Acme Ltd, ACME Limited and Acme Group walk into a database

We'd tell you the punchline, but they're still arguing about which one of them is real.

Ask a room full of people how many business customers their organisation has, and you'll rarely get one answer. Finance has a number. Procurement has a different one. Whoever looks after the data goes quiet for a moment and then says "it depends".

Nobody has done anything wrong. Business data just multiplies when nobody's looking. One company becomes Acme Ltd in billing, ACME Limited in a supplier list and Acme Group in a spreadsheet from 2021. Before long, one business has become five, and every report built on top of it is quietly off.

BusinessLinks fixes that. It finds every version of a business hiding in your data, resolves them into one trusted golden record, and enriches each one with verified Companies House data. All of it happens inside your own Snowflake account. It's messy data in, resolved companies out.

What that means for you, in 30 seconds

1. One true count of who you do business with. No more "it depends".
2. Every business verified against Companies House. You get the registered name, company number, current status and director details.
3. Hours of manual matching handed back. Automated matching replaces the spreadsheets, lookups and copy-and-paste.
4. Reporting and AI you can trust. Clean, resolved records sit underneath every number.
5. Your data never leaves Snowflake. There are no external APIs, nothing to extract and no new platform to govern.

Who's who? Entity resolution in plain English

Entity resolution is a grand name for working out which records belong to the same business. BusinessLinks does it in three steps.

It narrows the field. Comparing every record with every other one doesn't scale. So it first groups likely candidates together, such as records with similar names or the same postcode.

It scores each match. Clear rules settle the easy cases: same company number, same business. Probabilistic, ML-driven similarity scoring handles the trickier ones, like "Smith & Sons Plumbing" and "Smith and Son Plumbing Ltd" at slightly different addresses.

It verifies and joins. Confident matches become one golden record, checked against Companies House. Uncertain matches are flagged for a person to review. Every decision is kept, so you can always see why two records were linked.

Under the bonnet: for the practitioners

If you're the one who'll be asked "but how does it actually work?", here's what matters:

• It's a 100% Snowflake Native App. It installs from Snowflake Marketplace and runs inside your own account, under your existing security and governance. There are no external APIs and no fragile pipelines to maintain.
• Matching is probabilistic and ML-driven. Candidate blocking keeps comparisons efficient. Deterministic rules settle the exact matches, and ML similarity scoring handles the fuzzy ones.
• Confidence scores and thresholds are configurable. You decide what gets auto-merged and what goes to review, and you can tune both as you learn your data.
• People stay in control. Anyone can manually override a match, and ambiguous edge cases go into review instead of being silently merged.
• Model runs are idempotent and fully recreatable. You can rerun with confidence and reproduce any result, and there's a full audit history behind every link.
• The output is a golden record table in Snowflake, ready to feed reporting, AI workloads, or other systems such as your CRM through a standard Snowflake pipeline.
• The interface is a friendly React app, so business users can review and manage matches without needing Snowflake expertise or a data engineer.
• The matching engine is territory agnostic. BusinessLinks is UK-first, with Companies House enrichment built in. For organisations operating beyond the UK, Kickyr extends resolution to match your global footprint, giving you one resolved view across every market.

What it looks like in practice

For finance and credit teams

The pain: One customer's invoices are spread across three records. Each looks healthy on its own, so nobody sees the full exposure.

With BusinessLinks: Every record rolls up to one verified business, so you see the true total. Companies House status is checked, so you also know when a customer has changed its name or entered administration.

Ask yourself: How much of your debtor book is sitting on records that don't match up?

For procurement

The pain: Supplier lists grow over years, added to by lots of different people. Duplicates hide your true spend and make duplicate payments far easier.

With BusinessLinks: Every supplier is resolved to one verified entity, so you see exactly how much you spend with each business.

Ask yourself: What would knowing your real spend with your top 20 suppliers do for your next negotiation?

For compliance and onboarding

The pain: Know Your Business checks mean a lot of open tabs, a lot of copying and pasting, and a nagging worry that you're already dealing with this company under another name.

With BusinessLinks: Records arrive already matched and enriched with company IDs and director details from Companies House, so much of the groundwork is done before anyone starts.

Ask yourself: What would turning an afternoon of checks into a few minutes mean for your team?

For commercial leaders

The pain: You can't plan around your biggest accounts if you can't see how big they really are.

With BusinessLinks: Each business has one golden record, so reporting shows the true size of every relationship. That record can feed your CRM too, so sales sees the same picture finance does.

Ask yourself: Which of your top accounts are bigger than your reports say they are?

For data teams

The pain: Somebody built the matching logic years ago. Now somebody else maintains it, fixes it and gets asked why it's wrong.

With BusinessLinks: You get configurable rules, confidence scores, manual overrides and recreatable runs with full history, all inside Snowflake. You set the logic once and tune it as you go.

Ask yourself: What would you build with the hours you currently spend maintaining match rules?

Why BusinessLinks

• It's Snowflake-native. Your data stays exactly where it is, under your existing security and governance.
• It's verified against official records. Every resolved match is enriched with authoritative Companies House data.
• It's transparent. Every match comes with a confidence score and a full history, so people trust the result and actually use it.
• It's built around your business. You can tune the rules and thresholds to your own logic and data quality.
• It's ready to go global. The matching engine is territory agnostic, and Kickyr takes it beyond the UK.
• It's quick to start. Install it from Snowflake Marketplace and switch it on, with no data engineering project required.

It's built by Lumilinks, a Snowflake Premier Partner and winner of the 2025 Snowflake Startup Challenge, with a track record of getting clients to value fast.

Knowing who you actually do business with sounds simple. Once you do, everything built on top of it gets easier.