Free AB-250 practice test: 30 questions with answers

A chat channel plus an embedded widget on the website provides the customer-facing entry point. Emailing HTML, printed QR codes, and social-only publishing do not provide the web chat experience.

Proactive chat uses triggers based on page behaviour to invite visitors. Email invitations after the visit, printed pop-ups, and surveys afterwards do not engage the visitor in the moment.

Pre-chat surveys collect information before the chat is routed, which helps routing and preparation. Post-call surveys happen after, printed forms are not digital intake, and voice recordings are not pre-chat collection.

Real-time translation supports conversations across languages. A printed dictionary, separate channels without translation, and ignoring languages do not provide live translation.

Sensitive data masking protects card numbers in transcripts while preserving recordings for review. Deleting all recordings loses quality data, and unprotected or public transcripts expose sensitive information.

Copilot summaries condense conversations so representatives can take over quickly. Printed transcripts are slow to read, calendar reminders do not summarize, and random generators are not useful.

A voice agent with NLU and a defined handover handles self-service and escalates appropriately. Keypad-only bots, closed-office recordings, and bots that never transfer frustrate callers.

Connecting approved knowledge and testing against known questions keeps answers accurate and grounded. Internet-only answers, hard-coded responses, and agents without knowledge increase error risk.

Agent assist provides suggestions during the live conversation. Printed FAQs, monthly training, and post-call surveys do not help while the call is happening.

The Copilot analytics report shows usage and effectiveness indicators. Manual logs, coffee orders, and photographs do not measure assistance usage.

Solutions and ALM package configuration for consistent, repeatable deployment across environments. Manual recreation, unmanaged copies, and screenshots are error-prone and not repeatable.

A least-privilege security role gives representatives what they need without administrative capabilities. Admin roles for everyone, shared logins, and no role are unsafe or non-functional.

A supervisor persona with view-level analytics access and no configuration rights meets the requirement. Full admin rights, shared passwords, and no access prevent the role from working as intended.

The Agent hub is used to manage and roll out agents and track their status. A binder, email chain, and attendance sheet do not manage rollout.

Connectors bring external data into the contact center experience in a governed way. A wallpaper, manual USB copies, and manual chat copies are not integration methods.

Queue prioritization and work classification route urgent cases ahead of routine ones. A single unprioritized queue, random selection, and name-based sorting do not reflect business priority.

Operating hours with fallback behaviour route customers to a callback or message when the center is closed. Waiting indefinitely, signs, and personal phones do not provide the intended service.

Skills-based routing matches work to representatives with the needed skills, including best-match selection. It does not send everything to one person, does not require identical skills, and still uses queues in many designs.

Conversation diagnostics show how unified routing processed the conversation, which identifies the cause. A printed flowchart, asking customers, and restarting phones do not reveal the routing decision.

Intent-based routing directs work based on identified intent so that specialists receive the right cases. First-come, random, and length-based routing ignore intent.

Application tab templates in the experience profile open standard tabs automatically for conversations. Printed lists, manual shortcuts, and reminders are not consistent or governed.

Scripts with slugs insert customer and agent details automatically, which keeps greetings consistent and personal. Blank notes, unchanged recordings, and printed cards do not fill in details.

Macros automate repeated steps in one action, which saves time and reduces variation. Manuals, customer repetition, and memory-based steps are slower and less consistent.

Internal knowledge search surfaces approved content within the workflow. Unapproved internet searches, memory alone, and printed encyclopedias are not governed or quick.

Portal integration publishes approved knowledge to customers in a managed way. Email blasts, printed mailings, and private-only storage do not provide self-service access on the website.

The embedded Power BI editor allows KPIs to be customized in the analytics reports. Paper templates, chat messages, and edited screenshots do not change the live report.

Real-time dashboards show live queue activity so supervisors can respond. Monthly reports and printed summaries are too late, and static lists do not show activity.

The Quality Evaluation agent supports structured scoring of conversations against a rubric. Random numbers, attendance sheets, and holiday lists are not quality evaluation tools.

Conversation diagnostics show the steps and decisions for a conversation, which helps diagnose routing. Calendars, wallpapers, and user photos do not show routing steps.

Customizing historical analytics in Power BI shows trends over time from the contact center data. Printing daily scores, guessing trends, and using only one month do not show the year-long pattern.