Back to blog Customer story

How an industrial equipment maker turned its website into a second sales team

A working showroom floor, where the website is the second sales team
A working showroom floor: the website as a second sales team

The owner — call him Mr. Li — has sold industrial automation equipment for fifteen years. Deals routinely run into six figures. His belief, like most of his industry: big-ticket equipment is won at trade shows and over factory visits. The website? A digital business card that proves the company still exists.

Early this year, that playbook stopped working. Trade show leads halved. Paid search clicks cost a fortune and qualified nobody. At the Monday meeting his sales director kept slamming the table: "The lead quality gets worse every week. I can't run a team like this."

Out of options, Mr. Li gambled on rebuilding the website and taking content seriously. Nobody on the team believed industrial equipment could actually sell online.

Ninety days later, a dashboard landed on his desk: 41% of new qualified leads now came from AI search channels. The website had become a second sales team — one with no salary, no commission, and a 24-hour shift pattern.

Here is the full 90-day timeline, with the numbers.

Day 0: The decision — strip the site to its studs

When we took the project on, the site took eight seconds to load. The homepage was a full-screen animation — grand slogans, orchestral music, zero information.

We told him plainly at the kickoff: "Your buyers came to evaluate equipment, not to enjoy the soundtrack. AI crawlers can't even parse your headings. This needs a rebuild, not a facelift."

He winced at the cost, then signed off: "Do it. If it brings leads, I don't need the site to be pretty."

From showing equipment to guiding buyers: every path is designed for acquisition
From "showing equipment" to "guiding buyers": every path serves acquisition

Day 1–15: Turn the website into a spec sheet

For two weeks the team did exactly one thing: entity definition and structuring.

  • Cut the adjectives. Every "empowering synergy" sentence was deleted.
  • Published real spec tables: for the three flagship machines — power requirements, footprint, maintenance intervals, delivery windows — written like a maintenance manual.
  • Built scenario FAQs: a dozen question-and-answer blocks per product page. "Our plant voltage fluctuates — can your equipment handle it?" "What's the shortest lead time, and can it be expedited?"
The turning point: Mr. Li hated the first draft. "It reads like a spreadsheet, not a website," he said. We held the line: for a seven-figure industrial purchase, precision and transparency are the design.

Day 16–45: Build the trust layer

A clean skeleton means nothing if AI doesn't trust it. The next month we built corroboration across the web.

  • Deleted the vanity press. All the "dignitaries visited our booth" posts went.
  • Published verifiable anchors: "300+ units delivered, average 15% yield improvement for customers, service engineers on-site within 2 hours."
  • Distributed external corroboration: the CTO's buyer's-guide was placed in two industry forums and a trade publication.
An unexpected side effect: by month's end the sales director reported that the first WeChat message from prospects had changed. It used to be "how much is your machine?" Now it was "your materials mention a 15% yield gain — how do you achieve that?" Buyers had already been briefed by AI before they ever contacted us.

Day 46–75: The AI dividend arrives

These thirty days were when compounding kicked in. As AI assistants with live web access became the default research tool, queries like "automation equipment supplier, Yangtze delta, delivery within 30 days, automotive parts experience" began pulling structured answers straight from the site and the external guides.

  • Traffic quality shifted: daily unique visitors went from dozens to two or three hundred — nearly all problem-aware buyers.
A good problem: one rep complained to the sales director, "I'm answering inquiries on WeChat at 10 p.m. — can marketing turn down the lead volume?" The director relayed it to Mr. Li, who sent the web team a bonus that week.

Day 76–90: The review

At the 90-day mark, the numbers settled:

  • Total traffic up 400% over the pre-launch baseline;
  • 41% of new qualified leads attributable to AI search and recommendation;
  • Sales cycle for AI-sourced leads roughly 20 days shorter than trade-show leads.
100%75%50%25%0%DAY 0–15StabiliseDAY 15–45Trust layerDAY 45–75AI dividendDAY 75–90ReviewFrom Day 46, AI-sourced leads climb sharply90-day result: 41% of new qualified leads

Why did conversion get easier? Because buyers arrived pre-briefed. The specs, trade-offs and price bands had already been explained by an AI assistant using our client's own materials. Sales conversations skipped the "what is automation anyway" chapter and went straight to commercial terms.

What Mr. Li actually did right

At the celebration dinner he put it well: "I used to think a website was decoration. Now I see that buyers treat it like a dictionary."

Three things, done to extremes:

  1. Plain language: AI ignores rhetoric, so we gave it data and tables.
  2. Scenario depth: every way a plant can break was answered on the site before it was asked.
  3. Corroboration: positive, factual material across the web gives AI the confidence to name you.

Mr. Li now shows off his "second sales team" to anyone who will listen — one with no base salary, no social insurance, and a perfect attendance record, quietly filtering for the million-yuan deals that are actually winnable.

A question worth asking tonight: is your website a dusty digital business card, or a top-performing rep quietly working the night shift? Open an AI assistant after work today and ask about your own category. See who gets named.
Sam Gu

Co-founder of Ditorch and an AI acquisition consultant. He only writes about playbooks he has personally run — so your website can become the rep that closes.