The Order Is the Output
Say you have a list of 200 target factories and can make 20 serious calls a day. Working through the list takes two weeks — and who you call in the first three days all but determines your numbers for the month.
Most salespeople sort “top to bottom” or by “whatever looks right.” That's not an attitude problem — the human brain simply can't rank 200 companies across five or six dimensions at once. AI happens to be good at exactly that. Gong's research claims that sales teams making heavy use of AI tools generate 77% more revenue per rep than those that don't — take vendor-run research with a discount, but the direction is right: one of AI's biggest contributions to sales is turning “by gut feel” into “by signal.”
The five dimensions below are the ranking signals we've found most useful when evaluating factory customers.
Five Dimensions, from Hard to Soft
1. Size signals: do they have the money to buy what you sell?
Registered capital, social-insurance headcount, plant footprint — these hard facts determine a customer's purchasing scale. If you sell automation equipment, a factory with 300 insured employees and one with 30 call for completely different plays; if you sell raw materials, production capacity maps directly to your order volume. In the first sorting pass, cut by size and sink the obvious mismatches to the bottom.
2. Industry fit: can their production line actually use what you sell?
Two companies both registered as “plastic products” — one doing injection molding, the other blown film — use entirely different equipment, resins, and auxiliary materials. Industry fit has to go down to the process level — which is also why your list's data source matters: registered business scopes all read alike; only real primary product categories tell you who is actually your customer.
3. Regional density: can one trip string together three visits?
Industrial customers are heavy on in-person visits, so regional concentration directly affects your time cost. Cluster your list by industrial cluster: a route in Suzhou that covers eight factories in two days naturally outranks equally scored customers scattered all over the map.
4. Expansion signals: are they spending money right now?
A new plant going up, new production lines, intensive hiring, fresh funding — a factory in expansion has its buying window open. This kind of activity is exactly what AI should watch for you: have it periodically scan the companies on your list for developments and push the ones showing expansion signs to the front of the queue.
5. Decision-maker reachability: when the call connects, do you reach the right person?
Two equally ranked factories — with one you can directly reach the owner or the head of purchasing, with the other all you have is the front-desk landline. No hesitation about which to call first. Flag the availability of decision-maker contact information at the list stage and you skip a great deal of the futile “hello, could you transfer me to purchasing” loop.
Let AI Execute: From ICP to a Ranked List
Once the five dimensions are clear, execution can be handed to tools. The workflow goes like this:
Step one: spell out your ideal customer profile (ICP). No forms to fill — plain language: “I sell food-grade lubricants. My targets are beverage and dairy factories in East China, mid-size or larger, ideally expanding capacity.”
Step two: have AI produce an initial shortlist from your profile. In Tianxia Gongchang AI, that sentence turns directly into a multi-turn search over a database of 4.8 million real factories in active production — it first asks you a few narrowing questions (beverage or dairy first? region down to the province or the city?), then produces a list against your criteria, with the size and region dimensions already sorted at the search stage. The platform's filter system also includes “decision-maker contact information,” so the fifth dimension can likewise be settled at the list stage, rather than discovering there's no one to reach only after you dial.
Step three: do an incremental ranking of the top thirty to fifty. Hand the shortlist to an AI assistant for an activity scan, flag the factories with recent expansion, hiring, or contract-win moves, layer on industry fit, and you have your final calling order.
Step four: re-rank every two weeks. Forrester's research shows a typical B2B buying decision involves 13 internal stakeholders and runs on a timescale of months — a customer with no movement this month isn't a dead lead; the timing just isn't right yet. Have AI remember each account's blocker, and when the blocker changes (say, their new production line comes online), the account floats back up.
Don't Overrate the Tool — or Your Gut
A splash of cold water: in McKinsey's 2025 survey, more than 80% of companies said AI had yet to make a material impact at the profit level. The reason isn't that AI doesn't work — it's that most people use it for nice-to-haves and never change the actual order of their work.
Lead tiering is precisely the lever that “changes the order”: it adds nothing to your workload; it only changes the sequence of your calls. Same 200-factory list, same 20 calls a day — ranked by signal versus ranked by gut, a month later your pipeline quality traces two different curves.
Start with your next list: go to https://www.tianxiagongchang.com/ai and pull your target factories with one sentence, so the size, region, and decision-maker dimensions are sorted the moment the list is generated, then have AI layer the expansion signals on top. Look back at your call conversion rate two weeks later — the numbers will speak for the method.