SalesDots watches public tenders across the EU and beyond — TED, national portals in Poland and Germany, Ukraine's Prozorro, World Bank projects — reads them with AI, and delivers the IT deals that match your business: translated, classified, and ready to act on. Including the ones that haven't been published yet.
The EU publishes hundreds of billions of euros in technology procurement — but the tools to navigate it were built for procurement lawyers, not for IT sales teams.
TED publishes hundreds of notices every day in 24 languages. The IT deals you could actually win are hidden between road works and hospital catering.
By the time a tender is published, the incumbent has been positioning for months. Keyword alerts tell you when the race is already half over.
Staff-augmentation deals hide behind generic 'IT services' labels in every local language. Portals can't tell buying people from buying software.
Who is this buyer? Who won their last contract, at what price, against how many bidders? Without history, every bid is a guess.
Every TED notice across 85 countries — plus Poland's national BZP portal, Germany's Bekanntmachungsservice (DÖE) with the below-threshold notices that never reach TED, Ukraine's Prozorro, World Bank procurement, Moldova's MTender, and Kosovo's KRPP — continuously, including the authoritative raw XML that generic portals throw away. Scoped to technology procurement: IT services, software, hardware, networks, telecom, and staff supply.
Each notice is classified on three axes — offering, IT domain, and vendor — every label backed by a confidence score and a verbatim quote from the notice. Semantic embeddings link re-tenders across years and languages; titles are translated across all covered languages.
Your interest profile turns the stream into a personal To-Do inbox: new matches, approaching deadlines, changes on tracked tenders, and predicted re-tenders — plus a daily email digest, only when there's something new.
Teach SalesDots your business once — what you sell, who buys it, and what you never bid on. From then on, every incoming tender arrives pre-screened: with a verdict and the reasoning behind it.
Every tender gets a call — Pursue, Watch, or Skip — with plain-language reasons. Your inbox becomes decisions, not homework.
Every rejected tender shows exactly which of your rules cut it. Spot a wrong skip? Fix one rule — not the whole pile.
New tenders are screened while you sleep. The right ones are waiting in your To-Do before your competitors have opened their laptops.
Clone a ready-made profile, adjust the rules to your business, and you're screening in minutes.
Screening runs through odnoga, the AI gateway SalesDots runs on — nightly batches at the vendor's genuine 50% discount, every verdict on the cost ledger. How SalesDots uses odnoga
CPV codes tell you almost nothing. SalesDots classifies every notice along three axes — what is being bought, in which IT domain, and whose technology — with a confidence score and a verbatim quote from the notice behind every label.
Ministry of Digital Affairs · PL · €1.9M · services
The contracting authority requests the secondment of 12 senior Java developers to reinforce the delivery team of its central e-services citizen portal. Candidates must document 3+ years of experience with Oracle Database administration and be available on the authority's premises within 30 days.
“secondment of 12 senior Java developers”
“central e-services citizen portal”
“Oracle Database administration”
What kind of deal this is — the axis that decides whether it fits your business model.
Which part of IT the buyer is investing in.
Whose technology is named — tied to our canonical vendor registry.
Labels are earned, not guessed: the classifier must quote the line of the notice that proves each label, and an adjudication gate decides what counts as an IT contract at all — bundled software inside a construction tender doesn't make the cut.
Not another keyword-alert portal. A full intelligence layer over EU procurement, built around how IT companies actually sell.
Tenders come to you. New matches, deadlines, and changes land in one actionable inbox — no more daily portal safaris.
Know when a buyer will re-procure — months before the tender is published. Each forecast shows its own measured likelihood and time window, recalibrated nightly.
Every contracting authority profiled: procurement history, patterns, open and past tenders, and who they buy from.
1,040,000+ contract awards with named winners, awarded vs. estimated values, and bidder counts — winners matched to official registries wherever the notice allows, unverified links labelled as such — and every buyer's purchase history rendered as an analytics dashboard on their account.
AI detects when buyers procure people — developers, admins, specialists — across all languages. A first-class filter, not a keyword hack.
Price vs. quality weights parsed from every notice — and evaluation criteria mapped to one concept dictionary across 23 languages, with the buyer's original wording always shown — so you know how each tender is scored before you invest in a bid.
Embedding-based matching connects today's tender to its predecessors — even when the buyer renamed everything.
Notice titles and descriptions machine-translated at scale. Read a Dutch staffing tender in the language you work in.
Chat with any tender, run a bid / no-bid analysis, and get requirements checklists, pricing hints, budget context, and document analysis on demand.
What does this buyer already run? 53,000+ organization-to-technology links across 15,000+ organizations, read from tender notices, company websites, vendor reference pages and deep web research, every one stored with its evidence — vendors, versions, even cloud choices.
Market size, momentum since 2020, top buyers and top winners for any niche — EUR-normalized, on one dashboard.
Describe your niche in your own words. AI matches it against every open tender and feeds your To-Do automatically.
Search any supplier — by name or VAT / registry number — and see who they sell to, what they win, and when. Competitor and partner mapping from real award data.
Corrections, deadline extensions and re-publications fold into a single dossier, and the deadline you see is the one in force — never the original that a later notice replaced.
After a Polish deadline passes, SalesDots fetches the buyer's budget, the opening information and the bidder list with prices — tracked tenders first, automatically.
