A short story about a pipeline, a CFO and a Monday morning nobody wants.
It's 8.52am on a Monday. You arrive at the board meeting prepared. The board pack went out last Friday, and the meeting starts in eight minutes.
You've barely swallowed your coffee when, halfway down page three, your CFO stops. "Why doesn't this revenue figure match the one finance sent last week?" A little coffee dribbles down your chin and onto your crisp, freshly dry-cleaned white shirt.
An awkward pause hangs in the air until somebody says, "Let me check that and come back to you." Everyone knows what that means.
By lunchtime, the data team has the answer. A pipeline failed late on Thursday night. Nothing crashed and no alarm went off. The data simply stopped arriving, and the report carried on as if nothing had happened.
Nobody made a mistake. Nobody was watching the pipeline, that's all. And you're far from alone.
The cost of finding out last
74%
of data professionals say business stakeholders spot data issues before the data team does, all or most of the time.
31%
of revenue, on average, is exposed to data quality issues.
2 days a week
is roughly what data teams spend checking data quality: about 40% of their working time.
60%
of AI projects without AI-ready data will be abandoned through 2026, Gartner predicts.
Sources: Monte Carlo / Wakefield Research, 2023; Monte Carlo, 2022; Gartner, 2025.
Sentinel changes that. It watches your Snowflake platform continuously and tells the right person when something needs attention, before it reaches a report, a decision or a board meeting. It's built natively inside Snowflake, so there's no data movement and nothing new to run.
Want to see what Sentinel would catch on your platform?
Book a 20-minute walkthrough on your own dbt project, or install Sentinel straight from Snowflake Marketplace.
Sentinel watches two layers of your platform, so nothing slips through the gap between them.
The Floor: is the data arriving?
This is the plumbing everything else depends on. Are pipelines running, is data arriving on time, and are refreshes completing?
The Ceiling: can you trust it?
This is the layer your decisions sit on. Is the data complete, accurate and fit for purpose? Data can arrive perfectly on time and still be wrong, so Sentinel checks both.
When something needs attention, Sentinel tells the right person in Teams, Slack or email, with the lineage to trace it straight back to its source.
If you're the one who'll be asked "but how does it actually work?", here's what matters:
The pain: You've invested heavily in a modern data platform, and it only takes one wrong number in the wrong room for people to doubt all of it.
With Sentinel: Issues are caught and flagged early, and you get one clear view of platform health instead of a report after the fact. When something does go wrong, you hear it from your own team, with a fix already under way.
The numbers: Gartner puts the average cost of poor data quality at $12.9 million a year per organisation.
Ask yourself: When did a stakeholder last find a data problem before your team did?
The pain: Every morning starts with checking pipelines and dashboards by hand, just in case. It's tiring, it's invisible and it never really ends.
With Sentinel: Continuous monitoring of freshness, schema, volume and quality across your dbt project, with alerts only when something actually needs you, and lineage that takes you straight to the failing model.
The numbers: Monte Carlo's 2022 survey found data professionals spend 40% of their time checking data quality. The average organisation deals with around 61 data incidents a month, each taking about 13 hours to identify and resolve.
Ask yourself: What would your team build with two days a week back?
The pain: A dashboard looks fine on the surface, but it's built on data that stopped refreshing two days ago, and you're the one who gets the awkward email.
With Sentinel: Freshness and quality checks sit underneath every report, and lineage shows exactly which dashboards are affected when something upstream breaks.
The numbers: In a 2023 survey, more than half of data professionals said 25% or more of their revenue was exposed to data quality issues, with the average at 31%.
Ask yourself: How confident are you that every dashboard opened this morning is showing today's data?
The pain: You have dozens of dashboards, models and analyses, and no easy way to say which ones the business can trust today. Ownership is fuzzy, so alerts land in a shared inbox and nobody picks them up.
With Sentinel: Every data product carries an automated Trust Score based on the health of everything upstream. The Tagging Wizard assigns domains, medallion levels and severity tiers in a few clicks, so every model has an owner and every alert has a home.
The research: Gartner warns that organisations with basic or manual metadata management will struggle to make their data AI-ready.
Ask yourself: Which of your dashboards could you put a trust score on today?
The pain: A model will act on bad data with complete confidence. It won't stop to ask whether a number looks odd.
With Sentinel: A live view of whether the data feeding your models is complete, fresh and fit for purpose, so you can scale AI on foundations you trust.
The numbers: Gartner predicts that through 2026, organisations will abandon 60% of AI projects that aren't supported by AI-ready data. It names data observability as one of the practices that gets data ready.
Ask yourself: Would you know if the data behind your AI changed overnight?
It's built by Lumilinks, a Snowflake Elite Partner and winner of the 2025 Snowflake Startup Challenge, with a proven track record of getting clients to value up to 5x faster.
It's Thursday at 11.14pm and the pipeline fails. Within minutes, an alert lands in the data team's Slack channel, showing which pipeline failed and which reports depend on it. On Friday morning, an engineer fixes it over their first coffee, and the board pack goes out on time with the right numbers.
On Monday, you arrive at the meeting with two cortados in hand: one for you, and one for the CFO, who reads page three, nods and moves on to page four.
Nobody ever knows how close it came. That's exactly how it should feel.
See Sentinel on your own dbt project in 20 minutes.
We'll show you what it would catch on your Snowflake platform, how lineage traces it to source, and what that means for your reporting, your team and your AI plans.
Book your 20-minute walkthrough →
Prefer to explore first?
Get Sentinel on Snowflake Marketplace →
Want help setting up alert routes or tagging your dbt project? Just tell us on the call and we'll take it from there.
Sentinel, from Lumilinks. Because practical makes powerful.