Ask Your Business
Natural-Language Business Intelligence
Ask “what were margins by region last quarter?” and get the answer, with its source, in seconds — instead of filing a ticket and waiting weeks for a dashboard.
You bought Tableau, Looker, or Power BI. The exec team still asks the analyst. This puts the answer engine right in Slack — with the finance team's own definitions baked in.
Potential value on day one
+$234,000/yr freed
Example — 5 people × $75/hr × 15 hrs × 52 wks × 80% straight-through processing.
Model this with your own headcount below ↓
The value math
Before, after, and what it's worth.
On the left is what it costs to keep doing this by hand. On the right is what happens after the package goes live — and what that frees up on day one.
Today
Today · Analyst-gated reporting
How business questions get answered today
- 1
A business user has a question — 'what were margins by region last quarter?'
The question comes up in a meeting, an email, an exec ask. The answer is somewhere in the warehouse.
- 2$
They file a ticket with the BI team — and wait
The question joins a queue of 30 other requests. The BI team triages, estimates, schedules.
- 3$$$
An analyst spends hours writing SQL and building the report
SQL against the warehouse, joins across three tables, edge cases for currency conversion. Then chart-building on top.
- 4$
The report comes back stale — the question has already moved on
By the time the dashboard lands, the exec has already made the call from a napkin estimate.
- 5
Decision made from partial or delayed data
Or the report gets shelved because it no longer matters. Analyst work, wasted.
Human in the loop
A team of analysts, full-time, servicing an inbox of one-off questions.
The hidden cost — decision drag
Reports pulled at different times don't agree. Metrics are defined differently across departments — 'revenue' means net for finance, gross for sales, booked for ops. Decisions get made from stale numbers because the question moves faster than the analyst queue. Confidence in the numbers slowly erodes, and shadow spreadsheets multiply.
Yearly expense — today
−$292,500/yr
100% payroll burn on manual data entry.
5 people × $58,500 fully-loaded per seat.
After DEBCOR AI
With DEBCOR AI · Ask your business
How business questions get answered after the package ships
- 1
The business user asks in plain English — in Slack, Teams, or web
'Margins by region last quarter?' — typed the way they'd say it. No SQL, no ticket, no wait.
- 2
AI translates to SQL against your locked semantic layer
The finance-approved definitions of every metric and dimension gate the query. Bad SQL never runs.
- 3
Answer + chart returned in seconds, sourced from live data
The number, the trend line, and a one-paragraph read-out — with a link to the underlying SQL for audit.
- 4$
Ambiguous questions route to the BI team — but only once
When intent isn't clear, the BI team clarifies. That clarification becomes a promoted definition — every future asker gets it right instantly.
- 5
Every promoted definition compounds — the system gets smarter
The semantic layer grows. Coverage climbs. The BI team's role shifts from ticket-servicer to curator of a living knowledge asset.
Human in the loop
One BI lead, part-time, curating definitions — not answering the same question twice.
Metric drift disappears
One semantic layer, one source of truth. Every metric definition is version-controlled. Every answer cites its source. Reports agree across departments because everyone's asking against the same locked definitions. Decisions get made from live data, not week-old snapshots.
Yearly savings — with DEBCOR AI
+$234,000/yr
80% of your $292,500/yr labor pool redirected to real work.
Range: $204,750–$263,250/yr at 70–90% STP.
Directional example, not a quote. Fully-loaded rate assumes $75/hr — swap in your own if it's different. Freed capital = payroll cost the process stops burning, before package fees. Actual value depends on document mix and your straight-through target, set in the scope phase.
What's inside
Everything the package ships with.
A data warehouse layer over your existing sources — Supabase, BigQuery, Snowflake, or DEBCOR-hosted
A semantic layer workshop where finance locks the metric and dimension definitions — one source of truth, forever
A text-to-SQL agent with hallucination guardrails — Claude writes the SQL, DEBCOR validates it before it runs
Answer + chart + narrative — the exec gets the number and a one-paragraph read-out
Delivery in Slack, Teams, web, email digest, and iOS shortcut — wherever your team already asks
Row-level security honoring your IdP + a full audit log of every question and answer
Sizing & published pricing
Three sizes. One architecture.
