# AI-Powered Field Suggestions During Duplicate Merge in Dynamics 365 CRM InoWiz is the artificial intelligence engine embedded in DeDupeD that activates at the point of merge, the most critical decision moment in any duplicate detection workflow for Dynamics 365. When two or more records are identified as duplicates, InoWiz analyses every conflicting field value across all candidate records and recommends the optimal survivor value for each field, annotated with a confidence score and a plain-language rationale. The result is a faster, more accurate, and fully auditable duplicate merge operation in Dynamics 365 that preserves the highest-quality data rather than defaulting to the most recently modified record. This section documents the full InoWiz end-to-end workflow: how AI suggestions are triggered, how they appear within the DeDupeD merge interface, how confidence scores and rationale tooltips are surfaced, and how administrators can configure automatic suggestion loading at the entity level to support large-scale Dynamics 365 data cleansing programs. ## How InoWiz Works for Data Deduplication in Dynamics 365 CRM When a user initiates a merge of duplicate Dynamics CRM records, DeDupeD opens the merge screen listing the current record alongside all identified duplicates. InoWiz operates on top of this standard merge flow by evaluating every field where values differ across records. It calls the configured AI model (Azure OpenAI by default), passes the entity context, field metadata, and all candidate values as structured input, and receives a ranked recommendation with a confidence percentage for each field. The AI model applies the logic defined in the active Field Suggestions AI Prompt assessing completeness, format correctness, business context, and internal consistency across linked field groups (e.g., address components evaluated together rather than in isolation). This approach ensures that the surviving record represents the most logically consistent and data-complete view of the entity, directly improving Dynamics 365 data accuracy at the source. ### Step 1 - Initiating the Merge From the DeDupeD duplicate detection view, select the record to be processed and click Merge. The DeDupeD merge dialog opens, displaying the Current Record and all identified Duplicate Records for the same entity. Each row shows the Account Name, Modified On timestamp, Created On timestamp, and Owner, providing the contextual signals required to evaluate record quality before the AI model runs.
Select the duplicate records to be merged with the current record, then click **Merge**. DeDupeD advances to the merge field-selection screen where InoWiz operates ### Step 2 - The Merge Screen and AI Suggestions Toggle The DeDupeD merge screen presents all candidate records as side-by-side columns: a Preview column reflecting the current selection state, a Master column (the record that will survive the merge), and one column per identified duplicate labelled Set Master. Each column lists field-by-field values with radio buttons for manual selection. Three filter toggles are available in the toolbar: * Show All Columns Expands the field list to include fields where all records share identical values. * Include Same Value Columns Renders columns for fields where values are the same across all records, useful for auditing completeness. * AI Suggestions Activates InoWiz. When checked, the AI model is invoked,invoked and suggestions are loaded inline for every conflicting field.
By default, AI Suggestions is unchecked and users select field values manually. Enabling the checkbox triggers an immediate AI inference call. A "Getting AI suggestions…" loading indicator confirms the request is in-flight. ### Step 3 - AI Inference Loading State Once AI suggestions are enabled, InoWiz begins processing. The merge screen enters a loading state, field values in the candidate columns are dimmed and a spinner with the label **"Getting AI suggestions…"** is displayed. During this period, the AI model is: * Receiving the entity type and all field metadata from the active AI Prompt configuration. * Evaluating all candidate values per field across the current record and all selected duplicate records. * Scoring each candidate against completeness, format validity, and contextual consistency criteria. * Generating a ranked recommendation with a confidence percentage (0–100%) for every conflicting field.
{% hint style="info" %} **Note:** The loading duration scales with the number of fields and duplicate records being evaluated. For entities with many conflicting fields across five or more duplicates, expect a processing window of 3–8 seconds depending on Azure OpenAI service latency. {% endhint %} ### Step 4 - AI Suggestions Rendered in the Merge Interface Once the AI model returns its recommendations, the merge screen updates immediately. Each field for which InoWiz has a recommendation displays an AI Suggested (X%) badge rendered in orange alongside the radio button in the Master column. The badge confidence percentage reflects the model's certainty that this candidate value is the optimal survivor for the merged record. Simultaneously, the Master column is automatically updated with the InoWiz-recommended values pre-selected, and the Preview column reflects these choices in real time. A disclaimer banner "AI-generated suggestions may not always be accurate. Please review before applying" is displayed at the top right of the merge screen to reinforce that InoWiz recommendations are advisory, not final.
#### Address Field Groups Higher Confidence Scoring Logically linked field groups such as Address 1: City, Address 1: Country/Region, Address 1: State/Province, and Address 1: Street 1 are evaluated as a cohesive unit rather than independently. This grouped evaluation model prevents logically inconsistent outcomes (e.g., a city from one record being paired with a country from another). When address components are internally consistent across a single source record, confidence scores for that group typically score higher **(85–95%)** than isolated fields.
#### Communication and Contact Fields For contact fields such as Fax, Main Phone, and Website, InoWiz evaluates format completeness (e.g., standardized phone number formatting), value length, and cross-field consistency. Fields selected from a record that provides a complete, correctly formatted contact dataset score higher than fields sourced from records with partial or inconsistently formatted values. The Master column pre-selects the recommended values and marks each with the AI Suggested (X%) badge.
