AB-210 Accelerating Sales Pipelines with AI in Dynamics 365 Exam Topics and Questions

This topic is where the exam grounds you before it lets you anywhere near AI features. Every intelligent capability in later topics depends on a correctly configured foundation, and examiners know it. Get the plumbing wrong and the AI does nothing useful. Expect to account for mailbox setup, security roles, data flows, and catalog structures, because those are the levers that control whether downstream intelligence even has clean data to work with.

– Deployment readiness and connectivity

Before a single AI feature fires, the environment has to be ready. Evaluating prerequisites for deploying Sales means checking that the right licenses are in place, the org is in a supported region, and server-side sync is enabled. Mailbox configuration sits here too: if mailboxes are not approved and tested in Dynamics 365, email tracking and Copilot email features will silently fail. Microsoft 365 integration covers Teams, SharePoint, Outlook, and OneDrive. Each has its own enable path and its own permission surface. The exam will test whether you know which integration requires which admin action, not just that the integrations exist. Pay attention to how Teams meeting intelligence and SharePoint document management attach to records, because those specifics come up.

– Security, process design, and the user experience

The Sales security model in Dynamics 365 uses a layered structure of business units, roles, and field-level security. Knowing which layer controls record visibility versus which controls field visibility is the kind of precision the exam rewards. Business process flows are configured here, and the exam distinguishes between designing a flow that matches a sales stage model and simply knowing that flows exist. The timeline configuration matters because it affects what sellers see on a record and what Copilot can surface from activity history. Import and export options span the full range from Excel to Dataverse connectors, and the exam tends to test which option is appropriate for a given scenario rather than the mechanics of any single tool.

– Product catalog and pricing structures

The product catalog is the data backbone for opportunity line items, forecasting, and AI-driven deal insights. Creating and managing it means understanding the relationship between product families, products, and product bundles, and knowing how properties cascade from family to child. Price lists control which prices a seller can apply and in which currency. The exam tests currency configuration and the relationship between price lists and territories or customer segments. Getting this wrong has downstream consequences: opportunity revenue figures become unreliable, and any AI feature that reads product or pricing data will produce misleading output. Treat the catalog not as an admin task but as a data quality gate.

How Configure Dynamics 365 Sales core features for AI is tested

Items in this topic are scenario-driven. You will be given a business situation, a current system state, and a symptom, and asked to identify the configuration step that resolves it. The trap is that many options will describe valid Dynamics 365 actions that simply do not address the stated problem. Items reward the candidate who reads the symptom first and works backward to the cause, rather than the one who recognises a familiar feature name and picks it. Mailbox approval failures, security role gaps, and misconfigured price lists all appear as symptoms in scenarios. The exam also tests prerequisite sequencing: some features cannot be enabled until others are in place, and items reward knowing that order. Microsoft 365 integrations are a reliable source of missed marks because candidates assume admin consent is automatic. It is not, and the exam knows it.

Once the foundation from the first topic is solid, this topic is where the exam expects you to build the intelligence layer on top of it. The scope is wide: strategy, data modeling, Copilot configuration, agents, forecasting, and relationship intelligence all live here. Width is the risk. Candidates who know several of these areas well but have gaps in the less-discussed ones, such as capacity and billing for agents or goal metrics, tend to lose marks they could not afford to lose.

– Strategy, data model, and planning

Designing an AI-first sales strategy means making deliberate decisions about which AI features to activate and in what sequence, not simply turning everything on. The Dataverse data model underpins all of it: entity relationships, table ownership types, and how custom tables interact with the standard Sales schema all affect what AI can read and act on. Differentiating features across Sales plans is tested because the exam expects you to know which capabilities require a premium license and which are available in the base plan. Reporting and visualization options span from built-in Sales dashboards to Power BI, and the exam tests when each is the right choice rather than how to build one from scratch.

– Agent deployment and capacity management

Preparing to deploy agents involves more than enabling a toggle. Prerequisites include environment configuration, Dataverse capacity, and ensuring the underlying AI service connections are in place. Configuring capacity, billing, and credits is a topic many candidates skip because it feels administrative rather than functional, and the exam exploits that gap. You need to know how agent consumption is metered and what happens when capacity runs out. Configuring work assignment and distribution connects to agents because agents need routing rules to know which records they should act on. The exam tests the relationship between these settings and agent behavior, not just the settings in isolation.

– Copilot, intelligence features, and the Sales accelerator

Configuring AI-powered intelligence and insights features covers a range of capabilities that share a common dependency: clean, connected data. The Sales accelerator organises seller activity into a prioritised queue, and its configuration determines what signals drive that prioritisation. Conversational intelligence requires call recording to be enabled and properly connected to the environment. Predictive scoring, both for leads and opportunities, uses historical data to generate scores, and the exam tests what inputs the model uses and what a low-confidence score signals. Copilot features, including record summarization, have their own enable path and privacy considerations that the exam surfaces in scenario questions.