Scope compresses down, not features. S-tier is playbook-driven — a single senior engineer delivers against pre-built accelerators inside a fixed configuration boundary. M-tier extends the same systems into landscapes with multiple sources, custom rules, and real integration complexity. L-tier is the full enterprise engagement: dedicated environments, SOC 2 wrap, roadmap co-development, and — for SAP customers — clean-core BTP integration. Same architecture at every size.
Band
≤ 500 employees · one system, one team
Timeline
3–5 weeks
Published price
from $15k
Scope at this tier
One domain (finance OR ops OR sales), locked metrics, Slack + web delivery — a single senior engineer against pre-built accelerators.
Band
500–5,000 employees · multiple sources or rules
Timeline
6–8 weeks
Published price
$55–100k
Scope at this tier
3–5 domains, custom semantic layer, richer visualization, integration with existing dashboards.
Band
5,000+ employees · enterprise scope, SOC 2, roadmap co-development
Timeline
8–12 weeks
Published price
$180–325k
Scope at this tier
Enterprise scope, SOC 2 wrap, iOS/watch surfaces, exec digest email, and roadmap co-development.
Published bands are directional. Final commercial terms are per-client SOW. Managed operations are quoted separately.
Timeline
6–8 weeks from scope to live.
Every package follows the same four-phase delivery. What changes at each tier is the count of sources, rules, and integration surface — not the shape of the work.
Scope
Weeks 1–2
Semantic layer workshop — finance locks metric and dimension definitions.
Build
Weeks 3–4
Deploy the agent in Slack + web. Wire the semantic layer. Golden-set validation.
Validate
Week 5
Parallel run — CFO or lead analyst validates 50 questions before general rollout.
Live
Weeks 6–8
Rollout to exec team. Measure adoption. Add sources and metrics as demand shows.
For SAP customers · Bridge to SAP-native
For SAP-native reporting the answer is Joule + SAP Analytics Cloud. This package is for the questions your business asks across systems — SAP + your CRM + your warehouse. Same semantic-layer approach; same governance principles carry into SAC when you extend. See how DEBCOR extends this into SAP →
Package FAQ
Common questions.
No — it sits alongside them. Your dashboards keep running for canonical reporting. This handles the ad-hoc questions dashboards can't anticipate. Many customers use both: dashboards for the known KPIs, NL for the follow-up questions.
Every generated SQL query is validated against your semantic layer before it runs — schema, joins, filters, and aggregation rules must match approved definitions. Queries that fail validation are blocked, not silently truncated. And every answer cites the underlying query so finance can spot-check.
The agent runs queries under the asker's identity, honoring your existing IdP. If a regional VP can't see US-East data in the warehouse, they can't ask the agent for it either. Answers respect permissions end-to-end.
Per-question inference cost is small through a tiered model architecture. A frontier reasoning model runs during scoping to lock the semantic layer and validate the golden query set — a one-time cost. In production, most questions are structurally similar to prior questions and resolve through cached query patterns with no LLM call. When a novel question needs SQL synthesis, a high-efficiency low-cost model handles it at fractions of a cent. The frontier model is only re-engaged for genuinely novel analytical questions — rare after the first weeks. At production volumes, AI cost is a rounding error against the analyst time it saves.
Yes. We connect to Supabase, BigQuery, Snowflake, Redshift, and Postgres today. If you're on something else, we can adapt — the semantic layer sits above whichever warehouse you're on.
Scope this package
Fixed scope, fixed price. Tell us the document types and the systems, and we'll come back with a one-page proposal.
Start scopingNot sure this is the one?
Book a 30-minute fit call with a senior architect. We'll tell you if this is the right package, a different one, or nothing at all.
Book a 30-minute fit callOther packages
Also worth a look.
AI That Runs a Process
Custom AI Agent Development
A digital employee for exactly one job — matching invoices to purchase orders, working AR disputes, or triaging support tickets — all day, every action logged, a human approving anything unusual.
Kill Manual Data Entry
Intelligent Document & Transaction Ingestion
Your invoices, orders, and forms arrive as PDFs, EDI files, and emails — and today, someone retypes them. The system reads, validates, and posts them automatically; your people only touch the exceptions.
Trust Your Data
Data Foundation for AI
The same customer entered five different ways across your systems is why your reports disagree and your AI guesses wrong. We merge everything into one clean version of the truth — and keep it clean automatically.