Designing AI-Assisted Operations for an Enterprise Monitoring Platform

Role

Product Designer

Timeline

4 Weeks

Responsibilities

UX Strategy • Product Discovery • User Research • AI Experience Design • Interaction Design • Prototyping • Usability Validation

Overview

Enterprise operations teams depend on a wide range of monitoring tools to keep business-critical systems running. Dashboards, alerts, incident queues, logs and operational reports all provide valuable information, but they often exist across multiple disconnected systems.
As a result, engineers spend a significant portion of their day navigating between tools, investigating incidents manually, and piecing together information before they can make decisions.

This project explored how conversational AI could become a natural part of an existing enterprise monitoring platform—not by replacing established workflows, but by making them faster, simpler, and more intuitive.

The Platform

The platform was designed around a simple principle:

Reduce operational complexity without requiring organisations to replace their existing enterprise tools.

When I joined the project, the monitoring experience was already established. The opportunity was to explore where AI could create meaningful value within that workflow.

The platform was designed around a simple principle:

Reduce operational complexity without requiring organisations to replace their existing enterprise tools.

When I joined the project, the monitoring experience was already established. The opportunity was to explore where AI could create meaningful value within that workflow.

The Challenge

While the platform successfully consolidated operational data, finding answers still required users to manually navigate dashboards, inspect widgets, review alerts and analyse incident details before they could understand what was happening.

This created several challenges:

Investigating incidents required information from multiple screens.
Users manually searched dashboards for operational context.
Experienced engineers could interpret system behaviour quickly, while newer team members often needed additional support.
Valuable operational knowledge remained locked inside dashboards instead of being easily accessible.

The challenge wasn't a lack of data.

It was reducing the effort required to transform that data into actionable insight.

This created several challenges:

Investigating incidents required information from multiple screens.
Users manually searched dashboards for operational context.
Experienced engineers could interpret system behaviour quickly, while newer team members often needed additional support.
Valuable operational knowledge remained locked inside dashboards instead of being easily accessible.

The challenge wasn't a lack of data.

It was reducing the effort required to transform that data into actionable insight.

Understanding the Users

Rather than assuming AI was the answer, I wanted to understand how operations teams currently investigated incidents and whether conversational AI would genuinely improve their workflow.

I interviewed ten participants, including operations engineers, monitoring specialists and IT administrators who regularly worked with enterprise monitoring platforms.

The research focused on:

How incidents were investigated.
Which information users searched for most frequently.
Where investigations became slow or repetitive.
Existing familiarity with generative AI tools.

Several consistent themes emerged.

Most participants already used conversational AI tools in their daily work and expected the same interaction model inside enterprise software.

Instead of navigating multiple dashboards, they wanted to ask questions naturally, such as:

"Why did this batch job fail?"

"Summarise today's critical incidents."

"Which systems need immediate attention?"

Users also made one expectation very clear:

They didn't want another chatbot.

They wanted AI that understood the operational context of the platform they were already using. This insight became the guiding principle for every design decision that followed.

Rather than assuming AI was the answer, I wanted to understand how operations teams currently investigated incidents and whether conversational AI would genuinely improve their workflow.

I interviewed ten participants, including operations engineers, monitoring specialists and IT administrators who regularly worked with enterprise monitoring platforms.

The research focused on:

How incidents were investigated.
Which information users searched for most frequently.
Where investigations became slow or repetitive.
Existing familiarity with generative AI tools.

Several consistent themes emerged.

Most participants already used conversational AI tools in their daily work and expected the same interaction model inside enterprise software.

Instead of navigating multiple dashboards, they wanted to ask questions naturally, such as:

"Why did this batch job fail?"

"Summarise today's critical incidents."

"Which systems need immediate attention?"

Users also made one expectation very clear:

They didn't want another chatbot.

They wanted AI that understood the operational context of the platform they were already using. This insight became the guiding principle for every design decision that followed.

Designing AI Around Existing Workflows

Because users already relied on dashboards throughout their day, introducing an entirely separate interface would force them to change established behaviours.

Instead, I designed AI as an extension of the monitoring workflow.

Users could switch into an AI workspace directly from the platform, maintain the context of the dashboard they were viewing, and ask questions about incidents, alerts or operational data without losing their place.

This approach reduced context switching while making AI feel like another tool within the platform rather than a completely new product to learn.

Because users already relied on dashboards throughout their day, introducing an entirely separate interface would force them to change established behaviours.

Instead, I designed AI as an extension of the monitoring workflow.

This approach reduced context switching while making AI feel like another tool within the platform rather than a completely new product to learn.

Designing the Shift Handover Experience

The solution transformed AI from an information assistant into an operational assistant.

Instead of simply answering questions, AI could now analyse operational activity across the user's shift and generate a structured handover summary.

Before sharing the report, users could:

Select the incoming engineer.
Choose which operational data to include.
Review the generated summary.
Edit the content where necessary.
Share a consistent, AI-generated handover report.

This ensured engineers remained in control while eliminating much of the repetitive documentation work.

The solution transformed AI from an information assistant into an operational assistant.

Instead of simply answering questions, AI could now analyse operational activity across the user's shift and generate a structured handover summary.

Before sharing the report, users could:

Select the incoming engineer.
Choose which operational data to include.
Review the generated summary.
Edit the content where necessary.
Share a consistent, AI-generated handover report.

This ensured engineers remained in control while eliminating much of the repetitive documentation work.

Website Development for Nuvanta Financial Advisory ›

Got an innovative idea? Let's talk

Let's talk

Behance

Dribbble

Latifat Akinyemi

2025

Got an innovative idea? Let's talk

Let's talk

Behance

Dribbble

Latifat Akinyemi

2025

Got an innovative idea? Let's talk

Let's talk

Behance

Dribbble

Latifat Akinyemi

2025