Introduction
In this two-part blog series, we’ll identify the three main aspects for a successful Virtual Agent Implementation:
- Goal
- What you want to achieve
- Business Outcome
- What you want to get out of the virtual agent
- Objective
- The steps to get there
In today's blog, we're covering Business Outcome and Objective. To read part 1, click here.
Expected Business Outcome
Once you select the goal, it is important to determine what outcomes you want to achieve that support this goal. What are your Critical Factors to Success? What business value will be realized?
To assist in defining valuable and realistic business outcomes, you need to understand the capabilities of Virtual Agent, your potential to make use of them and to align them with your overall goal.
Understanding Virtual Agent Capabilities
Virtual agent and Now Intelligence use AI, machine learning, and advanced analytics to surface information, make relevant predictions and actionable recommendations, and automate repetitive, mundane, and manual error-prone tasks. These capabilities give back more time to your employees and customers so they can focus on the things they excel at — strategic and creative thinking, complex decision making, customer interactions, and unpredictable work.
Here are some more capabilities available through Virtual Agent:
- Personalized experiences. Master top-down planning to align your teams and investments with business goals.
- Conversational interfaces. Configure and deploy the look and feel of the chatbot for your organization quickly and easily.
- Virtual agent designer. Deploy pre-built conversations or build and test your own custom conversations.
- Natural language (NLU). Discern everyday language using models tuned for ServiceNow workflows, which can evolve over time.
- Issue auto resolution. Identify user intentions in submitted issues and proactively initiate a conversation for resolution.
- Conversational analytics. Gain user insight into user behavior and trends so admins can make quick, informed decisions.
- Topic recommendations. Use machine learning to identify critical topics then add suggested conversations with one click.
- Multi-language support. Detect and translate conversations dynamically into the end user’s language of choice.
- NLU workbench. Create and tune advanced no-code models. Define intents and map entities to Now Platform data easily.
- Multiple NLU providers. Use the built-in NLU engine or connect IBM Watson Assistant, Microsoft LUIS, or Google DialogFlow.
- Third-party channel integrations. Meet users where they are—on collaboration tools, portals, and apps like Slack and Microsoft Teams.
- Live agent hand-off. Transfer conversation history and context to the right agent so they can quickly resolve issues.
Determining Virtual Agent Business Outcomes
The business outcomes will be aligned with the goal and have the ability to be measured and evaluated against intended or projected results. For example, reduce cost by 10%.
For this example, we are looking at a KPI Tree to reduce the cost of our service desk.
- Business Goal: Reduce Service Support Cost
- Critical Success Factors:
- Maximize Positive User Experience (CSAT)
- Maximize Incident Resolution Efficiency (Productivity)
- Maximize Virtual Agent utilization (Adoption)
- Maximize Incident Deflection (Deflection rate)
- Various supporting metrics are shown below the Critical Success Factors in the KPI Tree.
For the full Scoping Document, visit this NowCreate article: VA Scoping Statement
Business Objectives
The business objectives are the measurable actions in the short term required to achieve the overall goal. For example, creating a topic for registering incidents. The objectives should help to achieve the business outcome that aligns with the goal.
Identifying Top Cases
Top use cases will be unique to your own organization, and this is where the value in analyzing your own instance comes into play. However, in addition to that analysis, we can look at the top use cases seen across all customers currently using VA.
Topic types
- Resolve a service bottleneck
- Enhance a complicated experience
- Frequently access KB’s & FAQ’s
For the full document, visit this community article: ITSM Pro VA Use Cases
Determining Virtual Agent Objectives
The primary tool in ServiceNow for identifying objectives in Virtual Agent is the Topic Recommendations app. It allows you to identify relevant, pre-built Virtual Agent conversation topics that you can quickly implement, as well as opportunities for creating new conversations suited for your organization. This app uses Intent Discovery to help determine the intents and corresponding topics that might be beneficial, based on incident or other task data in your organization.
A combination of both pre-built topics and custom topics will be needed. The following document, Virtual Agent Conversations will cover new topics in Virtual Agent.
After ensuring you have these plugins in place, you can run the app.
- Topic Recommendation
- ITSM Virtual Agent plugin
- NLU Workbench
- NLU Workbench – Advanced Features (Including Intent Discovery)
When you run the topic recommendation model against the incidents it produces a report that will recommend topics from existing (35-40) conversations that mapped to the incidents.
Once analysis is complete, you can see Topics that VA can handle along with User Intents that need a topic flow.
For the full document, visit this community article: Virtual Agent E2E Implementation
How do I get started?
How you can leverage the out-of-box conversations to jumpstart your Virtual Agent journey.
In ServiceNow®, there are Pre-packaged conversation topics – templates for most common IT, Customer Service Delivery scenarios to get started quickly.
You can use these setup topics out of the box, or as templates to get started and adjust or add enhancements as needed. How to take a template topic and customize it to meet your environment’s needs will be covered in detail in the Topics and Conversations Document.
For the full PDF, visit this community article: Getting Started
