Answer these questions to get a personalised data trust score

Overall progress

Category 1: Data Foundations

Completed

Category 2: Data Quality & Trust Operations

In progress

Category 3: AI-Ready Data

Not started

Answer these questions to get a personalised data trust score

How do you bring data stakeholders together across your organization?

We don’t have an established way of bringing data stakeholders together. Data stakeholders operate in silos.

We rely on informal chats or individual relationships.

We sometimes run syncs or working sessions for key initiatives.

We have a community of practice or recurring cross-functional forums.

We run a formal, cross-functional governance program with active stakeholder engagement and decision-making authority.

What types of cross-functional data bodies exist at your org?

None, there are no formal cross-functional data groups.

We have informal working groups or ad hoc collaborations.

We have a data governance committee or working group.

We have multiple cross-functional bodies (e.g., data governance, AI oversight, architecture council).

We have an integrated data governance structure with clearly defined roles, responsibilities, and regular engagement across business, data, and technology teams.

Do you classify your data according to sensitivity or privacy level?

We don’t classify data by sensitivity or privacy.

Some sensitive data is labeled manually, but not consistently.

We classify certain types of data (e.g., PII, PHI) in key systems.

We use standardized categories (e.g., public, confidential) across most data domains.

We have automated, policy-driven classification across systems, with enforcement tied to data use and access.

How do you manage the lifecycle of your data (retention, archiving, deletion)?

We don’t have defined lifecycle policies or processes.

Some teams set their own rules, but it’s inconsistent.

We’ve defined retention policies for certain systems or sensitive data.

We have organization-wide policies, with some automation in place.

We fully automate data lifecycle management across systems, with regular reviews and policy enforcement.

How do you manage privacy risk across your data ecosystem?

We don’t actively manage or assess privacy risk.

We rely on manual controls or one-off policies in certain systems.

We’ve identified sensitive data types and apply basic access restrictions.

We track privacy risks in a structured way (e.g., risk register, data classification, periodic reviews).

We have a formal, organization-wide privacy risk management program with automation, auditability, and regular compliance checks.

Where are data issues recorded and tracked across teams?

Answer 1

Answer 2

Answer 3

Answer 4

Answer 5

How do people across your organization raise data quality concerns?

There’s no formal process, people just flag issues as they arise.

Concerns are usually raised in meetings or chat, but it’s inconsistent.

We have a ticketing system or helpdesk process for reporting issues.

Reporting issues is part of our regular workflows or tools.

We have a structured, organization-wide process for reporting and triaging data quality concerns.

Answer all of the questions above

Next Section

How is your organization using data quality dimensions?

We’re not familiar with data quality dimensions.

We understand the concept but don’t measure them consistently.

We track a few dimensions manually for key datasets.

We regularly measure multiple dimensions with some automation.

We have a fully automated, organization-wide system to measure and monitor key data quality dimensions across all critical datasets, with thresholds and alerts tied to business impact.

How do you detect data quality issues?

We don’t actively detect them.

We rely on individuals to report issues manually.

We have basic monitoring with standard alerts.

We use automated tools to detect some issues, either in-flight or at rest.

We have automated detection in place for both in-flight and at-rest data issues across key systems.

Do you track how long it takes to resolve data issues once identified?

We don’t track resolution time at all.

It’s tracked informally by individuals or teams.

We use a ticketing system to log and close issues.

We monitor resolution trends over time to identify bottlenecks.

We track resolution metrics and actively use them to improve processes and team performance.

What are you doing to monitor and manage your data pipelines to ensure high data trust?

We don’t actively monitor our data pipelines.

We find out about issues when data is missing or someone reports it.

We track job success/failure and basic timing metrics.

We monitor for anomalies in volume, schema, and freshness across key pipelines.

We have end-to-end observability with automated alerts, impact analysis, and trust indicators tied to pipeline performance.

How does your team use data lineage today?

We don’t use lineage information.

Lineage is sometimes documented manually, but not consistently used.

We use partial lineage views to support audits or answer basic questions.

Lineage helps us understand downstream impact and supports change management.

We have automated, end-to-end lineage that drives auditing, impact analysis, and data trust across teams.

When a data issue is identified, how clearly is ownership assigned for resolving it?

Ownership is often ambiguous or missing. It's usually unclear who is responsible.

Ownership is assigned manually on a case-by-case basis.

Most issues are routed to known teams, but not always consistently.

Issues are assigned to documented owners through defined processes or tooling.

Ownership is clearly defined, automatically assigned, and tracked through resolution across all key domains.

