Agent-driven data profiling
Turn messy cost & pricing spreadsheets into a documented, trusted data asset
Beacon Data Prism runs a team of AI agents over the files you already have — discovering structure, explaining what each column means, scoring quality and telling you exactly what to fix first.
Files are parsed in your browser. Only profile metadata is ever stored.
Six capabilities, one workspace
Everything in the engagement scope, delivered as a platform your whole team can log into.
Autonomous discovery
Point an agent at an Excel or CSV file and every sheet, table and column is discovered and catalogued without manual mapping.
Semantic understanding
Agents infer what each column actually means in business terms — cost drivers, rates, identifiers, dates — not just its data type.
Automated quality checks
Deterministic rules plus agent review surface missing values, outliers, duplicates, inconsistent formats and possible personal data.
Natural-language queries
Ask the catalogue plain-English questions about structure, quality and priorities and get grounded, citable answers.
Dynamic documentation
Export a living Excel data profile — assets, column meanings, statistics and every finding with a suggested fix.
Multi-client workspaces
Each client gets an isolated workspace with its own team, roles and invitations. Nothing is shared between them.
How the process works
13 clear steps from your first project through reports and custom pivots to final questions about the completed analysis. Every stage is saved in the session so you can stop and resume.
Phase 1 · Set up
- 1Create a project
- 2Create a session
Phase 2 · Ingest
- 3Upload files
- 4Combine files
Phase 3 · Prepare
- 5Categorise & map
- 6Review & correct
- 7Exclusions
- 8Transformation lineage
Phase 4 · Analyse
- 9Pricing analysis
- 10Charts & waterfall
Phase 5 · Deliver
- 11Reports
- 12Pivot report
- 13Ask the catalogue
Step 1
Create a project
Name the engagement for a client. Opening it takes you inside, where its sessions live.
Step 2
Create a session
Inside the project, start a working session — a subfolder that holds its own files and analysis. Leave and resume it any time.
Step 3
Upload files
Inside the session, add transaction, cost, quote, contract and master files in any common format. Preview any file's columns and sample rows.
Step 4
Combine files
Join, merge or stack uploaded files on shared columns, preview the combined columns and sample rows, and save the result as a new file used in every next step.
Step 5
Categorise & map
Confirm which requested data item every file answers, then match its columns to the fields the analysis needs.
Step 6
Review & correct
Work the checklist: supply missing columns, answer open questions and clear quality issues, alone or in bulk.
Step 7
Exclusions
Leave out what should not count: intercompany, returns, non-sellable, international or named accounts, plus unused columns.
Step 8
Transformation lineage
See the full audit trail for any file — how it was categorised, renamed, mapped and excluded — and export it.
Step 9
Pricing analysis
Per-product price trends, discount levels and cost-to-price ratios across the cleaned data.
Step 10
Charts & waterfall
The quote-to-margin waterfall plus supporting charts, filtered by region, customer and period.
Step 11
Reports
Review and export the analysis workbook and printable findings report before the final data exploration step.
Step 12
Pivot report
Build custom cross-tabs by dragging dimensions and measures, then download the underlying rows for further analysis.
Step 13
Ask the catalogue
Finish by asking plain-language questions across the session and get answers grounded in its uploaded files and completed analysis.
Ready to see your data clearly?
Create a workspace, invite your team and profile your first file in minutes.
Create your workspace