The Augmented Consultant by Hugo Lourenço Silva
Notes & opinion · BI & Data Analytics

A consultant, augmented by AI.

Notes and opinion on BI & Data Analytics, AI agents, skills and augmented workflows, reshaping data work from raw data to the published report.

Latest writing Read the manifesto

Latest writing

All writing
Future of reporting

The static report is dying

Why the dashboard is no longer the default destination for business questions.

4 min read

Power BI

The Semantic Model is the New AI Contract

Why the quality of AI in Analytics will depend on the quality of the semantic layer.

 Meta

The Augmented Consultant Manifesto

Why AI didn't come to replace consultants. It came to raise the game.

The augmented BI lifecycle

  1. Business requirements

    Decisions to support, questions to answer, KPI definitions, analysis grain, stakeholders.

    AugmentedAn AI solution can turn interviews, docs and tickets into a traceable KPI definition document, with each metric mapped to a source and a confidence level. The consultant validates it with the business.

  2. Data discovery & profiling

    Source inventory: grain, keys, nulls, distributions, referential integrity, aggregation behaviour.

    AugmentedAI solutions can profile every dataset against the requirements, infer grain and candidate keys, and flag the risks that would break the model before a single table is built.

  3. Semantic model design

    Star schema, grain, relationships, naming. The data contract between business and model.

    AugmentedAn AI solution can propose the dimensional model from profiling and requirements, then challenge it: missing dimensions, fan-outs, ambiguous grain. The consultant signs the contract.

  4. Model implementation

    Tables, relationships, Power Query and TMDL, incremental refresh.

    AugmentedAn AI solution connected to the model through MCP can write the TMDL and M, and review each transformation against the approved design.

  5. Metrics & DAX

    Measures, KPIs, time intelligence, calculation groups, RLS logic.

    AugmentedAI solutions can write, explain and audit DAX measure by measure, with validation queries against the live model. Business logic stays the consultant's call.

  6. Quality & performance review

    Validated numbers, edge cases, BPA rules, VertiPaq footprint, refresh times.

    AugmentedAgentic workflows have the capability to run the full audit: BPA, deterministic checks over the model, storage metrics. They report a maturity score with evidence and the consultant decides what ships.

  7. Report design

    Narrative, visual hierarchy, chart selection, interaction design, accessibility.

    AugmentedAn AI solution can draft the storyline and page wireframes, propose chart types per question and data type, and flag visual anti-patterns. The designer keeps the pen.

  8. Publish & operate

    Workspaces, apps, refresh, RLS, lineage, usage, incidents.

    AugmentedAI solutions have the capability to monitor refresh and lineage, explain failures, watch adoption and propose fixes. The consultant approves every change to production.

  9. Documentation & knowledge

    Data dictionary, measure catalogue, lineage, decisions. All kept in sync with the model.

    AugmentedAI solutions can generate the documentation from the model itself (TMDL, PBIR, VertiPaq) and regenerate it on every change, so it never drifts from reality.
The manifesto

Why AI didn’t come to replace consultants. It came to raise the game.

AI doesn’t replace the consultant who thinks. It replaces the consulting that stopped thinking.
Read the manifesto