- 8 min read
- November 2026
AI Readiness Checklist for CRM Data: 10 Checks for Dynamics 365 (2026)
Deryn Crighton-Smith
Quick Answer: How Do You Measure AI Readiness?
To measure AI readiness, pick one AI use case. Then check the CRM data it relies on against six qualities: complete, valid, unique, current, consistent and explainable. Set rules for each, score your records against them, and keep re-measuring as your data changes. Readiness is an ongoing process, not a one-off clean-up.
Who Is This Checklist For?
This checklist is written for teams using Microsoft Dynamics 365 and Dataverse who are planning, or already running, AI features such as Copilot, AI agents or automated outreach.
The 10-Point AI Readiness Checklist at a Glance
| Check | Ask Yourself | |
|---|---|---|
| 1 | Define the AI use case | What will the AI actually do with our CRM data? |
| 2 | Identify the data it relies on | Which records, fields and relationships will it use? |
| 3 | Set your readiness rules | What needs to be true for that data to be trusted? |
| 4 | Measure completeness | Are the critical fields and relationships populated? |
| 5 | Check validity | Do important values pass the appropriate checks? |
| 6 | Measure uniqueness | Are duplicates fragmenting the customer picture? |
| 7 | Check currency | Is the information still up to date? |
| 8 | Test consistency | Does data follow the same standards across the CRM? |
| 9 | Make quality explainable | Can users see why data passed or failed? |
| 10 | Monitor continuously | Can you measure, act and improve as data changes? |
Why Should You Measure AI Readiness?
AI agents need reliable business context, not just access to more data. A CRM can hold millions of records and still be untrustworthy. AI can only summarise, recommend and automate from the signals it can see. If those signals are incomplete, duplicated, invalid or out of date, the output rests on an unreliable base.
In Dynamics 365 and Dataverse, those signals include:
- Contacts and accounts
- Leads and opportunities
- Activities and the relationships between records
- Dates, validation statuses and ownership
- Consent and customer status
Treat AI readiness as a data quality, data design and governance challenge. Quality should be measurable and visible inside your CRM, not left to an occasional clean-up or an external spreadsheet.
What Does AI-Ready CRM Data Look Like?
Trusted, AI-ready CRM data has six core qualities:
| Quality | What It Means |
|---|---|
| Complete | The critical fields and relationships needed for the scenario are present. |
| Valid | Email, phone, address and other important values pass the defined checks. |
| Unique | Duplicate records do not fragment the customer picture. |
| Current | Activity, ownership, status and lifecycle information reflect the present position. |
| Consistent | Users and systems follow agreed standards across records and tables. |
| Explainable | Quality scores, rules and fixes are visible and easy to understand. |
A mandatory field does not prove the data is trustworthy. It can hold a value that is invalid, outdated or at odds with another record. That is why AI readiness needs to be measured, not assumed.
1. How Do You Define Your AI Use Case?
Start by being clear about what you want AI to do. The data needed for one use case can be very different from another.
For example: an internal account summary depends on complete account relationships and recent activity. An automated email campaign depends on email validity, consent and customer status.
A record-level readiness score should show how confidently a record can be used for a specific use case. It is not a universal verdict on whether a record is good or bad.
Checklist question: What exactly will your AI use case do, and what could happen if the data it relies on is wrong?
2. Which CRM Data Does Your AI Rely On?
Next, map the information behind the use case. For Dynamics 365 and Dataverse, AI context is grouped into four areas:
- Tables: contacts, accounts, leads, opportunities and relevant custom tables.
- Relationships: links between customers, contacts, households, memberships, campaigns or opportunities.
- Signals: activities, dates, validation status, ownership and consent.
- Consumption: where the data is used, such as Copilot, AI agents, analytics, dashboards or automation.
Individual fields rarely tell the whole story. An AI-generated customer summary might draw on a contact, its related account and its activity history. If those links are incomplete, or the same customer exists in several records, the AI sees only part of the picture.
Checklist question: Which tables, records, relationships and fields does your AI use case depend on?
3. What Does "Ready" Mean for Your Data?
Turn readiness into measurable rules that reflect the purpose of each record. Not every field matters equally. For email outreach, email quality is critical. An empty optional field matters far less.
Data8 Data Integrity lets Dynamics 365 teams configure these rules in Dataverse. You choose the table and scoring field, which criteria feed the result, and which records each rule applies to.
Checklist question: Have you defined the rules a record must meet before your AI use case can rely on it?
4. How Do You Measure Completeness?
Completeness asks whether the information your AI needs is actually there. Avoid scoring every empty field equally. Otherwise, a missing optional value can pull a score down as much as a missing essential one.
Remember that complete does not mean valid. A populated email field passes a basic completeness check. It says nothing about whether the address works.
For example: a lead record with a phone number, email and company name scores as complete, even if the phone number has too few digits.
Checklist question: Are the critical fields and relationships for this AI use case populated?
5. How Do You Check Validity?
Validity asks whether a value can be relied on for its purpose. The address name@example.com looks like an email, but passing a format check is not the same as passing a validation rule.
Data8’s validation services check postal addresses, email addresses and telephone numbers when a record is created and when it is updated.
Checklist question: Can you show that important values meet the defined checks, rather than just confirming a field has something in it?
6. How Do You Find Duplicates That Fragment Your Customer View?
Duplicates are a particular problem for AI because they create several versions of the same customer. Summaries, segmentation and decisions all work better from one trusted customer view. Potential duplicates should be surfaced and reviewed.
