Skip to content
E
ES
Client story · Language education

How Business English Academy cut LinkedIn prospecting from 10 hours a week to 90 minutes

BEA's outreach to corporate HR and L&D teams ran on manual LinkedIn prospecting: one person finding prospects, writing messages in three languages, and checking for replies by hand. We built a system that prepares outreach, classifies replies by interest, and alerts BEA the moment a lead is hot. Within two months of launch, that same campaign closed BEA's first client.

Business English Academy logoSeptember 28, 2026
Watch Kate's video below

~35 hrs/month freed

Kate's estimate of time returned

Manual prospecting: ~10 hrs/week before → 1–1.5 hrs/week after

10–20 min

Typical hot-lead response time

Down from up to a couple of hours on a busy day

1 client → 5 groups, 22 learners

From the first campaign

Client closed from the campaign launched 29 Jun 2026; groups launched 1 Sep 2026 after negotiations through August

The challenge

Before the system

Before the project, BEA found and contacted corporate prospects on LinkedIn by hand: one person sending connection requests and follow-up messages across Ukrainian-, English-, and Russian-speaking audiences, then checking manually for replies.

That took Kate and her team an estimated 10 hours a week, and it meant hot leads could sit unanswered for a couple of hours on a busy day - exactly the moment a corporate L&D buyer is comparing options. Success meant getting that time back without losing the personal, language-appropriate tone that makes a cold LinkedIn message actually land.

The shift

Manual prospecting time

Before

~10 hrs/week

After

1–1.5 hrs/week

Hot-lead response time

Before

Up to a couple of hours

After

10–20 minutes

What changed

How it works now

  1. 01

    Outreach runs across three languages at once

    LinkedIn outreach campaigns for HR and L&D teams run in Ukrainian, English, and Russian in parallel, instead of one person managing each language by hand.

  2. 02

    Replies are read and classified automatically

    Every reply is scored Hot, Warm, or Cold based on interest signals, and greetings are localized so a message reads naturally in the prospect's language: including correct grammatical name forms, not a literal template swap.

    For example

    Hello, Дмитре!

    Hello, Dmytro!

    Name formatting matches the language of the message. "Дмитре" is the correct Ukrainian vocative form, but it doesn't belong in an English greeting: keeping "Dmytro" as-is is what actually reads naturally here.

  3. 03

    Hot leads reach Kate immediately: she still makes the call

    A hot reply triggers an email alert right away. The system flags and prioritizes; every reply, follow-up, and sales decision is still BEA's.

Behind the workflow: Expandi runs the LinkedIn layer, n8n connects the pieces, and an LLM handles language-aware name formatting and lead classification.

Results

What happened

The shift above has held since BEA's first campaign launched on 29 June 2026 - not a one-off week.

From that same campaign, BEA closed one client. Negotiations ran through August, and five English-learning groups covering 22 learners launched on 1 September 2026.

Attribution & limits

Kate notes that sales decisions at BEA can take up to six months, so this is an early result, not a stabilized conversion rate. Responding to interested leads and closing the sale remain BEA's own work; the system's job is making sure a hot lead never sits unanswered.

“What I value most is that they don't just install tools. They understand the business process, identify where time and resources are being lost, and build a system around it.”

Beyond this story

Metrics worth tracking for a system like this

If you're weighing something similar, these are the numbers we'd want visibility into from day one.

Response rate

The share of outreach that gets any reply - the earliest signal the messaging and targeting are working.

Not yet published for BEA.

Contracts closed

The commercial outcome that actually matters - replies and calls are activity, not revenue.

1 so far, within two months of launch (see above).

Time saved

Whether the system is actually returning hours, not just adding another dashboard to check.

~35 hrs/month, by Kate's estimate (see above).

LLM classification accuracy

How reliable the Hot/Warm/Cold scoring is - a misclassified hot lead is a missed opportunity, quietly.

Not yet measured for BEA; worth a periodic spot-check against human review.

Is manual outreach eating your team's week?

Let's Fix It
E

EsperaStudio

Initializing Systems
Core.v2.5.094%