Smartnet Technologies DOO
Jelene Glavaški 23, 18 000 Niš, Serbia
VAT ID: 110543210 · Reg. No.: 21361011
VAT ID: 110543210 · Reg. No.: 21361011
How Smartnet Technologies designed and built a quoting platform for a national B2B distributor: a speed layer on top of a legacy ERP, and an AI-driven engine that turns competitor quotes into counter-offers in minutes instead of days.
Smartnet Technologies DOO is a software development company based in Niš, Serbia: 25+ senior engineers, 25+ years in business. We work with companies across Europe and the US as an independent IT vendor or as an embedded part of their engineering team. Every client we have ever worked with came through a referral. We have never used cold outreach.
This document describes a production platform we analyzed, designed, and built for a national distributor: an AI-powered quoting workspace and competitive cross-reference engine, running on top of the client's existing legacy ERP.
"Our people are not losing deals because our offer is worse. They are losing them because we are slower."
Our client is one of the largest distributors of healthcare products in the United States: tens of thousands of SKUs, a nationwide field sales organization, and a market where the competitor's quote is usually already on the customer's desk.
Client identity is withheld by agreement. The architecture, the workflow, and the results described here are accurate and unmodified.
The legacy ERP was imperfect, but it was the system of record: every customer, contract, price level, and purchase history lived in it, and it was not going anywhere. Replacing it would have meant years of risk before a single salesperson got faster.
Instead, we built a focused speed layer on top: a modern web application that reads from and writes back to the existing landscape, giving every sales representative everything on one screen — catalog, pricing strategies, contract terms, live prices, alternatives, attachments. The ERP keeps doing what it does; the people who race competitors every day got a tool built for racing.
This pattern, wrap the legacy core, attack the highest-value workflow first, delivered value in months instead of years, and built the client's trust incrementally toward the platform's most ambitious feature: the cross-reference engine.
Everything a rep needed lived in different places. Assembling one quote meant switching between the ERP, spreadsheets, and email.
Deals frequently start with "here is what they offered me, can you beat it?" — in PDFs, scans, spreadsheets, and email attachments.
Matching competitor line items to the client's catalog meant manually re-keying hundreds of items across catalogs with no shared item numbers or units of measure.
The competitive cross-reference was the most valuable activity in the sales process, and the one the old workflow made too expensive to do at all.
PDF, scanned image, or spreadsheet, dragged onto the import screen and tagged to a session
A specialized document-extraction service reads item number, description, quantity, price, manufacturer, and unit of measure, asynchronously, with a real-time notification when done
Every item is checked against accumulated cross-reference knowledge: known competitor item numbers, industry code tables, manufacturer numbers. Exact hits are traced to their source
Whatever the tables don't know is scored against the catalog using a domain-specific matching approach, and presented as ranked suggestions with confidence scores
Matches are reviewed in an editable grid; corrections feed a curation flow so the engine improves with every session it processes
One click pushes the entire reviewed session, hundreds of matched line items, into an existing or brand-new quote
The quote detail screen gives a sales representative everything on the table: one editable grid holds the quote items, with pricing breakdowns, costing strategies, related-quote groupings, promotions, attachments, and a full integration log for transparency around it.
Individual page sections are isolated behind error boundaries, so a failure in one panel degrades that panel, not the rep's entire quote.
The ordering is a design principle, not an accident. Pass one checks exact cross-reference knowledge the distributor has already accumulated. Pass two handles everything the tables don't know, using a domain-specific scoring approach backed by a hand-built industry vocabulary that grew with the project.
A human stays in the loop at every step: a match popup shows the competitor's line next to the proposed item, ranked candidates, alternatives, and match history. Rejected and corrected matches feed a curation flow, so the engine's knowledge base improves with every session it processes.
Buy the commodity capability, build the domain moat. Document extraction is a solved, competitive market. Accumulated cross-reference knowledge is not, and that is exactly why it is defensible.
| Parameter | Value |
|---|---|
| Founded | 25+ years in business |
| Location | Niš, Serbia (GMT+1, same as Zagreb and Berlin) |
| Team | 25+ senior engineers, 80%+ MSc / BSc |
| Client acquisition | 100% through referrals |
| Travel | 1–3h from EU capitals, direct flights |
| Overlap | Full working day with all EU markets |
| Engagement model | Dedicated team / Team extension / Custom project / White-label SaaS |
This engagement: AI-powered quoting and competitive cross-reference platform for a national B2B distributor, wrapping a legacy ERP with a modern speed layer.
Other industries: Healthcare & Telemedicine · Veterinary Tech · Telecommunications (EU, Latin America) · Insurance · Medical Device Software · Embedded Systems
We adapt this pattern, legacy ERP wrap plus AI-driven workflow acceleration, to any distribution or enterprise sales environment.
Saša Tančev, CEO, Smartnet Technologies