Duplicare is Data8’s Queen’s Award winning deduplication solution for Dynamics 365. It finds, merges and helps prevent duplicate CRM records.
Proof it works: manufacturer Tensar used Data8 Data Integrity and Duplicare to reduce its database size by 40% and to identify nearly one million records for remediation.
Checklist question: Could duplicate records be giving your AI conflicting or fragmented versions of the same customer?
7. How Do You Check Your Data Is Still Current?
Good data becomes poor data over time. People change jobs. Companies move. Contact details, account owners and customer statuses all change. Currency is therefore a core part of AI readiness.
Current data is information where activity, ownership, status and lifecycle dates reflect the present position. Some checks need recalculating even if nobody edits the record. Data8 Data Integrity supports batch recalculation and scheduling for relevant rules, which matters most for date-based assessments.
For example: a contact with a valid email and a recent activity date last year may still be out of date if their account owner has since left.
Checklist question: Does your AI have enough evidence that the information it uses still reflects the current position?
8. How Do You Measure Consistency?
AI needs information to mean the same thing across your CRM. Inconsistent data makes matching, segmentation, analytics and AI outputs harder to trust.
Apply your standards to every route that creates or changes data: web forms, imports, integrations and manual entry. Cleaning one source while leaving another unchecked only moves the problem.
Checklist question: Are the same data standards followed across teams, tables and data-entry routes?
9. How Do You Make Data Quality Explainable?
An AI readiness score should never be a mystery number. Data8 Data Integrity’s Data Quality Score Indicator shows the score alongside weighted improvement suggestions. You can customise these so users see each finding in terms that fit your CRM processes. It also shows recent score changes and historical performance, so teams can see how records change over time.
Checklist question: Can users see what is lowering a record’s readiness and what to do about it?
10. How Do You Monitor AI Readiness Continuously?
Do not treat AI readiness as a project you finish once. Quality monitoring is a cycle: validate, deduplicate, enrich, score and govern.
Data8 Data Integrity stores a snapshot whenever a score changes and keeps rule version information during recalculation. This matters because a higher score does not always mean better data. The scoring rules may simply have changed. You can also use score data in reporting tools such as Power BI, once your reporting environment is set up.
Checklist question: Can you see when data quality improves or gets worse, and act on it?
How Can Data8 Help You Measure AI Readiness?
Data8 has helped more than 1,000 organisations worldwide keep customer data accurate, compliant and useful. For Dynamics 365, our AI-ready data loop has five steps:
1. Validate
Check key information at the point of capture and when it is corrected. Data8 validation covers email, phone and address data.
2. Deduplicate
Duplicare supports duplicate detection, prevention and controlled resolution, to build a single customer view.
3. Enrich
Add useful context so segmentation and decisions work better. Business Insights enriches company data with access to over 2.5 million UK businesses.
4. Score
Data8 Data Integrity measures CRM records against your own quality and readiness rules and makes the scores available in Dataverse.
5. Govern
Use scores, history, reporting and improvement processes to keep quality visible as your data changes.
The result is a continuous approach where you can find issues, fix them and measure the effect, instead of relying on assumptions about your CRM.
Why Does AI Readiness Start with Trusted CRM Data?
AI can make CRM data more useful, but it does not make poor data disappear. Before focusing on what an AI agent can automate, look at the information it depends on. Measure it. Validate it. Deduplicate it. Improve it. Then keep monitoring it. That is how AI readiness becomes something you can demonstrate, not just claim.
Even with high-quality data, check important AI-generated content before it drives a decision. AI output can still be wrong when the underlying data is good, so always keeps a human check in place.
Ready To Find Out How AI-Ready Your CRM Data Is?
Our data quality experts can help you measure your current data quality, find where the biggest risks sit and build an ongoing approach to validation, deduplication, scoring and governance.
AI-Readiness for Dynamics 365 FAQs
What is AI-ready CRM data?
AI-ready CRM data is information that is trustworthy enough for the specific AI use case relying on it. Data8 assesses this through six qualities: completeness, validity, uniqueness, currency, consistency and explainability.
How do you measure AI readiness?
Identify the CRM records and fields your AI use case depends on. Define readiness rules, measure records against them, review the failed rules and keep monitoring as the data changes.
What is a good AI readiness score?
There is no universal pass mark. Read a score alongside the rules assessed, how important they are, the records included and the intended use of the data.
Why do duplicates matter for AI?
Duplicate records split the customer picture, leaving activities, details or statuses across several records. That gives AI incomplete or conflicting context, so Data8 treats uniqueness as a core part of trusted CRM data.
Is a CRM ready for AI if all required fields are complete?
Not necessarily. A required field can hold a value that is invalid, stale or inconsistent. A populated field is not evidence that a record is ready for AI.
How often should you measure AI readiness?
Continuously. Records and the rules applied to them change over time, so Data8 recommends ongoing scoring and governance rather than a one-off exercise.
Does this checklist work for CRMs other than Dynamics 365?
Yes. The ten checks apply to any CRM. The Data8 tools referenced here (Data Integrity and Duplicare) are built for Microsoft Dynamics 365 and Dataverse.
What is the difference between data quality and AI readiness?
Data quality describes how good your data is in general. AI readiness asks whether the data is good enough for a specific AI use case, measured against rules that matter for that use.