SalesDots reads contract durations and each buyer's purchasing rhythm to forecast when they'll return to market. Then it does something no portal does: every night it checks its past predictions against what actually got published. Every forecast carries its own measured likelihood — '95.8% likely · ±9 days' — so you know exactly which ones to bank on.
forecasts whose measured likelihood puts them in the top tier — a tender for the same subject really appeared in the predicted window, verified matches only
across forecasts spanning months to years
and counting — the engine keeps learning
repeat procurements of something the buyer already bought — the market you can see coming
A hit means a tender from the same buyer and for the same subject — the same system, product, or service, verified by our subject engine, not just a similar category — was actually published within the predicted window. Forecasts for buyers that published nothing at all after the window opened are set aside as unobservable — they count in neither half of the rate. These scores recalculate automatically as new outcomes arrive — you're looking at today's numbers, not a press release. Top-confidence forecasts are the ones the engine's own calibration puts in its highest likelihood tier; the strict rate for every forecast is shown right beside them, never hidden.
The pipeline is scoped to technology procurement from the first byte — every classifier, every filter, and every prediction is trained on how the public sector buys technology.
Find delivery projects that match your stack and references — filtered by subject, country, and value.
The only feed that reliably separates 'buying people' from 'buying products' — in every EU language.
Track where your product — or your competitor's — is being procured, renewed, and replaced.
Spot managed-service contracts approaching expiry and position before the re-tender goes public.
These counters are live production data — not marketing estimates.
Every notice AI-classified
With named winners & values
Contracting authorities
From contract expiry
Across 7 official platforms
15,000+ organizations · 1,289 technologies · tenders, websites, vendor references
SalesDots harvests directly from the official procurement portals — no resellers, no scraped aggregators. Every notice links back to its origin.
From seven official platforms: TED (Tenders Electronic Daily), the EU's public procurement journal; Poland's national BZP portal (ezamowienia.gov.pl), which adds below-EU-threshold tenders TED never sees; Germany's Bekanntmachungsservice (oeffentlichevergabe.de), which adds the German below-threshold notices that by law never reach TED; Ukraine's Prozorro; World Bank procurement notices; Moldova's MTender; and Kosovo's KRPP. SalesDots harvests every notice — including the authoritative raw XML — across 85 countries and keeps it continuously updated.
The entire pipeline is scoped to technology procurement: IT services (CPV 72), software (CPV 48), hardware, networks, telecom, and staff supply (CPV 7962). AI classifies every notice by subject (vendor, product, technology) and engagement type — project delivery, managed service, or body leasing. It even reads the technology out of tender documents, mapping what 15,000+ public organizations already run.
You describe your business in a screening profile: what you sell, which buyers matter to you, and your hard no-gos. SalesDots then screens every tender — past and incoming — and hands you a verdict: Pursue, Watch, or Skip, each with the reasoning spelled out. Skipped tenders always show which of your rules cut them, so you can trust the filter and tune it in seconds. Start from a ready-made template or build your own; new tenders are screened automatically every night.
SalesDots reads contract durations and end dates from notice text, then forecasts when a buyer will re-procure — before the new tender is published. That gives you months of lead time to position against the incumbent. Re-tendering is the norm, not the exception: in the last 6 months alone we identified 17,000+ tenders as repeat procurements of something the buyer had bought before. And we grade ourselves in the open: every forecast is checked nightly against what actually got published — 4,900+ outcomes verified so far — and a hit only counts when the new tender is for the same subject, not just a similar category.
Yes — this is a first-class filter. AI detects when a buyer is procuring people (developers, admins, specialists) rather than a product or project outcome, across all languages and local terminology, and lets you filter on exactly that.
All 85 countries publishing on TED are covered, with the deepest volumes in Germany, Poland, Spain, France, and the Netherlands. Poland and Germany get double coverage: TED plus the national portals — Poland's BZP (41,000+ national notices) and Germany's Bekanntmachungsservice for below-threshold tenders. Ukraine is covered end-to-end through Prozorro, and World Bank-financed projects add tenders beyond Europe. Everything is translated to English and Polish.
Define an interest profile (countries × IT subject categories × engagement type) and SalesDots delivers matches to a proactive To-Do inbox, tracks changes on tenders you follow, and sends a daily branded email digest — only when there is something new.
Yes. Since August 2026 SalesDots harvests Germany's official Bekanntmachungsservice (oeffentlichevergabe.de), which publishes the below-threshold notices (VOB, VOL, UVgO) that German law never sends to TED. Each German-only tender is deduplicated against TED on the eForms notice identifier, so you see one tender with links to every platform that published it.
Yes. The Winners view lets you search any supplier by name or by VAT / registry number and shows the buyers they sell to, what they win, the attributed contract values and the years — built from hundreds of thousands of award lines. Where a winner cannot be matched to an official registry, the link is labelled unverified rather than guessed.
Measured, not assumed. The engagement classifier (project delivery vs. managed service vs. body leasing) is a trained model over each notice's embedding; on a hand-judged set of 485 production notices it reached 85.9% accuracy, with 95.5% precision on high-confidence labels — up from 33.5% for the keyword-anchor approach it replaced. It runs at no AI cost per notice, and every label shows its confidence.
Join the waitlist and get tender intelligence that works for you — matched to your countries, technologies, and engagement model.
Join the WaitlistTender data sourced from official procurement platforms: TED (Tenders Electronic Daily), Poland's BZP (ezamowienia.gov.pl), Germany's Bekanntmachungsservice (oeffentlichevergabe.de), Ukraine's Prozorro, World Bank procurement, Moldova's MTender, and Kosovo's KRPP. Counters reflect live production data and update continuously.