– Forecasting, goals, and relationship intelligence

Forecasting in Dynamics 365 Sales is configurable at a granular level: forecast categories, hierarchy, rollup fields, and adjustment permissions all need deliberate setup. The exam tests the relationship between opportunity stages and forecast categories, because a misconfigured mapping produces forecasts that do not reflect pipeline reality. Setting up relationship intelligence, which surfaces connection health and activity gaps, depends on Exchange data being available. Goal metrics define what counts as progress toward a goal, and the distinction between amount-based and count-based metrics is the kind of detail that appears in scenario items. Candidates who treat forecasting as a reporting feature rather than a configuration domain tend to underperform here.

How Optimize AI-driven sales is tested

This is the widest topic on the exam, and the item style reflects that. Expect scenario questions that combine two or three sub-areas: a seller workflow problem that implicates both Copilot configuration and the Sales accelerator setup, for instance. Items reward candidates who can trace a business outcome back to a specific configuration decision rather than identifying individual features in isolation. Capacity and billing questions tend to be straightforward if you have studied them and costly if you have not. Predictive scoring items often describe a model behaving unexpectedly and ask what configuration or data condition explains it. Items reward understanding the inputs the model depends on. Forecasting questions test mapping logic rather than UI navigation. The breadth here means timed practice matters: candidates who work slowly through familiar areas run out of time before reaching the questions they would have answered correctly.

Lead qualification is where AI moves from supporting sellers to acting on their behalf, and this topic tests whether you understand both sides of that shift. The Sales Qualification Agent can research, score, and engage leads autonomously, but only if it is configured correctly and monitored effectively. Candidates who treat this topic as a continuation of traditional lead management miss the agent-specific configuration and governance questions that separate passing scores from failing ones.

– Lead generation, nurturing, and the lead experience

Generating and nurturing leads in Dynamics 365 involves configuring the channels and forms that create lead records, and ensuring that the data captured is sufficient for AI scoring to work. The lead and opportunity experience configuration controls what sellers see when they open a lead, which fields are visible, and which business process flow guides them. This matters for AI because the Qualification Agent reads from the record, and a poorly structured record limits what the agent can do. The predictive lead scoring model is configured here too, and fine-tuning it means understanding which attributes the model weights and how to improve model confidence by improving data completeness.

– Sales Qualification Agent: configuration and operating modes

The Sales Qualification Agent has two distinct operating modes, and the exam draws a hard line between them. Research-only mode means the agent gathers information and surfaces it to the seller without sending any outbound communication. Research and engage mode allows the agent to contact the lead directly. Choosing the wrong mode for a given business scenario is a meaningful error, and the exam tests your ability to match mode to use case. Configuring the agent involves setting its scope, connecting it to the right data sources, and defining the criteria it uses to assess lead quality. Prerequisites must be in place before configuration can succeed, and the exam tests that sequencing.

– Agent actions, interpretation, and monitoring

Once the Sales Qualification Agent is running, the exam expects you to interpret what it does and catch problems early. Interpreting agent actions means reading the agent's activity log and understanding why it took a specific step, not just that it acted. Monitoring the agent covers the dashboards and signals that indicate whether it is performing as intended, whether it is consuming capacity as expected, and whether any leads have been mishandled. Candidates who configure agents but do not study the monitoring surface tend to miss questions about what to do when an agent behaves unexpectedly. The exam rewards knowing how to diagnose agent behavior, not just how to enable it.

How Qualify and prioritize leads by using AI is tested

Items here split roughly between configuration scenarios and operational judgment calls. Configuration items give you a business requirement and ask which agent setting or mode satisfies it. The Research-only versus Research and engage distinction appears frequently because it has real business consequences and is easy to get wrong under time pressure. Operational items describe an agent doing something unexpected and ask what the most likely cause is, or what the administrator should check first. Items reward candidates who understand the agent as a system with inputs, logic, and outputs, rather than as a black box that either works or does not. Predictive scoring fine-tuning questions test whether you know what data conditions improve model performance. Candidates lose marks by conflating the scoring model configuration with the agent configuration: they are related but distinct, and the exam tests that distinction.

Lead qualification hands off to opportunity development, and this topic picks up exactly there. Three distinct agents operate in this space, the Sales Opportunity Agent, the Sales Close Agent, and the Sales Research Agent, and the exam treats them as separate systems with separate configuration paths, not as one feature with three names. Candidates who blur the boundaries between them will lose marks on questions that hinge on which agent does what and under which conditions.

– Opportunity management and pipeline configuration

Optimising opportunity management starts with the pipeline view, which is configurable in ways that affect what sellers and agents can see and act on. The opportunity pipeline view setup controls column display, filtering defaults, and the signals that surface at-risk deals. Adding opportunity products and configuring pricing connects back to the product catalog and price list work from the first topic: errors there propagate into opportunity revenue figures and affect AI deal scoring. The exam tests the relationship between these configurations and the quality of AI output, not just the mechanics of adding a line item.

