Can sales and finance get answers from Snowflake without writing SQL?
Why it's hard
The answer is three joins away, and the only people who can write them already have a backlog. A simple question waits days.
Snowflake + Querri
Snowflake analytics,for everyone who doesn't write SQL.
Your warehouse already holds the answers. Querri lets sales, finance and operations ask for them in plain language, with your team's definitions applied, while the data team gets out of the ticket queue.
Querri only ever reads from Snowflake, through a read-only role and a key it never shows anyone. SOC 2 Type II, HIPAA and ISO 27001:2022 certified.
Illustrative data
Where Snowflake reporting stalls
The data is in the warehouse. The people who need it aren't.
Snowflake does its job well. The trouble starts between the warehouse and the person with the question, and most of it is about people, not compute.
01
SQL is the only way in
Snowsight is a SQL editor. Anyone who can't write a join files a request, and the data team spends its week on one-off pulls instead of the models it was hired to build.
02
The meaning lives somewhere else
What counts as an active customer or a net sale sits in dbt models, a semantic layer or one analyst's memory. Two queries over the same table give two different numbers.
03
Exploring costs credits
A warehouse bills per second with a 60-second minimum each time it starts, and it runs until it suspends. A day of ad-hoc questions keeps it awake for work nobody planned.
04
Every viewer needs a BI tool, and a builder
Dashboards on top of the warehouse mean BI licenses for everyone who looks, plus someone who can model the data before a single chart appears.
05
It all ends in a spreadsheet
Finance and ops copy query results into Excel, add their own columns and send them around. By the second week, nobody knows which copy matches the warehouse.
What you can finally answer
Eight Snowflake questions the business stops waiting on
Each is a request data teams field every week, why it stalls today, and what Querri does with it.
Which accounts renewing this quarter have stopped using the product?
Why it's hard
Product events sit in Snowflake and renewal dates sit in the CRM. Lining them up means an export from each and a match on account IDs that rarely agree.
With Querri
Connect Salesforce next to Snowflake and Querri matches accounts across both, saying how it matched. The at-risk renewals playbook turns it into a weekly list.
Can everyone use the same definition of an active customer?
Why it's hard
The definition is buried in a dbt model or a LookML file. Business users never see it, so their spreadsheet version quietly differs.
With Querri
Say it once in plain words in Facts & rules, or define it as a KPI. Every chat answer, dashboard and report after that uses the same meaning.
How long does it take a new signup to reach first value?
Why it's hard
The event tables are huge and the funnel steps are argued over. Each time someone asks, the query gets rewritten from scratch.
With Querri
Build the funnel once as a project on your synced event tables and rerun it whenever you like. The signup to first value playbook shows the steps.
Which marketing channels bring customers who keep paying?
Why it's hard
Campaign data lives in HubSpot, revenue in the warehouse. The join is an email address or a company name that's spelled differently in each.
With Querri
Connect HubSpot beside Snowflake and Querri links contacts to accounts, with a caveat when a match is uncertain. The steps are in the channel to revenue playbook.
Which metric moved overnight, and where?
Why it's hard
Nobody runs every metric against its own history every morning. A drop in one region shows up in the monthly review, weeks late.
With Querri
Nightly insights checks each KPI against its own history with arithmetic every night, and leaves what's worth a look in your Inbox. Nothing changes without your OK.
Can each client or region see a dashboard of only their own rows?
Why it's hard
Doing it in a BI tool means row-level security rules per viewer, kept in step with the warehouse by someone technical.
With Querri
Ask for the dashboard, then set a row filter per person once in Workspaces & access. It follows them everywhere, including an embedded dashboard in your own app.
Can I ask Claude or ChatGPT about our Snowflake data?
Why it's hard
Hand an assistant a schema and it guesses at joins and definitions. It can sound sure and still be wrong.
With Querri
Add Querri to Claude or ChatGPT and the assistant answers from your Library, with your rules and each person's access in Querri, and says which reading it used when a question is ambiguous.
Snowflake plus the rest
Not everything made it into the warehouse
Usage and events are in Snowflake. Renewal dates, invoices, campaigns and the planning spreadsheet often aren't. Querri keeps them together in one Library, so nobody waits for a new pipeline to answer a cross-system question.
