Decisions, backed by evidence — not guesses.
SignalLense is context-aware decision intelligence. It turns raw business data into deterministic findings, ranked explanations, and real-world context — so every "why did this change?" gets an answer you can trust and trace.
Connect SignalLense to the tools your team already uses.
- Tableau
- Power BI
- Raw CSV
- SnowflakeSoon
- Amazon RedshiftSoon
- PostgreSQLSoon
- In-app
- SlackSoon
Dashboards tell you a number moved. They stop exactly where the real question begins.
The investigative work of connecting a business change to its likely causes is where the effort actually is — and where every conventional BI tool leaves you opening five tabs by hand. SignalLense makes that step the product.
"What happened?"
- Shows the number and the chart
- The "why" is left to a manual, ad-hoc correlation in spreadsheets and Slack
- No real-world context, no evidence trail, rarely repeatable
"What happened → why → what to decide."
- Detects meaningful change automatically, with a baseline comparison
- Ranks explanations, each graded by the strength of its evidence
- Aligns internal events and external context to the exact window — sourced
Detect
A tracked metric moves. SignalLense flags a meaningful change from ordinary variation — before you have to notice.
Investigate
One click opens a structured investigation: quantified change, segment breakdown, and history.
Understand
Ranked explanations, each with an evidentiary status and the exact evidence behind it.
Decide
Record the decision it informed, then track whether it actually achieved its intended outcome.
See why — not just what.
Every investigation opens with the change quantified against its own history, then lays out the evidence and real-world context before it ever shows a conclusion — so you build your own read of the facts first.
- What changed, computedObserved value, trailing-average baseline, percent change and z-score — all deterministic.
- Actual vs baseline, in viewThe investigated period highlighted against its baseline band — no chart re-computes analytics.
- Relevant real-world contextExternal events aligned to the same window and place, each dated and sourced.
A disciplined pipeline — deterministic where it computes, AI only where it reasons.
The separation is structural, not a promise in a prompt: the two AI stages propose and narrate, every other stage is deterministic code. That boundary is what makes the answer trustworthy enough to act on.
1 · Detect & scope
A statistically significant change is detected on a tracked metric, or you select a metric, segment and window by hand. Segment analysis and historical comparison run first — all deterministic.
Deterministic2 · Gather evidence & context
Internal business events and external context events are retrieved through one shared interface — dated, scoped things that might explain the change — then scored for relevance to this exact window.
Deterministic3 · Rank & grade
AI proposes candidate explanations; deterministic code then links each to concrete evidence, scores confidence, and assigns an evidentiary status from a fixed vocabulary. "Insufficient evidence" is a first-class, honest answer.
AI proposes · evidence decides4 · Decide & revisit
Investigations are persisted and shareable, never silently recomputed. Record the decision they informed, and follow its outcome over time in a decision history.
Persisted & auditableEverything an investigation needs — in one system.
Not a chart builder with an "explain" button bolted on. Each surface exists to move a decision forward.
Automatic change detection
Versioned metrics are watched continuously and labeled ordinary variation vs. meaningful change — so you notice what matters without writing a query.
Evidence-graded investigations
Ranked explanations, each carrying a confidence band and an evidentiary-status label — expandable down to the exact source behind every claim.
Context engine with provenance
Holidays, news and other real-world events are aligned to the metric's time and place, deduplicated, and scored — every item stored with its source and retrieval time.
Cross-platform comparison
Ask whether Power BI's Revenue equals Tableau's — a deterministic, auditable answer to "do these platforms agree, and if not, exactly where."
Threshold & anomaly alerts
Threshold, percent-change and anomaly alerts with in-app and email notifications — so a change reaches you before it becomes a fire.
Decision history
Decisions recorded from metrics, investigations, alerts and comparisons — with owner, status and whether each achieved its intended outcome.
AI that explains — and never computes.
