Metal Zbiorniki

Metal Zbiorniki manufactures steel tanks and pressure vessels — most of it made to order, where every enquiry used to start as free text and turn into a slow email thread. Quotes lived in a shared inbox split between the owner and one colleague, so enquiries got duplicated, follow-ups slipped, and pricing a single tank could take a day or two of digging through an Excel cost sheet and past offers.

The brief was never just "a new website" — it was the whole sales pipeline. We started with the site (WordPress → Payload CMS + Next.js, in Polish, English and German with AI-assisted translation) and paid-search growth, then worked backwards from the enquiry to fix how quotes actually get made.

Tech wrap-up

The pipeline, stage by stage:

• Website rebuild — WordPress → Payload CMS + Next.js, live in production, trilingual (PL / EN / DE) with AI-assisted translation; conversion tracking and paid search wired in end to end.
• Google Ads run semi-automatically — we manage the paid-search account with an LLM wired straight into Google Ads and GA4 (over MCP): it reads the campaign, search-term and conversion data, then drafts the changes — negatives, budgets, bids, ad copy. A human approves every one before it goes live.
• Instant-quote configurator — on made-to-order product pages, tank parameters (capacity, material, fittings, inspection body) turn into a ballpark price range and a complete, structured enquiry — no more blank-text emails.
• Searchable quote history — years of email offers and the client's own cost sheet mined into a product catalog and a pricing model calibrated to their real calculations.
• CRM — a sales pipeline with clear stages, owner assignment, and reminders (a tank with no quote after three days, a follow-up due tomorrow), plus win/loss analysis — replacing the shared-inbox guesswork.
• Inbox integration — incoming free-text enquiries are parsed into structured leads automatically, with a preliminary quote ready before the owner even opens the email.
• AI-drafted quotes — the offer is drafted from the cost model and similar past quotes; the salesperson confirms and sends instead of building it from scratch — minutes instead of 1–2 days.