Book a callLogin

Customer storyIndustrial roboticsKewazo

Their buyers live at plants. Their CRM only knew company names

Kewazo's Liftbot is a battery-powered lifting robot that replaces cranes and manual handling on industrial turnarounds, maintenance, and capital projects. A single deal runs through several people spread across a site, not one contact at a company. A standard sales database sees a flat list of names. Kewazo's world is a hierarchy. The studio rebuilt their RevOps data around it.

Plant sites analyzed
531
Contacts recovered
202
Reply rate
20–25%

The work

Selling to a plant, not a company

Kewazo builds Liftbot, a battery-powered lifting robot that replaces cranes and manual handling on turnarounds, maintenance, and capital projects at petrochemical and industrial sites. The buying unit is not a company — it is a plant, where several people shape the decision rather than a single contact at headquarters.

Each one shapes the deal differently, so a rep has to know which of them to talk to — and none show up cleanly in a standard sales database, which knows a contact works at Dow Chemical but not which plant he stands in. In field sales, the site matters more than the company.

The buying unit is a plant. The tools only knew the company name.

Before Extruct

Four ways the tools fell short

No location context. Tools resolve the employer, not the site. For a field team, the plant is the unit of work — and it was invisible.

No plant or project layer. The real buying unit is a plant — a site with its own people, contractors, and projects — and standard databases have no concept of that hierarchy.

Field intelligence is hard to capture in the moment. Reps are on job sites, not at desks, so what they learn on a visit lives in voice notes and memory — with no easy path into the CRM.

And the data moves constantly. Industrial workers change contractors and projects all the time, and without automated tracking, no team could keep a contact list current by hand.

The same employer showed up three different ways — Apex Scaffold, Northgate Industrial, Apex Access — and nobody could see it was one company.

What changed

Six weeks

The studio took a sample of Kewazo's CRM: 1,200+ contacts in flat lists, no plant or site context, the same employer scattered under multiple aliases. We cleaned it, verified every contact against live sources, and restructured it around the way Kewazo actually sells — plants as first-class records, each with the right people, projects, and location context attached.

On top of the rebuilt foundation, the studio shipped the tools the field team runs on: a site-intelligence map that replaced manual region lookups, Claude Code automations for the repetitive admin, and a Slack agent that turns a rep’s post-visit voice note into CRM records, contact lookups, and a follow-up queue — answering “what’s our status on Dow Freeport?” in seconds, from any device. Slack is where the team already worked; the agent extended the habit instead of adding a tool to log into.

“Our buyers don't live at companies, they live at plants. Extruct rebuilt our data the way we actually sell, then put an agent on top of it.”

Artem Kuchukov CEO Kewazo

At scale

What the rebuild produced

  1. 531 plant sites mapped

    Identified across three accounts — Kiewit, Brand, and Dow — each a first-class record with the right people, projects, and coordinates.

  2. 1,206 contacts verified and re-tiered by plant

    Each verified with real-time data, assigned to the right plant, and re-scored A, B, or C based on the site they actually work at.

  3. 500+ new contacts discovered via ICP scoring

    Right-fit decision-makers the flat list never surfaced.

  4. 202 contacts recovered from the “unreachable” bucket

    17% of LinkedIn profiles re-verified — the right contacts reps had written off as gone.

  5. 7% of stalled contacts removed

    Left the company or the wrong contact entirely — removed so reps work from an accurate list.

  6. 50 min → 3 min conference prep per contact

    For a 60-meeting industry conference.

The tools

What the field team runs on

Two of the systems the studio shipped: every plant as a first-class record, and a Slack agent that turns field notes into CRM updates and answers account questions on the spot.

Site intelligencePlants ranked by tier, each with a verified, current record — the map that replaced manual region lookup.
Gary, the Slack agentField notes logged to the CRM, and in-channel answers like “what’s our status on this plant?”

In the field

Sharper outreach, fewer wasted touches

  1. 20–25% reply rate on warm, verified contacts

    Reaching the right, verified person at the plant — not a name off a generic list.

  2. 5% reply rate even on cold outreach

    Versus near-zero on the old generic lists — relevance moved cold replies off the floor.

  3. 0% → engaged when the message was relevant to the site

    Outreach tied to the rep’s own patch read as relevant, not generic — and got replies instead of silence.

  4. 5–7 → 3 job-change emails to immediate actions

    A job-change outreach test: three contacts acted immediately, two more showed interest.

What we shipped

Data foundation. Site intelligence. Claude Code skills. Gary, in Slack

One six-week engagement, four interlocking systems the field team now runs on every day.

Talk to us

Want this built for your team?

A 30-minute call. We ask about your setup, share how we’d approach it, and see if Extruct is a fit.