### Step 5 - Confidence Score and Rationale Tooltip Every AI Suggested badge is interactive. Hovering over or clicking an AI Suggested (X%) badge opens a "Why This Suggestion?" tooltip that surfaces two pieces of information: **Rationale:** A plain-language explanation of why InoWiz selected this specific value. The rationale references the evaluation criteria applied for example, format completeness, data richness, or cross-field consistency making the AI decision transparent and auditable. **Confidence Score:** A numeric percentage (e.g., 85%) reflecting the model's certainty in the recommendation. Scores of 90%+ indicate high-confidence selections where all evaluation criteria aligned on a single candidate. Scores of 70–89% indicate moderate confidence where the model identified a best-fit candidate but encountered competing values of similar quality.
Users retain full control. Any AI-suggested value can be overridden by selecting a different radio button in the same field row. Overriding an AI suggestion does not suppress suggestions for other fields the remaining InoWiz recommendations stay active. This design supports a human-in-the-loop approach to duplicate Dynamics CRM data resolution, where AI provides the analytical baseline and users apply business context where needed. ### Step 6 - Preview and Completing the Merge The Preview column on the far left of the merge screen provides a live composite view of the surviving record as field selections are made. Each field in the Preview column reflects the currently selected value from the Master column, giving users an accurate picture of the post-merge record before committing. Once all field values have been reviewed and any desired overrides to AI suggestions have been applied click Finish to execute the merge. DeDupeD consolidates all selected field values into the master record, deactivates the subordinate duplicate records, and reassigns all related records (activities, contacts, opportunities) to the surviving master. The Finish button is highlighted in the toolbar to indicate the merge is ready to be committed.
{% hint style="info" %} **Note:** Clicking Finish is irreversible for the merge operation. Duplicate records are deactivated, not deleted. They can be reactivated by a System Administrator if a merge needs to be undone, but field values on the surviving master record will need to be manually corrected. {% endhint %} ### Auto-Enabling AI Suggestions at the Entity Level For high-volume duplicate detection and deduplication in Dynamics 365 workflows, requiring users to manually check AI Suggestions on every merge screen adds unnecessary friction. DeDupeD provides an entity-level configuration option that automatically enables InoWiz whenever the merge screen loads for a given entity type. **Steps To configure AI suggestion on load of merge UI :** * **Step 1:** Navigate to Administration > Entity Configuration in the DeDupeD left navigation pane. * **Step 2:** Open the entity configuration record for the target entity (e.g., Account, Contact, Lead). * **Step 3:** Select the Merge Settings tab. * **Step 4:** Locate the AI suggestion toggle labelled "Automatically enables AI to analyze conflicting field values when the merge screen loads and suggest the most relevant values for the merging record." * **Step 5:** Set the toggle to Yes and click Save.
Once enabled at the entity level, InoWiz activates automatically whenever a user opens the DeDupeD merge screen for that entity without requiring the user to manually check the AI Suggestions checkbox. The AI inference call fires in the background while the merge screen renders, minimising total merge processing time for operators handling large queues of duplicate Dynamics 365 records. {% hint style="info" %} **Tip:** Enable the auto-suggestion toggle on entities where duplicate merge frequency is high such as Account, Contact, and Lead to maximize throughput during scheduled Dynamics 365 data cleansing runs. For lower-frequency entities, leave the toggle disabled and allow users to activate AI Suggestions on demand. {% endhint %} ### InoWiz Merge Interface UI Elements Reference
ElementDescription
AI Suggestions checkboxEnables/disables InoWiz for the current merge session. When checked, triggers an AI inference call against the configured Azure OpenAI model. Automatically checked when the entity-level auto-suggestion setting is enabled.
AI Suggested (X%) badgeOrange badge displayed on individual field rows in the Master column where InoWiz has pre-selected a value. The percentage is the model's confidence score for that specific field recommendation.
Why This Suggestion? tooltipInteractive tooltip triggered by hovering/clicking an AI Suggested badge. Displays the plain-language rationale and numeric confidence score for the recommendation.
{% hint style="info" %} Note: AI suggestions will not work for fields configured under Field Merge Criteria And Address merge criteria. {% endhint %} ### InoWiz Behaviour Notes • Suggestions are advisory: InoWiz never automatically commits a merge. All suggestions require a user to click Finish to take effect, maintaining compliance with human-in-the-loop data governance requirements. • Partial suggestions: If the AI model cannot produce a high-confidence recommendation for a specific field (e.g., all candidate values are equally complete or equally incomplete), that field will not display an AI Suggested badge. Users must select the value manually for those fields. • Token consumption: Each InoWiz inference call consumes tokens from the configured AI model's token budget. Monitor consumed tokens in the AI usage logs to track utilization across high-volume duplicate merge tool operations in Dynamics 365. • Disclaimer banner: The "AI-generated suggestions may not always be accurate" banner is always visible when AI suggestions are enabled. This is a non-discretionary compliance indicator reminding operators to apply professional judgment before finalizing the merge. • Multi-record evaluation: InoWiz evaluates all selected duplicate records simultaneously, not pairwise. In a merge involving five duplicate records, the model compares all five candidate values per field in a single inference call, producing a globally optimal recommendation rather than an iterative pairwise best. {% hint style="success" %} For further queries, reach out to us at [crm@inogic.co](mailto:crm@inogic.com)m {% endhint %} --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.inogic.com/deduped/ai-powered-field-suggestions-during-duplicate-merge-in-dynamics-365-crm.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.