ASK • CITE • ACT
AskEdi works from governed metadata, not raw data. Each chat starts with your selected table and columns, ensuring controlled, explainable, multilingual analysis - without exposing more than you intend.
Grounded in governed data
You choose the table and columns for each chat. Nothing else is exposed.
Explainability baked in
View the generated query, the executed read-only query, and the actual results.
Per-chat guardrails
Select your model provider and privacy mode (Minimal / Balanced / Full) before asking.
You
“How much pipeline will we lose if we pause spend in LATAM?”
AskEdi
“Projected dip: -$1.8M over 30 days. LATAM SMB accounts drive 62% of pipeline within that cohort. Recommended action: scale down paid by 35% while increasing partner plays.”
Cited: revenue_cube.ltv_bands, spend_attribution.v3
Guardrails
What you get
CONTEXT MODES
Minimal, Balanced, or Full Context governs metadata, sample rows, and outputs. Pick a mode before you share anything.
Minimal
Anonymized names, real types, column descriptions only. No samples or outputs.
Balanced (default)
Governed names, descriptions, and 5-row samples. No full query outputs.
Full Context
Balanced plus approved query outputs for deeper analysis when allowed.
ASK • CITE • GOVERN
AskEdi turns governed datasets into instant analysis with lineage, masking, and model guardrails baked in. Business teams move faster while security keeps every question explainable.
Mode-based context
Minimal, Balanced, or Full Context governs what metadata, samples, or results reach the model.
Per-chat provider choice
Choose Anthropic, OpenAI, or DeepMind each chat within your plan’s allowed providers.
Transparent queries
View the AI-generated query and the executed read-only query to validate results.
Full audit trails
Prompts, responses, and context exposure log automatically for compliance and post-mortems.
Multilingual answers
Chat in English, Hindi, Marathi, or Tamil as needed.
FROM PROMPT TO PROOF
AskEdi wraps each conversation in policy-aware context. Teams get instant analysis, while admins see exactly what was shared, which model was used, and how the logic ties back to governed data.
Select the dataset
Pick one governed table and choose which columns to include. Exclude PII or sensitive fields up front.
Decide what context to share
Choose a privacy mode (Minimal, Balanced, Full) to control metadata, sample rows, or outputs.
Assign guardrails + models
Pick the LLM provider allowed by your plan for this chat and confirm read-only execution.
Ask and audit
Questions, responses, and queries stay logged so teams can trace every decision.
WHY IT MATTERS
AskEdi replaces ad-hoc questions, but keeps the controls. Each chat starts with table/column selection, mode, and provider choice.
Security teams keep audit trails while operators get instant clarity. Export answers as PDF/PNG/CSV, and remember: quality depends on your metadata.
Get Analysis When It Matters
No dashboards. No delays. AskEdi delivers structured, live insights - even seconds before the decision that matters.
Natural Language, Not SQL
Chat in English, Hindi, Marathi, or Tamil. AskEdi returns governed, explainable results.
Column-Smart Context
Column insights are mandatory (and better when validated). Pick the table and columns per chat to exclude PII.
Governance Built In - Not Bolted On
Privacy modes, per-chat provider choice, and read-only execution keep every request scoped and logged.
Export-Ready Answers
Download full chats as PDF, and export visuals as PNG or data as CSV when charts are included.
Secure by Design. Auditable by Default.
All logic runs inside Edilitics; queries are read-only, and every action is tracked - garbage in, garbage out still applies.
FREQUENTLY ASKED QUESTIONS
Here’s what teams ask before they centralize their stack. Need something more specific?
Talk to us.
AskEdi is an AI-powered assistant that lets you get real-time analysis from your unified, analytics-ready data - directly inside Edilitics. Instead of writing SQL or building dashboards for every question, you initiate natural-language analysis and get structured, explainable results, powered by your governed data and the latest enterprise LLMs.
AskEdi closes the “decision gap” - providing live, actionable analysis exactly when the question arises, whether in a meeting, on the go, or at the executive level. No more waiting for custom dashboards, chasing down analysts, or wrangling with query tools. AskEdi empowers anyone to unlock value from trusted data at the speed of business.
Dashboards and SQL work well for recurring, predictable reporting. AskEdi is for the questions you can’t anticipate - one-off, ad-hoc, or critical “what if” scenarios. There’s no code, no setup, and no dependency on engineering or data teams. You get instant, explainable analysis - directly from your governed data, not from raw, unreliable sources.
AskEdi only operates on datasets you explicitly connect and enable. You control which tables, columns, and metadata are shared. For best results, AskEdi should be used with data that’s already cleaned, transformed, and documented - though you can technically query any source. The more analytics-ready your data, the more meaningful and accurate your analysis.
You need integrated data sources and mandatory AI-driven column insights on the tables you use. Quality improves when users validate or edit those insights. Column selection is per chat, so you can exclude PII or sensitive fields up front.
Each chat starts by picking one table and its columns, then selecting a privacy mode (Minimal, Balanced, or Full Context) and the LLM provider. Nothing is exposed by default beyond what the mode allows. The model never touches your source directly; only governed metadata, optional sample rows, or approved results are shared.
AskEdi is powered by the latest enterprise-grade language models from Anthropic, OpenAI, and DeepMind. You can select which model to use for each session - giving you flexibility and control over performance and output. There is no auto-routing; the choice is always yours.
AskEdi is architected so that your data, prompts, and analysis outputs are not used to train or improve the underlying models from Anthropic, OpenAI, or DeepMind. We use only enterprise-grade API endpoints, which - according to the latest official documentation - do not use API-submitted data for training or product improvement. However, ultimate data handling policies rest with the model providers, and we recommend you review their latest documentation if you have specific compliance concerns.
Once your sources are integrated and metadata is completed, you can start chatting with your data in under five minutes. No code, no technical setup, and no dependency on data engineering. The onboarding flow is designed for instant time-to-value.
AskEdi is built for decision-makers, business teams, and anyone who needs timely insight without technical overhead. C-suite, product leads, ops, and analysts can all get governed, trusted analysis - without bottlenecks or handoffs. For analytics and data teams, AskEdi eliminates ad-hoc request overload, freeing them to focus on higher-value work.
READY FOR GOVERNED AI?
Bring a question that keeps coming up. We’ll show how AskEdi uses governed context, masking, and audit trails to analyze it in seconds.
