Reading Time: 4 minutes

Becoming data-ready is a journey of maturity. In the first two parts of our Tableau Next blog series, we explored use cases for Salesforce’s next-level AI analytics platform and why organizations need to make their Salesforce data clean, governed, and semantically structured to accelerate AI decision-making instead of automating confusion. The organizations that get it right advance through the three deliberate Crawl/Walk/Run phases outlined below, each building the trust and infrastructure the next one requires. 

Skipping a phase does not accelerate the timeline. It guarantees a return to the beginning. The most common sequencing mistake is attempting to build governance (the Walk step) before the data itself has been audited and agreed upon. Organizations that do this find themselves governing definitions nobody trusts, which is indistinguishable from having no governance at all.

CRAWL: Collect. Refine. Clean.

Before any platform enters the conversation, the data itself must be trustworthy. This phase is unglamorous and non-negotiable. Data stewards and subject matter experts across Sales, Service, Marketing, RevOps, and IT work together to reach a shared, signed-off understanding of what the data means and who is accountable for its quality.

A Center of Excellence can drive this, but it does not need to be a large or formal body. The right people already exist in your organization. The work is to give them ownership of this problem and the mandate to resolve it. That means auditing Salesforce fields, picklists, and reporting logic for inconsistencies; identifying data quality gaps with named, accountable owners; and defining the core KPIs that the business will stake its decisions on, signed off by the stakeholders who use them.

This work will surface disagreements that have existed for years. That is the point. In practice, this phase exposes political friction as often as technical debt. Some definitions will not be resolvable immediately, and that is not failure. It is a signal. Data readiness is not about forcing agreement on day one, but about establishing ownership, escalation paths, and a shared commitment to resolve ambiguity deliberately rather than letting it persist invisibly. Better to surface them in a data audit than to discover them embedded in an AI output that leadership has already acted on.

WALK: Document. Unify. Govern.

Clean data is necessary, not just sufficient. Before AI can deliver reliable insights, there needs to be a stable analytics layer that the business already trusts and uses. This is the proving ground where metrics are validated, dashboards tell consistent stories, and teams develop confidence in what the numbers are showing them.

This phase is about institutionalizing definitions. KPIs must be documented with their full logic: the filters applied, the time windows used, the attribution rules followed, and the edge cases handled. Master data principles need to be applied so that Account hierarchies, territory structures, and product taxonomies are consistent across every report and every team.

The smarter move during this phase is to build on what already exists. If teams are using Tableau, deepen it. If CRM Analytics or Tableau is in place, build on it. The goal is not to introduce new tools, but to lock in the core BI metrics, stabilize the analytical stories the business relies on, and enforce access controls and row-level security that match organizational reality.​​ If dashboards are multiplying faster than decisions improve, the organization is stuck in Walk.

An AI layer is only as credible as the dashboards and metrics beneath it. That stability is what AI will build on. 

RUN: Activate. Govern. Act. 

The Walk phase stabilizes analytics and reveals where AI will have the most immediate impact. By the time an organization reaches the Run stage, the opportunities are already visible. The question shifts from whether to introduce AI to how to govern it well.

A semantic layer must be built: the layer that tells AI not just what the data is, but what it means. Entities must be defined. Hierarchies must be established. Time logic must be encoded. Fiscal calendars, territory models, and product hierarchies must be explicit so that AI can reason over them correctly rather than approximate them incorrectly.

And governance must travel with AI. Every AI output needs traceability. Every automated action needs a human-in-the-loop checkpoint. Every use case needs an approval process and a rollback plan. The organizations that skip these steps find themselves explaining to leadership why the AI recommended something the business would never have sanctioned.

The AI Activation Checklist

Use this as your phase-by-phase readiness guide. Complete each stage before moving to the next. Each unchecked box is not a gap to paper over, but your next investment priority.

CRAWL: Clean the data

☐ Core KPIs defined and signed off by business stakeholders

☐ Data quality gaps identified with accountable owners assigned

☐ Salesforce fields, picklists, and reporting logic documented and agreed upon

WALK: Structure and govern

☐ Governed BI datasets built in CRMA or Tableau with documented lineage

☐ Business metrics stabilized and consistently telling the right story across dashboards

☐ Access controls and row-level security mapped to your organizational structure

RUN: Bring in AI

☐ Semantic layer in place exposing entities, hierarchies, time logic, and business context

☐ AI governance policy defined: approvals, audit trail, explainability standards, and fallback procedures

☐ Human-in-the-loop decision points established for every automated action

The question that changes everything

Stop asking: “Which AI tool should we use for analytics?”

Start asking: “Are we data-ready for Tableau Next’s AI analytics?”

Invest first in definitions, semantics, and governance. The right AI will meet you where your maturity is and take you further, faster, and more safely than any shortcut ever could.

The organizations that win with AI are not the ones who launched first. They are the ones who built the foundation right and then moved with confidence.

That foundation starts with a 90 to 180-day commitment to data readiness. Not to a tool. Not to a vendor. To the truth of your own data. Connect with our team to develop the data strategy that will open up a new world of possibilities through the power of Tableau Next.