How Brasta Glass built its own automated B2B lead generation in 3 months
Brasta Glass built its own automated B2B lead generation with AI website research and got 371 replies and 153 positive answers from buyers in about 3 months.
- Published
- Updated
- Reading time
- 7 min read

TL;DR
- Brasta Glass, a Kaunas maker of custom shower enclosures and glass partitions, now runs its own automated B2B lead generation.
- An AI check of buyer websites replaced manual list cleaning. In about 3 months the team got 371 replies and 153 positive answers.
- Lean Sales Systems trained the team and set up the tools. The Brasta Glass team runs the campaigns itself.
On this page
Brasta Glass makes custom glass shower enclosures and partitions in Kaunas. With Lean Sales Systems, its own sales team learned to run automated B2B lead generation. In about 3 months it got 371 replies and 153 positive answers from buyers, with AI doing the list research that people used to do by hand.
| Client | Industry | Service | Market | Period |
|---|---|---|---|---|
| Brasta Glass | Manufacturing, glass | Sales systems integration and Training | Lithuania, plus an export test | May 2025 to October 2025 |
Key results
| Decision makers reached | Replies | Opportunities | Hours of manual work saved |
|---|---|---|---|
| 4,000+ | 371 | 153 | 670+ |
Opportunities are replies that Reply.io's AI tagged as positive. Decision makers reached is calculated from 371 replies at an 8% reply rate. Hours saved is an estimate (see below).
About the client
Brasta Glass is part of the Brasta group in Kaunas. It has made glass solutions for homes and businesses for 30 years: shower enclosures, restroom partitions and custom glass partitions. It sells to builders, installers and businesses in Lithuania and abroad, and has 51 to 250 employees.
The challenge
Brasta Glass had a strong product and a sales team, but finding the right buyers took too much manual work.
| Challenge | Why it hurt |
|---|---|
| Bought company lists were mostly the wrong fit | Only about 1 in 10 construction companies works with showers or bathrooms. Salespeople had to open each website to check. |
| Export lists often had only company names | Without a website or a contact person, about half of each list was lost before any outreach. |
| No clear outbound plan or targets | The team did not know what a good campaign looks like, so it could not tell what to fix. |
| Generic messages | Buyers ignored mass emails that did not mention their own projects. |
| Risk of spam folders | Sending a lot of cold email from the main domain could damage its reputation. |
| Know-how stuck with an outside agency | The group wanted its own people to run the system, not depend on a supplier. |
Markets and focus
Lithuania was the main market. It produced most of the positive conversations.
| Market | Who we reached | Companies | Volume | Result |
|---|---|---|---|---|
| Lithuania | Owners, directors and buyers | Construction companies, installers, interior designers and hotels that work with bathrooms | Just under half of the 4,000+ contacts, by email | The main source of the 153 positive replies |
Export campaigns to the Nordics, Poland and other countries ran from the same system. They were a low-cost test of foreign markets, and they showed that those buyers expect messages in their own language and a local presence.
The solution
Lean Sales Systems did not run the campaigns for Brasta Glass. We built the system with their team, trained them, and handed it over.
| Building block | What it did for the client | Why we used it | Tools |
|---|---|---|---|
| Lead research agents | AI research assistants read each company's website, kept only companies that work with showers, bathrooms or glass, and wrote a short note on each one for a personal opening line. | Bought lists were 90% the wrong fit, and checking them by hand wasted sales time. | Clay, a data platform that combines 100+ data sources and AI in one table, with OpenAI models |
| Email verifiers | AI found the manager's name on the company website. An email finder then found the work email and checked it exists before sending. If none was found, AI picked the best general email on the site. | Export lists often had only company names, so half of the leads were being lost. | Clay email finders, Rekvizitai.lt, LinkedIn Sales Navigator |
| Automated sequencer | An automated email and LinkedIn outreach tool (a sequencer) sent messages and follow-ups on schedule and stopped when someone replied. AI sorted every reply into categories such as interested, not now or referral. | Writing and following up by hand took too much time, and replies got lost. | Reply.io |
| Connected CRM | This project did not connect a CRM. Replies landed in 1 shared inbox, sorted by AI, and each salesperson answered from their own mailbox. | The team needed to see every warm reply the same day. | Reply.io inbox |
| Sending setup | New sending domains, DNS settings and warm-up: using new inboxes gently for 2 to 3 weeks so email providers trust them. | To keep emails out of spam and protect the main domain. | Reply.io |
| Team training | An outbound strategy with targets, video lessons, ready Clay templates and checklists for the whole Brasta group. | So the team could build new lists and campaigns without us. | Notion, video guides |
What made the difference. The AI website check. Instead of sending to a whole bought list, the team only wrote to companies whose website showed they work with showers or bathrooms. Each first email could say what the company actually does.