Ask questions in plain language and get answers grounded in your own Metrics, Investigations, Comparisons and Alerts — not a generic chatbot guessing at numbers.
- Grounded, not generative: answers are built from already-computed results, so figures stay traceable.
- A structural boundary: AI only proposes hypotheses and writes narrative for an already-scored explanation.
- Honest by design: when the evidence is thin, it says so — no unqualified causal claims.
Modern BI answers "what changed." SignalLense answers "why" — with the evidence attached.
Tableau Pulse and Power BI Copilot already detect anomalies, attribute drivers, and answer questions in plain language. SignalLense starts where that ends — assembling external context, your own business events, and graded evidence into an investigation you can reopen and defend.
| Capability | BI dashboards | BI-native AI | LLM chat on data | SignalLense |
|---|---|---|---|---|
| Dashboards & reporting | BI dashboards Core strength | BI-native AI Built in | LLM chat on data Not the focus | SignalLense Not a dashboard tool |
| Natural-language questions | BI dashboards Basic Q&A | BI-native AI Yes, with citations | LLM chat on data Yes | SignalLense Yes — Ask AI |
| Anomaly detection & driver attribution | BI dashboards Built in | BI-native AI Built in | LLM chat on data Varies | SignalLense Built in |
| External context (holidays, news) tied to the change | BI dashboards Custom build | BI-native AI Custom build | LLM chat on data Unsourced recall | SignalLense Built in, source-linked |
| Your business events as evidence | BI dashboards Manual annotation | BI-native AI Not modeled | LLM chat on data Not modeled | SignalLense First-class evidence |
| Evidence grading & hypothesis handling | BI dashboards Not modeled | BI-native AI Facts, not graded | LLM chat on data Confident prose | SignalLense Graded, incl. "insufficient" |
| Reproducible, re-openable investigation | BI dashboards Saved views | BI-native AI Conversational | LLM chat on data Non-deterministic | SignalLense Persisted with its inputs |
| Spans multiple BI platforms | BI dashboards Single platform | BI-native AI Vendor's own stack | LLM chat on data Varies | SignalLense Power BI + Tableau |
Competitor capabilities reflect vendor documentation for Tableau Pulse and Power BI Copilot as of September 2026. SignalLense is designed to run alongside your BI platform, not replace it.
For the people who currently investigate by hand.
Defensible answers for the weekly review
Explanations you can put in front of leadership without doing the correlation work by hand — each one graded by the evidence supporting it.
Know why an operational metric moved
Attrition, leave patterns, ticket volume, fulfillment delays — understood before they become a fire, with context aligned to the moment they shifted.
The answer and its confidence
Sales, CS and Marketing leaders get the ranked explanation and its confidence band — with drill-down available the moment they want to challenge it.
Investigation without an analyst
At companies too small for a dedicated analyst, the investigation mostly happens for you — far less manual effort than five tabs and a spreadsheet.
Start free. Add AI, context & your team when you're ready.
The deterministic core is yours on every plan. Pro adds AI-assisted interpretation, external context, and real team collaboration.
Core individual deterministic decision intelligence.
- Deterministic Metrics, Investigations & Comparisons
- Threshold, percent-change & anomaly Alerts
- In-app and email notifications
- Public sharing via share links
AI + external context + real team collaboration.
- Contextual "Ask AI" — grounded in your data
- AI-assisted explanations for investigations
- Historical & recent news context
- Organization & team membership with roles
- Comments, threads & @mentions
Up to 25 team members per organization.
Talk to the team behind the evidence.
Questions about investigations, connectors, or rolling SignalLense out to your team? Send a note and we'll get back to you.
-
Email contact@signallenses.com
-
WhatsApp +91 88060 90033
Your message is sent securely and never shared.
Stop guessing. Start investigating.
Turn the next unexplained metric move into a defensible, evidence-backed answer — or an honest "not enough evidence yet." Both beat a guess.
Free to start · No credit card required to explore the deterministic core