– Sales Opportunity Agent: setup and coordination

The Sales Opportunity Agent is configured to assist sellers with researching and progressing open opportunities. Setting it up requires the same prerequisite checks as other agents, plus specific data connections that give it access to account and contact intelligence. Coordinating opportunity management with the Sales Close Agent is a distinct skill: the two agents have overlapping scopes, and the exam tests whether you understand where one agent's responsibility ends and the other's begins. Candidates who configure both agents without understanding their interaction risk scenarios where both are active on the same opportunity without clear handoff logic.

– Sales Close Agent: configuration, actions, and collaboration

The Sales Close Agent is designed to drive deals toward a close decision, and its configuration is more involved than its name suggests. Setting it up requires defining the conditions under which it engages, the actions it is permitted to take, and the escalation path when human judgment is needed. Interpreting the actions of the Sales Close Agent means reading its activity trail and understanding the logic behind each step. Monitoring it covers performance signals and capacity consumption. Collaborating with the agent, where a seller and the agent share responsibility for a deal, is a scenario the exam tests because it requires understanding the agent's decision boundaries.

– Sales Research Agent and performance analysis

The Sales Research Agent gathers intelligence on deals from internal and external sources and surfaces it in a structured format. Evaluating its use cases means knowing when it adds value, typically on complex, high-value deals with multiple stakeholders, and when it is unnecessary overhead. Configuring it involves connecting the right data sources and defining the scope of research. The research canvas is where the output is presented and analysed, and the exam tests how to read and act on that output rather than how to build the canvas itself. Analysing sales performance through the research canvas connects research output to coaching and process improvement decisions.

How Develop deals by using intelligent opportunity research is tested

The three-agent structure is the main source of difficulty in this topic. Items frequently describe a scenario involving an opportunity and ask which agent is appropriate, or describe an agent action and ask what configuration decision produced it. Items reward precise knowledge of each agent's scope and trigger conditions. Candidates who memorise agent names without understanding their distinct functions will consistently pick the plausible-but-wrong option. Pipeline view and pricing configuration questions tend to be scenario-based, testing consequences rather than steps. The research canvas questions reward candidates who understand what the output signals and what action it should prompt. Monitoring questions across all three agents follow a similar pattern: a symptom is described, and items reward identifying the correct diagnostic path. Time management matters here because the scenario stems are longer than average.

The final topic shifts from configuring what Dynamics 365 Sales provides out of the box to extending it with the broader Microsoft ecosystem and Power Platform. The exam tests integration judgment: knowing when to use a native feature versus a Power Automate flow, and when embedding a Power BI report is the right answer versus using a built-in dashboard. This is also where mobile, calling, and SMS channel configuration appear, and those specifics are narrow enough that candidates who skipped them in preparation tend to notice the gap immediately.

– Channels and communication configuration

The Sales mobile app has its own deployment and configuration path, separate from the browser-based application. The exam tests what can be configured in the mobile app specifically and what requires the full client. Teams calling setup involves connecting a telephony provider, configuring call queues, and ensuring that call data flows into Dynamics 365 for conversational intelligence to process. Configuring an SMS channel requires a provider connection and a sender number, and the exam tests the prerequisite steps rather than the provider-specific UI. These three channels share a dependency on correct user permissions and environment settings, and missing that shared dependency is a common source of errors.

– Power Platform integration and extensibility

Expanding the sales footprint with Power Platform means knowing which tool solves which problem. Power Automate flows automate processes that Dynamics 365 cannot handle natively, and the exam tests scenario matching: given a business requirement, identify whether a flow is the right tool and, broadly, what trigger and action pattern applies. Embedding Power Apps components and controls in Dynamics 365 forms extends the UI without custom code, and the exam tests when that approach is appropriate versus a native form customisation. Embedding Power BI reports brings external or aggregated analytics into the Sales experience, and items test the configuration steps and the data connection requirements. The exam rewards candidates who treat Power Platform as a deliberate architectural choice rather than a fallback.

How Extend and enhance Sales is tested

Items in this topic are almost always scenario-based requirement matching. A business need is described and you are asked to identify the right extension mechanism. The trap is that several options will describe valid Power Platform capabilities: the item rewards the one that fits the stated requirement most precisely, not the one that is technically possible. Mobile app questions tend to focus on what is supported in the mobile experience and what is not. Teams calling questions test the integration path and the data that flows from calls into Dynamics 365. Power Automate flow questions reward understanding of trigger types and the difference between automated, instant, and scheduled flows in a sales context. Candidates who have not practised under timed conditions often spend too long on Power Platform items and compress their time on the channel configuration questions, where the answers are faster to reach if the prerequisites are known.