OI-Qi for Inquiries

What it's for

It adds AI-powered insights to your existing views, helping you quickly understand what's going on beneath the surface of customer interactions - not just that a metric moved, but what's actually happening in the conversations behind it.

How to use it

Accessing OI-Qi

OI-Qi is launched from the Inquiry and Sub-Inquiry layers.

Hover over the inquiry you're interested in and choose the lightbulb icon.

Running an analysis

1. Choose an analysis mode. You'll see three predefined prompts, each designed to help you understand what's affecting a key customer experience metric:

   • Customer Effort
   • Sentiment
   • Resolution

   These prompts are automatically tailored to the selected inquiry, so you can quickly understand the "why" behind changes in customer behaviour for that inquiry.

    

2. A modal opens over the dashboard and the job is queued. You can resize and drag the modal to wherever suits you.

Things to know when running

• Minimising the modal doesn't stop the job - it keeps running in the background.
• You can launch several jobs at once. Each runs independently.
• OI notifies you when a job completes.

Working with the outputs

• Open any in-progress or completed job from the widget menu.
• Once a job has finished, you can have a natural language conversation with its output - ask follow-up questions, drill into specifics, or request a summary.
• Chat memory carries across multiple OI-Qi runs, so you can bring the outputs of different jobs into a single conversation and compare or combine them.
• You can download the conversation, or expand it to fill the window for easier reading.
• A context window indicator shows how much capacity your conversation has used - if it's still low, you can keep asking follow-up questions.
• To finish up, click the collapse (down) button to return to your dashboard.

Use cases

Understanding what's driving effort on a specific issue

Let's say the inquiry is "Can't log into the internet banking app." To explore what's driving customer effort for this issue:

1. Click the prompt for Customer Effort.
2. The system analyses the relevant interactions for the selected time period.
3. You'll get a clear, readable breakdown of what's driving effort on those interactions.

You can run this analysis across any of the three metrics mentioned above.

Turning the output into something you can share

Say you've run a Resolution analysis on an inquiry and OI-Qi has given you an overview of what's driving resolution and the common call types that aren't being resolved. Rather than copying the raw output, ask a follow-up question to reshape it - for example: "Rewrite this as an email for my customer care team leaders. Include detail on the call type analysed and the time frame, and split it into two sections of bullet points: what's working and what's not working."

OI-Qi returns a ready-to-send email, complete with a subject line, a summarised body, bullet points split across the two sections, any assumptions made in the analysis, and a sign-off. Copy it straight into your email client, or keep asking follow-up questions to refine it further - the context window indicator shows you how much room you've got left to do that.

Finding overlap between two metrics

Run resolution and sentiment on the same Sub-Inquiry, then ask the chat: "What scenarios have both low resolution and low sentiment?"

Comparing effort across related inquiries

Run customer effort on two related Inquiries and ask the chat to compare the main sources of effort between them.

Pro tip: filter for specific time periods. OI-Qi works best when paired with relevant time filters. For example, if you know a spike in inquiries happened during a specific outage, filter the chart to that time period first, then run your analysis. You'll get much more targeted insights.

Want to see more? 

See the Insights #19 module in OI University.