Use case 01
Close variance review
Where does actual spend differ from the approved monthly plan?
Compare ledger exports to operating data before close.
The AI data workspace
Your data speaks
we translate.
Stop waiting on dashboards. Your AI data team explores every source, answers every question, and moves faster than any queue.
Kole / researchReady
Question
Which segments drove net revenue this quarter?
Recommended questions
01 / Compare net revenue by segment
02 / Find the largest quarter-over-quarter change
03 / Check the contributing accounts
finance.ordersaccountsQ3 close
Kole / revenue reviewDetermined output
K
Enterprise accounts drove the increase, led by expansion revenue.
Revenue is up 18.4% from the prior quarter. Growth contributed the second-largest increase.
Net revenueUSD / Q3
Enterprise482k
Growth242k
Self-serve96k
quantum-lab / sqlwarehouse-prod
03 / SQL block38 ms
WITH q AS (
SELECT segment,
date_trunc('quarter', booked_at) qtr,
SUM(net_revenue) rev,
LAG(SUM(net_revenue)) OVER (
PARTITION BY segment ORDER BY 2) prev
FROM finance.orders
WHERE status = 'posted'
GROUP BY 1, 2
)
SELECT segment, rev,
ROUND(100.0*(rev-prev)/NULLIF(prev,0),1) qoq
FROM q ORDER BY rev DESCAction needs approval1 of 1
Save this query for the revenue review.
Kole proposes a saved query. Review the name and SQL before it is written.
ApproveReject
30%
Less time resolving complex operational issues at Farmako
Read
Previews can run without an approval pause
Review
Consequential agent work waits for a human
Record
Keep the notebook, query, and result together
Farmako result reported by Farmako leadership. Product-control statements describe the current interface, not an availability or compliance claim.
How it works
Work the way you do.
System layers / scroll to inspect
Understands your data and business.
Data sources become business context for human-approved agents inside the SyneHQ platform.
- 01
Data sources
PostgreSQL, Snowflake, Databricks, Sheets, and ClickHouse bring warehouses, files, and operational records into one governed workspace.
- 02
Context
Business logic, metadata, and configured semantic models make the data interpretable.
- 03
Agents
Kole and connected workflows propose decisions; people approve consequential actions before they run.
- 04
SyneHQ
SyneHQ keeps governed data, business context, human-approved agents, and reusable results together.
01 / Data sources
02 / Context
03 / Agents
04 / SyneHQ
Exploded system view
Roles
Stay ahead of every decision.
Analyst, planner, or engineer: SyneHQ is the governed workspace where Kole proposes work and people approve what matters.
The analyst
Ask the warehouse in natural language. Generate the report. Share the notebook, query, and result with the team.
The planner
Keep the question, the query, and the written recommendation together so a planning review starts from evidence. A saved query carries its runs, variables, and a captured snapshot; the markdown note stays attached, so the number you take into a meeting still has its source. Kole can propose the next cut. You decide what gets kept.
The engineer
Consequential SQL, schema changes, and agent writes wait for a person. Read-only previews run without a pause; anything that changes a table, a ticket, or a saved object waits on an approval card you can edit. Schema work stays a proposal until a responsible person advances it, with team scope and an inspectable trail.
Use cases
Start with the question your team already has.
Each workflow pairs a relevant analyst role with the sources needed to inspect the answer. Kole coordinates the work and keeps consequential steps ready for review.
Use case 02
Cash collection watch
Which overdue invoices now need a collection owner?
Keep aging, account context, and next steps together.
Dashboards / layout
Put the chart beside the context that explains it.
A dashboard can combine chart and Markdown blocks in a responsive 12-column layout. Blocks can be moved and resized, with the resulting layout retained for the dashboard.
Illustrative dashboard canvas12-column layout
Revenue by week
Chart block
Actual revenue, $kEight-week trendSample values only
Decision context
Markdown block
Why the change matters
Use the release note and owner decision beside the chart so a number has working context.
Review the segment before changing the target.
Plan mix
Chart block
Enterprise24%Growth18%Team15%Starter12%Scale10%Business8%Pilot6%Education4%Other3%
Regional pipeline
Chart block
QualifiedReview (hatched)Eight sample regional totals
Sample blocks and positions - no live dashboard changes
Governed analyst workflows
Trusted AI workflows for every data question.