What changed
| Metric | Result | Tracked in |
|---|---|---|
| Decision makers reached | 4,000+ (calculated from replies and reply rate) | Reply.io, September 2025 |
| Replies, all campaigns | 371 | Reply.io, September 2025 |
| Reply rate, all campaigns | 8% | Reply.io, September 2025 |
| Opportunities | 153 replies tagged positive by AI | Reply.io, September 2025 |
| Interest rate | Over 3% of all contacts showed interest | Reply.io, September 2025 |
| Referrals to a colleague | 22 replies | Reply.io, September 2025 |
| Time to first campaigns | About 1 month from training to launch | Project plan and emails |
| Who runs it | The Brasta Glass sales team, in house | Project plan |
These numbers come from the September 2025 review, about 3 months after launch. At that point the team had many warm conversations, but they had not yet turned into orders. Cold buyers needed more education and follow-up before talking about prices. The next step was a slower follow-up path and more message tests.
Extra benefits
- The sales team stopped checking company lists by hand. The AI does the first filter.
- Brasta Glass owns the system: lists, templates, prompts and campaigns stay with the company.
- A quick, low-cost test of export markets showed where local language and a local presence are needed.
- The team learned to recycle contacts: "not now" replies can be contacted again in 3 to 12 months.
- The training covered the whole Brasta group, so other units can reuse the same method.
What this could be worth (estimates)
The numbers below are estimates, not results reported by Brasta Glass.
Work the system did
| Metric | Estimate |
|---|---|
| Value of outreach work done | About €18,000 of activity (4,000 contacts x €4.50) |
| Hours of manual outreach work saved | About 670 hours (4,000 contacts x 10 minutes) |
| Full-time months saved | About 4 months (670 hours / 160) |
| Hours to qualify a 10,000-company list by hand | About 330 hours (2 minutes per website) |
| Manual research time cut by the AI website check | About 95% |
Assumptions: €4.50 per manually processed contact, 10 minutes per contact for research, writing, sending, 2 follow-ups and logging, 160 working hours a month, 2 minutes to check 1 company website by hand. The value of outreach work is activity value, not return.
Projected deals within 12 months
We do not project deals for this case. At the September 2025 review the conversations had not yet become orders, and the team took over all follow-up itself. Brasta Glass does not share deal sizes, so we do not estimate revenue.
What happened next
In October 2025 we helped the team start a new campaign for Baltic companies, with leads from a business directory filtered in Clay and sent through Reply.io. The export team paused AI campaigns in early 2026.
Why this is a system problem
Most companies treat slow lead generation as a hiring problem. It is a system problem. When AI does the research and a sequencer does the follow-up, your salespeople spend their time on buyers who already said yes to a conversation. Lean Sales Systems can build this with your team or run it for you. See more results on our case studies page.
FAQ
How long did it take to get the first results?
About 1 month. Training and setup ran in May and June 2025, and the first campaigns went out in June. By September 2025 the team had 371 replies and 153 positive answers. Most of the time went into building clean lists, not into sending.
How much work did it take from the Brasta Glass team?
The team joined training sessions and working sessions, then ran the campaigns itself. Lean Sales Systems built the templates, prompts and the plan, and answered questions by email. After that, building a new list and campaign became a routine task for the team.
Who did the campaigns target?
Companies that work with bathrooms: construction companies, installers, interior designers and hotels. The AI website check removed companies that do not work with showers or glass. In each company the team wrote to owners, directors or buyers.