Snowflake pairs well with
Illustrative data
Answers you can stand behind
"Active" can mean two things. The answer names the one it counted
"Active customers in September" can mean accounts that logged in during the month or accounts paying on the last day of it. Each is a fair reading, and the two counts differ by hundreds. Querri states which reading it used and puts the other count beside it, so nobody has to reverse-engineer the SQL to know what the number means. Read how Querri checks its answers.
Illustrative data
Wake up to what changed
Querri reads the overnight numbers before you do
Each night Querri measures every KPI against its own past and looks for the questions nobody thought to ask. Whatever it finds is sitting in your Inbox when you log in, and it changes nothing until you say OK.
Illustrative data
Custom Snowflake dashboards
Snowflake dashboards the business can change without a ticket
Describe what you want to see, or start from a screenshot of a dashboard you already use. After that, the person looking at it can change it by asking, every chart draws on the same Library as chat, and each viewer can be limited to their own rows.
- Ask for a chart, ask again to change it
- Warehouse tables next to CRM, finance and spreadsheet data
- A row filter per viewer, so a region sees its region
Recurring requests, retired
Six Snowflake reports that rerun themselves
Weekly metrics pack
The numbers leadership asks for every Monday, rerun on the latest sync and emailed.
See reportsSignup to first value
How long new accounts take to reach the moment that predicts they'll stay.
Read the playbookCustomer health score
Usage, support and billing signals combined into one score per account.
Read the playbookAt-risk renewals
Accounts near renewal whose product usage has dropped.
Read the playbookChannel to revenue
Which acquisition channels turn into revenue that stays.
Read the playbookMonth-end finance workbook
Warehouse figures delivered as a formatted Excel file, not a CSV someone has to fix.
See Excel toolsWhat comes over
The tables you choose, copied and kept current
Querri copies the tables and views you pick into its own storage on a schedule and answers questions from that copy, not live against Snowflake. It can only sync what the connecting role is granted.
What you can sync
- Base tables
- Views
- Every schema in the database
- A chosen subset of tables
- Shared databases
- Custom SQL queries
- Each table arrives as SCHEMA.TABLE; row counts in the picker come from metadata, so nothing is scanned.
Full or incremental
Full refresh by default. Pick a cursor column, such as an updated-at timestamp, and incremental syncs append only rows newer than the last run, as often as hourly.
Built for big tables
Rows stream in batches into Parquet, so no table is held in memory at once. There's no fixed row cap; one test sync pulled more than 520 million rows.
Read only
Querri runs only reads, under a read-only role, tagged querri-connector in QUERY_HISTORY. It never creates or changes a table. Syncs use your warehouse's credits.
No password involved
Key pair by default: Querri generates the key, stores the private half encrypted and never displays it. A programmatic access token also works.
How to connect Snowflake
One setup script, run once, and the agent does the rest
- 1
Ask the Connector Agent
Open Connect and say "connect our Snowflake warehouse," then enter the account identifier, a username and the warehouse to use.
- 2
Generate the key and script
Click Generate key pair & SQL. Querri creates the key, keeps the private part encrypted and writes a ready-to-run script.
- 3
Run it once in Snowsight
An ACCOUNTADMIN runs the script with Run All. It creates a service user, a read-only role and its grants, plus a warehouse if you need one.
- 4
Pick tables and approve
Take every table or choose a few, set an incremental cursor where it helps, and pick a schedule. No data moves until you click Approve.
- 5
Ask your first question
Once the first sync finishes, your Snowflake tables sit in the Library and anyone with access can ask about them, chart them or schedule them.
Snowflake FAQ
Snowflake connector questions, answered
How do I connect Snowflake to Querri?
Ask the Connector Agent to connect Snowflake and enter your account identifier, a username and a warehouse. Click Generate key pair & SQL, and have an ACCOUNTADMIN run the script once in Snowsight. It creates a service user and a read-only role. Then pick tables, approve the load and choose a schedule.
What Snowflake data does Querri sync?
Every table and view in the database you choose, across all its schemas except INFORMATION_SCHEMA, or just the tables you select. Shared databases work too, and you can add your own SQL queries as extra sources. Each table arrives named SCHEMA.TABLE. Querri can only sync what the connecting role is granted.
How often does Snowflake data sync?
Full refreshes can be scheduled daily, weekly or monthly. If you pick a cursor column, such as an updated-at timestamp, incremental syncs can run hourly, daily or weekly and add only rows newer than the last run. A manual sync can run at any time between schedules.