Let Kole inspect the source, propose the next step, and leave consequential work for your review. The query, decision, and result stay attached to the question.
schemaqueryreviewdecision
Every action has a visible trail
Kole AgentLoading
DATA QUESTION
Which regions need a renewal review this week?
Kole checks saved revenue work
Discovering…
3 steps01 — Connections
warehouse-prod
02 — Schema
orders, accounts
03 — Saved query
weekly expansion
APPROVAL_QUEUE — 1 of 2
run_sql — update revenue staging tableneeds approval
APPROVE · runs the edited proposal
REJECT · returns guidance to Kole
kole (analyst)
I found the weekly expansion query and its latest run. The next step changes the staging table, so it is waiting for your approval.
Inspect the proposal ›
Arguments remain editable
ACTION_TRAIL
RUN: lab_01J6X9
preview_sql — 25 rowscompleted
NEXT_ACTION
WAITING — human decision required
Quantum Lab
SQL for the question. Python for the analysis. Markdown for the decision.
Build a reviewable line from source data to a written recommendation. Queries, calculations, and the conclusion stay together in one lab document.
SQLsource questionPythonanalysisMarkdowndecision
regional-revenue-review — Quantum Lab
SQLwarehouse-prod
WITH q AS (
SELECT
region,
date_trunc('quarter', created_at) AS qtr,
SUM(net_amount) AS revenue,
LAG(SUM(net_amount)) OVER (
PARTITION BY region ORDER BY 2
) AS prev_q
FROM finance.orders
WHERE created_at >= :start_date
AND status = 'posted'
GROUP BY 1, 2
)
SELECT
region, revenue,
ROUND(100.0 * (revenue - prev_q) / NULLIF(prev_q, 0), 1) AS qoq_pct
FROM q
WHERE qtr = date_trunc('quarter', :as_of)
ORDER BY revenue DESC4 rows · 218 ms
Pythonlocal runtime
rev["share"] = rev.revenue / rev.revenue.sum()
rev["qoq"] = (
rev.sort_values("qtr")
.groupby("region").revenue.pct_change()
)
cut = rev.share.sort_values(ascending=False).cumsum()
focus = rev.loc[cut.le(0.61) | (rev.qoq > 0.08)]
focus.assign(
flag=np.select(
[focus.qoq.gt(0.15), focus.share.gt(0.2)],
["expand", "watch"],
default="hold",
)
)[["region", "revenue", "share", "qoq", "flag"]].round(3)Decision note
West and Northeast account for 61% of revenue this quarter.
Prioritize the renewal review for these regions before the planning meeting.
3 blocks · Draft
Determined outputs
Keep the query, the runs, and the snapshot together.
A result needs more than a number. Keep the source query, each run, its variables, and the captured output in the same reviewable record.
01
Saved with the query
Keep the query definition with the output it produced, so the result starts with its source rather than a copied number.
02
Run by run
Review the variables, timing, row count, and result for each run. The latest output is easier to trust when its history is present.
03
Snapshot the result
Capture the output you intend to reference, then return to the same record when the question comes back.
04
Share deliberately
A shared snapshot is a specific captured result, not a moving target behind a dashboard refresh.
saved queryreviewable recordrun historyreviewable recordcaptured snapshotreviewable record
weekly-expansion-pipeline — saved query
Latest snapshot
842 accounts · August 31
AccountSegmentARR
Aster LabsGrowth$48k
NorthstarScale$36k
Cedar HealthGrowth$29k
Captured output — retain the source, run, and snapshot in one record.
A change has a visible state.
change-142.sql
1 ALTER TABLE contracts
2 ADD COLUMN renewal_date date;
3 -- requested by Revenue Ops
4 -- awaiting reviewer approvalPut the SQL, its purpose, and its review state in one ticket. A proposed schema change remains a proposal until a responsible person advances it.
Change ticketReview is part of the path.
01
DraftSQL proposedLoading
ReviewAwaiting approval01
ScheduledReady to deployreview_sql — renewal_date migrationneeds review
review_sql — account flags updateneeds review
schedule — weekly expansion runapproved
The queue distinguishes work awaiting review from work that is ready. It does not replace your deployment controls.
