AB-410 Cheat Sheet: Task-to-Portal-Path Reference
Not organized by domain — organized by what you're trying to do, so you can find the answer in the last hour before the exam without hunting through prose.
Exam Snapshot
Sourced from Microsoft's AB-410 certification page and study guide, reviewed September 7, 2026. MSCertQuiz sells a 500-question AB-410 practice bank; this sheet was written by the same team and states only published facts.
If You Need To… Build the Data Foundation
| Task | Where in the maker portal | Quick note |
|---|---|---|
| Create or edit a Dataverse table | make.powerapps.com → Tables → + New table | Choose "standard" unless the scenario needs an elastic or activity table type. |
| Add a prompt column | Table → Columns → + New column → AI → Prompt | Runs an AI Hub prompt against the row's own field values on save/refresh — for AI-generated field content, not user-facing chat. |
| Set up a table relationship | Table → Relationships → + New relationship | One-to-many vs. many-to-many changes which side owns cascading delete/assign behavior — read the scenario for who "owns" the relationship. |
| Configure a row summary | Table → Row summaries → + New row summary | Produces an AI-generated, read-only recap of a record — different feature from a prompt column, which writes to a field you define. |
| Lock down who sees which rows | Environment settings → Users + permissions → Security roles | Table-level security is where Domain 1's "Configure security" objective lives — think privilege depth (User/Business Unit/Org), not just table permissions. |
| Package work for ALM | make.powerapps.com → Solutions → + New solution | Everything you build for the exam scenario — tables, apps, flows, prompts — should conceptually live inside a solution, not the default/unmanaged layer. |
If You Need To… Build the App Experience
| Task | Where in the maker portal | Quick note |
|---|---|---|
| Generate a page from a description | App designer → Pages → + New → Describe what you want | This is the "generative pages by using natural language" objective — separate from manually building a form or view. |
| Control who sees a form or view | App designer → component → Access | Distinct from table-level security roles — this controls app-surface visibility, not row-level data access. |
| Reuse logic across canvas apps | Canvas app → Insert → Component library | Named formulas replace app-open OnStart variable assignments; user-defined functions replace repeated formula logic — know when each applies. |
| Catch a failed data call gracefully | Canvas app formula → IfError / IsError | Exam scenarios describing "the app should keep working if the connector times out" point to explicit error handling, not a Notify() afterthought. |
| Validate app behavior before shipping | Canvas app → File → Monitor | Monitor traces formula evaluation and network calls live — the tool named explicitly in the "test canvas apps" objective. |
| Turn a canvas app into an agent | Canvas app → Create a Copilot Studio agent | A one-way handoff from app to agent, called out as its own bullet under canvas apps — not the same as embedding an existing agent inside an app. |
If You Need To… Automate and Add AI
| Task | Where in the maker portal | Quick note |
|---|---|---|
| Build a cloud flow trigger | Automate → Cloud flows → + New flow | "Recommend cloud flow triggers" is its own tested bullet — automated (event) vs. instant (manual) vs. scheduled changes the rest of the design. |
| Route a request for sign-off | Flow designer → + Add action → Approvals | First-choice vs. everyone-must-approve changes both the action used and the resulting flow branch logic. |
| Build a reusable AI prompt | AI Hub → Prompts → + New prompt | Start from a template or blank; add knowledge sources and inputs here before consuming the prompt elsewhere. |
| Call that prompt from a flow | Flow designer → + Add action → AI Builder / AI Hub prompt | "Consume a prompt in cloud flows" is listed separately from consuming it in apps — the action and its inputs differ by surface. |
| Swap or tune the underlying model | AI Hub → Prompts → Prompt settings → Model | Model selection lives inside prompt settings, not as a separate top-level configuration step. |
| Enforce a rule regardless of the app used | Table → Business rules → + New business rule | Business rules run at the table level across every app surface; canvas app error handling and business process flows do not. |
If the Question Says X, Think Y
| The scenario says… | Think… |
|---|---|
| "Generate a field value automatically using AI, from other fields on the same record" | Prompt column, not a calculated column |
| "Let a maker build a page just by describing it" | Generative pages, not manual form/view design |
| "The same instruction/prompt needs to run from three different apps" | A reusable AI Hub prompt, not three separate raw model calls |
| "Logic must apply no matter which app or interface touches the record" | A business rule at the table level, not app-specific formulas |
| "Guide the user through a required multi-step sequence" | A business process flow, not a canvas app screen sequence |
| "The app should stay usable if a connector call fails" | Explicit error handling (IfError), not default behavior |
Acronyms & Terms
| Term | What it means on this exam |
|---|---|
| ALM | Application lifecycle management — solutions, environments, and pipelines used to move work from dev to production. |
| AI Hub | The Power Platform area for building and managing reusable AI prompts and models, consumed by apps and flows. |
| UDF | User-defined function — a named, reusable Power Fx function you write once and call from multiple places in a canvas app. |
| Prompt column | A Dataverse column type that runs an AI Hub prompt against a record's own data to populate its value. |
| Row summary | A read-only, AI-generated recap of a Dataverse record, configured per table. |
| BPF | Business process flow — a guided, stage-based sequence enforced at the table level, independent of which app opens the record. |
| Power Fx | The low-code formula language behind canvas apps, named formulas, and user-defined functions. |
AB-410 Quick Answers
How is this cheat sheet different from the AB-410 study guide?
The study guide explains a week-by-week plan and where the exam puts its weight, in narrative form. This page skips the narrative for portal-path tables you scan the morning of the exam.
Where is AI Hub if I can't find it in my environment?
AI Hub availability depends on your environment's region and licensing. If it's not visible in the maker portal navigation, check environment settings or your tenant's Power Platform admin center rollout status before assuming it's missing entirely.
Do I need to memorize exact menu paths for the exam?
No — Microsoft tests whether you know which feature solves a scenario, not pixel-exact navigation. The paths above exist so you recognize the feature by name when a question describes its behavior.
Can I print this AB-410 cheat sheet?
Yes — every table here is plain HTML, so a browser print or "print to PDF" renders cleanly with no layout breakage.
Keep Studying
The week-by-week plan and domain-weight breakdown behind these tables.
Scenario questions by difficulty, with full rationale for every choice.
A free timed mock exam across all three domains.
Full domain list and exam facts in one place.
Tables reviewed — pressure-test them next.
Start with 40 free AB-410 questions covering every table above.
Start Practicing AB-410