How do teams assess whether a dataset is trustworthy or ready for use?

They use it without checking, there’s no process for trust evaluation.

They ask around or rely on tribal knowledge.

They manually inspect the data for issues (e.g. values, formatting, freshness).

They check available metadata like ownership, lineage, and last update.

They follow a standardized process or framework to evaluate trust before use.

How easy is it for business teams to find and use trusted data for their work?

Very difficult, most people don’t know what data exists or where to find it.

Some documentation exists, but teams mostly rely on asking others.

We have a catalog or directory, but it’s incomplete or underused.

A searchable catalog is available and used by most business teams.

Discovery is fully self-serve, with clear ownership info, and usage guidance for business users.

Answer all of the questions above

Next Section

What is your data team doing to ready itself for reliable AI deployment?

Answer 1

Answer 2

Answer 3

Answer 4

Answer 5

Do you have an internal AI governance framework or council?

No governance efforts in place for AI today.

Some informal discussions or working groups, but nothing documented.

We have a named council or initiative, but its role is limited or unclear.

We have an AI governance framework or council that defines policies, reviews use cases, and advises on risk.

We have a well-established AI governance program that drives policy, risk management, ethical standards, and cross-functional accountability.

How are you tracking how data is used in AI/ML projects?

We don’t track or control how data is used in AI/ML projects.

Some teams review usage manually or set their own policies.

We have basic controls in place, like approved tools or sign-off requirements.

We monitor which datasets are used in AI/ML projects and review usage periodically.

We have centralized oversight with access controls, audit logs, and policies that outline how data is and can be used in AI/ML applications across the organization.

How do teams know which data is approved for use in AI/ML projects?

They don’t, there’s no visibility into what’s approved or safe to use.

Teams choose datasets based on past usage or personal judgment

Some datasets are labeled or recommended, but it’s inconsistent.

We have a defined list or tagging system for approved AI-ready datasets.

AI-ready data is clearly marked, easily discoverable, and tied to quality, privacy, and governance metadata.

Do you have a process to certify data as ready for use in AI or advanced analytics?

No, data is used as-is without any certification.

Some datasets are manually reviewed before use.

We have criteria to certify data for AI, but it's not consistently applied.

Most high-impact datasets go through a defined validation or approval process.

We have a formal, repeatable process to certify readiness for AI/analytics use.

How do you ensure data used in AI models is high quality and contextually accurate?

We don’t apply any specific checks before using data in AI.

We rely on teams to choose reliable data sources.

Some AI use cases include manual data validation steps.

We check for quality, completeness, and relevance before using data in models.

We use standardized QA checks, documentation, and approval steps before data can power AI models.

Do you have a program to govern the full lifecycle of data used in AI/ML applications?

No, there’s no program for governing AI-related data.

Governance happens informally or on a per-project basis.

We’re starting to define policies for data use in AI workflows.

We have a partial program that governs key areas like access, updates, or model input review.

We have a comprehensive AI data governance program covering sourcing, validation, lineage, use, and ongoing monitoring.

What scanning or discovery tools do you use to identify sensitive data in AI training datasets?

We don't use any tools to scan for sensitive data in AI datasets.

We rely on manual reviews or spot-checks when teams remember to do them.

We use basic scanning tools occasionally for key datasets or when required.

We regularly scan most AI datasets with automated tools that detect common sensitive data types.

We have comprehensive, automated scanning integrated into all AI data workflows, with real-time detection and alerts for sensitive data across all datasets.

Do you maintain records of which specific datasets and data versions were used to train each AI model?

No, we don't track which data was used to train each of our models.

Some teams keep informal notes or documentation about training data, but it's inconsistent.

We document training datasets for key models, but version tracking is limited.

We maintain detailed records of training datasets and versions for most models, with some gaps.

We have complete, automated lineage tracking that captures all datasets, versions, and changes used in every model throughout its lifecycle.

How do you track and manage ongoing privacy risks and violations related to AI data usage?

We don't formally track privacy risks or violations related to AI data usage.

Privacy issues are handled informally when they're discovered or reported.

We maintain a basic log of privacy incidents, but follow-up is inconsistent.

We track privacy risks systematically and have processes to investigate and resolve violations.

We maintain a comprehensive privacy risk registry with automated monitoring, incident tracking, and systematic remediation processes for all AI data usage.

Answer all of the questions above

Submit

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form.

See Bigeye in Action

Answer these questions to get a personalised data trust score

Answer these questions to get a personalised data trust score