Product controls
Security and trust guardrails built into the work.
Product controls help bound access, review consequential work, and retain a visible record of what changed.
Product controls, not a certification claim.
01
Team-scoped access
Product resources are checked against the active team before a request proceeds.
ACTIVE_TEAMMATCHED
team_8fd2allowed
resource scope confirmed
02
Human approval gates
Supported consequential Kole tools pause for a person to review, edit, approve, or reject.
WRITE REQUESTQUEUED
Save weekly revenue query
ApproveReject
03
Configured masking
Configured column masking applies on supported result paths before data reaches the interface or agent context.
COLUMN POLICYSUPPORTED PATH
emaila•••@company.comcard_last4•••• 4821
04
Inspectable records
Agent activity metadata, explicit sharing controls, and reviewable change tickets leave the work open to inspection.
CHANGE RECORDDB-041
Add index to ordersreview
agent event metadatarecorded
Security status
Current evidence, stated without certification language.
01
Report not claimed
SOC 2
No SOC 2 report is claimed. Product controls and their assessment status are described plainly so teams can review the available evidence.
02
Compliance not claimed
GDPR
GDPR compliance is not claimed. This status does not represent a certification or compliance determination.
03
Documented control
Credential encryption + Infisical
Designated connection credentials are encrypted before storage with a tenant-scoped key outside development. db-api is configured to authenticate with Infisical at startup through Universal Auth for its configured project and environment.
04
Deployment-specific
Service boundaries
Reference deployments route selected traffic through an auth proxy and internal service names. Boundary exposure depends on the deployment: Compose ports may be published. PostgreSQL TLS is configurable, not enforced.
Security statuses are not certifications. Controls shown here apply within configured product paths and connected services.
Loading
Self-hosting is a future direction.
We're exploring a self-hosted deployment option for organizations that need a different operating model. Not available today — details will be published when ready.
Customer proof
Teams trust the work they can inspect.
Reported result
Farmako
30%less time resolving complex operational issues.
Read Farmako's user storyCustomer perspective
Farmako
“Before SyneHQ, we were constantly going back and forth with the data team just to get basic insights. It slowed us down and honestly got frustrating. Now, with SyneHQ, our Ops and field teams get the data they need in seconds—just by asking in plain English. It’s been a total game-changer for us.”
Customer perspective
Thirdbase
“With their robust platform, we've seamlessly integrated multiple data sources, unified analytics, and automated standard reporting, saving us countless hours each week.”
Customer perspective
Mintair
“Within days, it streamlined our workflow. The balance of security and innovation it offers is unlike anything we’ve seen.”
Customer statements reflect each organization's experience. Results vary by team, workflow, and use case.
Launch pricing
Start with one analyst. Keep the work reviewable as the team grows.
Every plan keeps consequential work visible and ready for a human decision.
Core
Solo
A governed starting point for independent analysis.
$9 / mo
- 1 seat
- 3 active data connection
- 2 active Quantum Labs
- 50 approved agent actions / month
- 50 saved queries
- Approval controls on consequential actions
Scale
RecommendedTeam
A shared workspace for analysts who need reviewable work.
$49/ seat / mo
- Shared Quantum Labs and notebooks
- 500 approved agent actions / seat / mo
- Saved queries and determined outputs
- Approval controls for database actions
- Team activity and change history
- Priority onboarding
Command
Company
A governed workspace for the team responsible for data work.
$399/ month
- 5 seats included
- 2,500 approved agent actions / month, pooled
- Shared Quantum Labs and notebooks
- Saved queries and determined outputs
- Approval controls and change history
- Guided setup for your team
Custom
Enterprise
A governed workspace for large teams with custom needs.
Custom
- Custom number of seats
- Custom approved agent actions / month
- Shared Quantum Labs and notebooks
- Saved queries and determined outputs
- Approval controls and change history
- Dedicated support and onboarding
Quantum Lab
Make the work inspectable before it becomes a decision.
Bring analysis, notebook work, and human approval into one governed workspace.
Open Quantum LabLaunch pricing
Launch pricing. Agent-action allowances reset monthly. Taxes may apply.
Approval model
You approve consequential work before it runs. The notebook keeps the evidence.