, comparing effective communication volume and inquiry quality before and after. That thinking isn't wrong, but it assumes your bottleneck is at the finding-clients stage. The small and medium trading companies I deal with—the problem is mostly that after finding them, they can't handle them. Leads pile up in spreadsheets with no one cleaning them, emails get written with no one following up, inquiries come in and get answered two days later. Lead-gen software doesn't touch these tasks; WorkBuddy handles them all.
As for how to compare these two types of tools, let me just lay it out item by item. Finding clients, channel coverage, database scale—lead-gen software wins, no contest; WorkBuddy doesn't have this function at all, and it's not meant for it. But lead cleaning, email drafting, inquiry categorization, quote sheet organizing, multilingual replies—WorkBuddy handles all of these, and lead-gen software basically ignores them. I uninstalled Teyi and Xiaolanben long ago; the stage that really bottlenecks me isn't theirs. So my judgment is direct: lead-gen software determines your ceiling, WorkBuddy determines your floor. The ceiling is built with budget, the floor with process. When budget is limited, shoring up the floor first is more worthwhile.
WorkBuddy isn't without flaws. The knowledge base part I still haven't fully gotten working—scanned PDFs dropped in often can't be found; I suspect it's the chunking method, or maybe my files are just too old. Another time, to save effort, I wrote three needs—cleaning, categorizing, drafting emails—all in one paragraph and threw it at it, and what came out was a mess. Later I learned my lesson: have it do one thing at a time, and it's actually faster and more accurate. I've mentioned these two points in other posts too; they're not new pitfalls.
Its biggest limitation is that it doesn't know your business. It doesn't know which model has high profit margin for you, doesn't know which client dragged on payment last year, doesn't know whether your production line can take rush orders this month. These it can't give, and you shouldn't expect it to. So my way of using it now is very fixed: anything repetitive, formatted, requiring digging through old materials—hand to it; anything involving judgment and trade-offs—do myself.
Back to that old client in building materials. Over 800 leads, just over 500 ended up able to receive a first email, sent over three days, about twenty replies, five video calls scheduled. That conversion rate isn't high, but better than having him blast out 800 randomly. Lead-gen software helps you meet more people; WorkBuddy helps you turn the people you meet into deals that land. The former can be bought with money; the latter you have to grind out yourself bit by bit.
Physix Frontier