Does Querri write to or change my Snowflake data?
No. Querri only reads from Snowflake and never creates, changes or deletes a table or row. The one-time setup script, which you run yourself, creates only a service user, a read-only role, its grants and, if you ask, a warehouse. Querri's queries are tagged querri-connector in QUERY_HISTORY, so you can audit them.
How do Snowflake roles and permissions carry over to Querri?
Querri sees only what the role you connect with is granted, so give that role exactly the databases and schemas you want in Querri. Inside Querri, who sees which rows is set by Querri's sharing and per-person row filters, which follow each person into chat, dashboards, Claude and the API.
Does Querri query Snowflake live when someone asks a question?
No. Querri copies the tables you choose into its own storage on a schedule and answers from that copy. Questions don't wake your warehouse; only syncs run queries in Snowflake, and those use your warehouse's credits. The data is as fresh as the last sync.
How do I combine Snowflake and Salesforce data?
Connect Salesforce next to Snowflake; both land in the same Library. Querri matches warehouse accounts to Salesforce accounts, ideally on an ID stored in both, and tells you how it matched. You can then ask about usage next to renewals, pipeline or open cases without building a new pipeline.
How do I combine Snowflake and HubSpot data?
Connect HubSpot alongside Snowflake. Querri links HubSpot contacts and companies to the customers in your warehouse, so you can follow a campaign through to revenue that stays. When a match might be wrong, such as a different email domain, the answer flags it.
How do I get a customer health score from Snowflake data?
Sync the usage tables from Snowflake and connect support and billing sources if they live elsewhere. Define the score once in plain words and Querri tracks it per account, with history from the day you start. The customer health score playbook walks through each step.
How do I get a signup-to-first-value report from Snowflake?
Sync your event and account tables, say which event counts as first value, and build the funnel once as a project. Rerun it on every sync or schedule it to email the team. The signup to first value playbook covers it end to end.
Can I build custom dashboards on Snowflake without a BI tool?
Yes. Ask for the charts in plain words, or begin with a screenshot of a dashboard your team already relies on. Edits are requests too. One dashboard can put Snowflake tables beside CRM or finance data, restrict each viewer to their rows, and sit inside your own app as an embed.
Can I ask ChatGPT or Claude about my Snowflake data?
Yes. Add Querri as a connector in Claude or ChatGPT. Questions about Snowflake tables are then answered from your Library, with your team's rules applied and limited to the rows that person can see in Querri. Instead of guessing joins from a raw schema, it uses the same definitions your dashboards and reports use.
How do I export Snowflake data to Excel automatically?
Save the analysis as a project and schedule it. On each run Querri works from the latest synced data and emails a formatted Excel workbook to the people who need it, with a separate version per person or per client.
Can I connect to Snowflake with a password, SSO or OAuth?
No. Snowflake blocks passwords for service users and requires a second factor for people, so Querri uses key pair or a programmatic access token. OAuth isn't offered because it needs a person to approve in a browser and its tokens expire, which would break scheduled syncs.
What are the limits of the Snowflake connector?
Incremental syncs only catch rows whose cursor value goes up, so deletes and edits that don't change the cursor need a full refresh. Custom SQL sources always fully refresh. There's no PrivateLink or VPN option. Setup needs an ACCOUNTADMIN once, and access tokens expire, so key pair suits schedules.
Can I limit Querri's access to Snowflake by IP address?
Yes. Add Querri's outbound IP addresses to a Snowflake network policy for the service user. Querri reaches your account over Snowflake's public endpoint, in any cloud or region. PrivateLink isn't supported, so an IP allowlist plus the read-only role is how you lock it down.
Does Querri handle large Snowflake tables?
Yes. Data streams in batches into Parquet, so a whole table never sits in memory, and there's no fixed row cap. One test sync pulled more than 520 million rows. For big, growing tables, an incremental cursor keeps each sync to the new rows.
Is my Snowflake data secure in Querri?
Querri is SOC 2 Type II, HIPAA and ISO 27001:2022 certified, and your data is never used to train a model. The private key Querri generates is stored encrypted and never shown, the connecting role is read-only, and the Connector Agent never sees the secret itself.
Keep exploring
Where to go after you connect Snowflake
Let the rest of the company ask the warehouse.
Run one setup script and connect in minutes. Querri reads from Snowflake and never changes a table or a row.