chief-data-officer-advisor
by alirezarezvanichief-data-officer-advisor is a strategic CDO skill for startup data decisions: AI training data rights, warehouse vs lakehouse vs mesh strategy, customer-data asset valuation, M&A readiness, and data team hiring. Includes references and Python tools for decision support, not tactical data engineering.
This skill scores 84/100, which means it is a solid listing candidate for directory users who want strategic Chief Data Officer decision support rather than generic prompting. The repository evidence shows clear triggers, focused decision frameworks, and executable helper scripts, though adoption would be easier with a README/install quick start and more end-to-end examples.
- Highly triggerable frontmatter: it names concrete use cases such as AI training data rights, lakehouse vs mesh, data asset valuation, M&A readiness, and data hiring decisions, and excludes tactical data engineering.
- Substantial operational content organized around four strategic CDO decisions, with dedicated references for training-data rights, data product strategy, customer data valuation, and data team evolution.
- Includes three stdlib Python tools with documented JSON schemas and sample usage for training-data audits, architecture selection, and data asset valuation.
- No install command or README is present, so users must infer installation from the repository path rather than a packaged quick start.
- The AI training-data rights workflow explicitly says it is not legal advice and should be used to surface issues for counsel, not replace legal review.
Overview of chief-data-officer-advisor skill
What chief-data-officer-advisor is for
chief-data-officer-advisor is a strategic data leadership skill for founders, startup executives, and AI teams that need CDO-style judgment before making irreversible data decisions. It focuses on four board-level questions: whether data can be used for AI training, which data architecture fits the company stage, how customer data should be valued or productized, and what data role to hire next.
This is not a SQL, pipeline, schema, or dashboard-building assistant. The chief-data-officer-advisor skill is best used when the decision has legal, organizational, fundraising, product, or M&A consequences.
Best-fit users and situations
Use this skill if you are deciding whether to train models on customer data, move from warehouse to lakehouse, resist premature data mesh adoption, quantify a customer-data moat, or sequence data hires after founder-led analytics stops scaling.
It is especially useful for B2B SaaS, AI startups, marketplaces, and data-rich products where customer contracts, consent provenance, data exclusivity, and productization risk matter. The strongest fit is chief-data-officer-advisor for Strategic Planning, not day-to-day engineering execution.
What makes it different from a generic prompt
A generic prompt may give broad data-strategy advice. This skill is more decision-oriented: it uses explicit frameworks, stage-based thresholds, and supporting Python scripts for repeatable analysis. The repository includes references for AI training data rights, data product strategy, customer data as an asset, and data team org evolution, plus scripts for audits, architecture selection, and valuation.
Important boundaries before install
The skill surfaces strategic risks and decision options; it does not replace legal counsel, security review, data protection impact assessments, or technical architecture design. Its AI training data guidance is especially useful for issue spotting, but legal approval is still required for regulated, PII-heavy, partner-licensed, scraped, or externally shared data.
How to Use chief-data-officer-advisor skill
chief-data-officer-advisor install context
Install from the GitHub skill path if your agent supports skill installation from repositories:
npx skills add alirezarezvani/claude-skills --skill chief-data-officer-advisor
The upstream skill lives at:
c-level-advisor/skills/chief-data-officer-advisor
After install, preview SKILL.md first, then read the reference file matching your decision. For practical runs, inspect:
references/ai_training_data_rights.mdreferences/data_product_strategy.mdreferences/customer_data_as_asset.mdreferences/data_team_org_evolution.mdscripts/ai_training_data_audit.pyscripts/data_product_strategy_picker.pyscripts/data_asset_valuator.py
Inputs that make the skill useful
The chief-data-officer-advisor usage quality depends on concrete company facts. Include stage, ARR if relevant, data sources, consent model, customer contract constraints, data volume, internal data consumers, ML production status, and the decision deadline.
Weak prompt:
“Should we use a lakehouse?”
