SalesAI automationCRM

WorkWeek

Industry
Workforce software
Stage
Startup
Year
2026
Deliverables
4 workstreams
WorkWeek success story film: how Hivory built an AI prospecting engine for WorkWeek's sales team

01 The opportunity

WorkWeek is the HR and clocking system for field teams. Construction firms, civils contractors and scaffolders use it to clock crews in on site, run payroll and manage leave. Their sales team is brilliant on the phone, but too much of its week went on working out who to call. They asked us to take prospecting off their plate.

The vision

A steady stream of good prospects, researched properly and waiting inside their CRM every week.

02 The challenge

WorkWeek sells to a very specific kind of company: one with crews on several sites, hourly workers and timesheets still on paper. Those businesses rarely show up in bought lead lists, and finding them by hand meant hours of searching, spreadsheets and calls to reception to ask who handles payroll. Every hour spent researching was an hour not spent selling.

03 Our approach

We built WorkWeek a prospecting engine that runs on our own agent infrastructure, so their team had nothing to install and nothing to learn. It starts from the official company register, filters to the trades and provinces WorkWeek sells to, reads every company's website and scores it against their ideal customer. The best of them land in Attio every week with the decision maker attached.

01

Define who fits

We sat with WorkWeek's team and turned what they knew about their best customers into rules an agent can apply. A company qualifies on three of four tests, and still using paper timesheets is a must. Every qualifying company is then tiered A, B or C from what its own website shows.

The fit test: more than 20 employees, paper timesheets (required), hourly workers, several sites. A company qualifies on three of four
Tier A large contractors, Tier B specialist firms, Tier C small firms, each read from the company's website

02

Build the engine

Agents work through the company register, cut it down by industry and province, check each firm's CIDB grade and find its website. They read the site, score the fit and the tier, then find the person who signs and a way to reach them. Only companies that pass every step move on.

The engine in seven steps: company register, industry and province, CIDB grade, their website, tier and fit score, decision maker, into Attio

03

Deliver into Attio

Prospects arrive in WorkWeek's Attio every week through the API, with fields we set up together: tier, fit score, CIDB grade, the signals behind the tier, the source and when the record was last enriched. The team opens their CRM and starts calling.

An example prospect record in Attio with tier, fit score, CIDB grade and decision maker, details masked

04

Learn every week

Every batch gets feedback from the people making the calls, and a monthly review tightens the fit rules and the tiers. Each batch is sharper than the one before it.

The film, frame by frame

Tools and technology

Our stack

The tools and data behind the engine.

Hivory Managed Agents

Our agents ran the whole pipeline on our infrastructure, from the register to the CRM.

Managed Agents

CIPC and CIDB registers

Official company and contractor registers gave a complete starting list and a reliable read on size.

Firecrawl and Brave Search

Found each company's website and read it, so every tier is based on evidence.

Attio

WorkWeek's CRM. Every prospect lands there through the API with the fields their team works from.

04 The results

Before the full build we ran a pilot on 400 companies. It found 36 strong leads, 17 of them Tier A, without buying a single contact from a data provider. Once the engine was live it delivered around 50 enriched, qualified prospects a week, roughly 500 over the engagement. WorkWeek's team ended up with more good prospects than they could call.

~50

Qualified prospects every week

17

Tier A leads from a 400-company pilot

The pilot: 400 companies, 36 strong leads, 17 Tier A

A sales team that spends its week calling, not researching.