Use cases
Five problems Data Foundations is built for
Each one starts with a question a data leader is asked and cannot answer quickly — what is in this estate, what does this table mean, can we prove this number, where will AI pay off. The platform answers from the evidence, then helps you act on it.
Move a BI estate to Amazon Quick
For heads of BI and data platform leads with hundreds or thousands of Qlik or MicroStrategy dashboards.
The problem
Nobody knows which dashboards matter, what they read, or what the move will cost — and interviewing owners does not scale to thousands of apps.
What the platform does
- Parses every load script, object and report definition into one inventory
- Traces each dashboard to the tables, connections and files it reads
- Triages every asset by complexity, with the reasons and an effort band
- Designs the gold-layer model the dashboards need on the target warehouse
What you get
- Estate inventory workbook
- Estate lineage and most-shared sources
- Gold-layer model with DDL
- Amazon Quick starter pack per Qlik app
The migration plan is built from what the estate actually contains, so waves are ordered by shared sources and effort is estimated before the first sprint.
Understand a warehouse you inherited
For data engineering teams taking over a platform with little documentation.
The problem
The tables are there but the meaning is not: undocumented columns, unknown quality, and transformation logic spread across SQL, SAS and application code.
What the platform does
- Connects to Postgres, MySQL, SQL Server, Snowflake, Databricks and BigQuery
- Profiles every column and proposes data-quality rules for review
- Reads SQL, SAS and .NET code to draw lineage between jobs and tables
- Converts legacy SQL and SAS logic to the target dialect in a reviewable workbench
What you get
- Column profiles and a data dictionary
- Data-quality rules with validation runs
- Code-derived lineage
- Converted SQL with the differences flagged
Weeks of archaeology become a documented, profiled and traceable baseline your team can plan against.
Evidence data quality for regulated reporting
For data owners and governance teams in financial services, insurance and other regulated sectors.
The problem
Controls exist in spreadsheets and people's heads. When audit or a regulator asks how a number is produced and checked, the answer has to be assembled by hand.
What the platform does
- Turns profiling results into data contracts and quality rules
- Runs validations and keeps every run, result and change
- Links each report back through lineage to its sources
- Packages rules, runs and lineage as an evidence pack
What you get
- Data contracts
- Validation history
- Lineage per report
- Evidence packs
The evidence is produced by the work itself, so it is current when someone asks for it rather than rebuilt for each review.
Build a medallion platform with change control
For platform teams standing up Land → Stage → Persist → Unify layers on Databricks, Snowflake, Redshift or Synapse.
The problem
Every team hand-writes similar pipelines differently, and it is hard to see which layer is healthy.
What the platform does
- Generates pipeline code per layer for the target platform
- Designs dimensional, Data Vault or one-big-table models from profiled data
- Tracks runs, versions and approvals for each pipeline
- Shows layer health in one cockpit
What you get
- Pipeline code per layer
- Models with DDL
- Run history
- Change log
Consistent pipelines from day one, with every change and run on record.
Decide where AI will pay off
For executives and transformation leads asked to back AI use cases with evidence.
The problem
AI business cases are written on assumptions about data readiness, governance and cost that nobody has checked.
What the platform does
- Runs a structured readiness assessment across data, platform, governance and people
- Reads the documents you provide and proposes evidenced answers for review
- Models ROI per use case with sensitivity ranges
- Benchmarks against published industry surveys
What you get
- Readiness scores
- ROI model per use case
- Assessment report
- Funding workbook
A costed, evidenced decision in weeks instead of a slide deck of opinions — and a clear list of what to fix first.
What they have in common
The same principles run through every engagement. See how it is built.
Deterministic first
Parsers, profilers and lineage builders produce the facts. Models are used where judgement helps, and their output is always reviewable.
Evidence by default
Every run, result and approval is kept, so the audit trail is a by-product of the work.
Versioned and reviewable
Models, rules and generated code carry versions and a change log; nothing ships without a person saying so.
Code you own
Outputs are SQL, DDL, pipeline code and definitions in open formats, not a proprietary runtime.
Built for estates
Tested on 5,000-dashboard estates: resumable background ingestion, paged reads and batched writes throughout.
Accelerate, don't decide
Automated work is marked as such and anything needing judgement is listed for a person rather than dropped.