Better prompt:
“Use chief-data-officer-advisor to decide our data architecture. We are Series A B2B SaaS, 55 employees, 3-person data team, 8 weekly data consumer groups, 4.5TB product and event data, one churn model in production, Snowflake today, S3 logs unused, board wants self-serve BI and ML feature reliability within 12 months. Recommend warehouse vs lakehouse vs mesh, build-vs-buy by layer, kill criteria, and a sequencing plan.”
Suggested workflow for strategic planning
Start by naming the decision, not the tool you expect. Ask the skill to classify the decision into one of its four domains: training data rights, data product strategy, customer-data asset value, or org evolution.
For AI training questions, prepare a JSON-like list of data sources with origin, data_class, and use_case, then compare the answer with scripts/ai_training_data_audit.py. For architecture questions, prepare a company profile compatible with data_product_strategy_picker.py. For M&A or monetization questions, prepare a corpus profile and run it against the valuation framework.
Practical prompt pattern
A strong chief-data-officer-advisor guide prompt has five parts:
- Context: company stage, product type, customers, regulatory exposure.
- Decision: the exact choice you need to make.
- Current state: data stack, team, contracts, consent, volume, ML use.
- Constraints: budget, timeline, buyer diligence, customer carve-outs.
- Output format: recommendation, risks, mitigations, sequencing, open questions.
Ask for “decision memo” output when presenting to executives, and “audit table” output when comparing data sources or architecture options.
chief-data-officer-advisor skill FAQ
Is chief-data-officer-advisor only for companies with a CDO?
No. The skill is often more valuable before a company has a CDO. It helps founders, CTOs, heads of product, and AI leads avoid premature architecture moves, risky data monetization claims, or mis-sequenced data hires.
When should I not use this skill?
Do not use it for writing ETL code, optimizing queries, designing schemas, configuring dbt, selecting exact cloud services, or debugging pipelines. It can recommend whether to buy or build a layer, but it will not produce a production-ready technical implementation plan.
How does it compare with ordinary strategy prompting?
Ordinary prompts often reflect fashionable advice: “adopt a lakehouse,” “build a data mesh,” or “hire data scientists.” This skill is more constrained. It ties recommendations to stage, data volume, consumer count, ML maturity, consent provenance, contractual restrictions, and organizational readiness.
Is the chief-data-officer-advisor skill beginner-friendly?
Yes, if the user can describe the business situation. You do not need to be a data architect, but you do need to provide accurate facts. If you cannot answer where the data came from, who consented, who uses it weekly, or what decision is blocked, the first output should be treated as discovery, not a final recommendation.
How to Improve chief-data-officer-advisor skill
Improve chief-data-officer-advisor inputs
The fastest way to improve results is to replace vague goals with decision evidence. Instead of “Can we train on customer data?”, list each source separately: support tickets, product telemetry, uploaded files, call transcripts, partner feeds, synthetic data, and scraped data. For each, provide origin, consent wording if known, data class, retention rules, deletion process, and intended model use.
Common failure modes to watch
The most common failure is asking for a confident strategic answer while hiding key constraints. Missing customer data carve-outs can distort valuation. Missing ML production status can lead to premature lakehouse recommendations. Missing consent provenance can make training-data advice too optimistic. Missing company stage can produce a hiring plan that is too senior or too early.
Iterate from recommendation to decision memo
After the first output, ask the skill to separate “recommendation,” “assumptions,” “risks,” “mitigations,” and “questions for counsel or board.” This turns an advisory answer into an executive artifact. For high-stakes decisions, also ask for a red-team pass: “What would make this recommendation wrong?”
Extend the skill with local context
For better chief-data-officer-advisor results, add company-specific templates: approved consent language, data processing agreements, architecture standards, cloud constraints, security review requirements, and board memo format. Keep these as local context rather than changing the core framework, so the skill remains reusable while its outputs reflect your operating reality.