info@ovizen.com
info@ovizen.com

Work /

Scaling Insight Creation with AI Summaries

Embedding intelligent insight generation directly into the reporting workflow.

gray concrete wall inside building
gray concrete wall inside building

CONFIDANCE NOTICE

This case study contains information from work completed under non-disclosure agreements. Sensitive details have been modified or omitted to respect confidentiality obligations. The content represents my personal analysis and work contributions, and does not necessarily reflect the views or positions of Whatagraph.

As agencies scaled their client portfolios, performance summaries became an increasingly important part of reporting. While data visualization was already automated, interpretation and written insight creation still required repetitive manual effort.

INTRODUCTION

This initiative introduced AI-powered summary generation directly into the reporting workflow — transforming a manual reporting task into a more scalable and intelligent product capability.

gray concrete wall inside building
gray concrete wall inside building
white and black abstract painting
white and black abstract painting
Monthly reports required account managers to review performance across channels, identify meaningful changes, and turn them into structured narratives for clients.

CONTEXT

Even when the insights themselves were straightforward, writing them still took time and repeated mental effort.

At scale, that created real operational overhead. Across 20+ clients, teams were repeatedly spending 10–15 minutes per report on a task that was valuable, but often repetitive.

IMPACT

Increased reporting efficiency and reduced the time spent creating manual performance summaries.
red blue and black abstract painting

MY ROLE

I led the design of the AI summary experience — defining how prompt-based generation, tone options, and language controls should work within real reporting workflows. I also shaped how AI generation would integrate directly into the existing text widget so it felt like a natural extension of the workflow, not a separate tool.

Working closely with product and engineering, I translated the concept into a scalable and intuitive solution ready for development.

CHALLANGE

Agencies needed to deliver thoughtful, high-quality performance insights, but doing it manually for every report slowed teams down. The challenge was not only reducing the time spent writing summaries, but introducing AI in a way that still felt useful, trustworthy, and easy to control inside an existing workflow.

gray concrete wall inside building
gray concrete wall inside building

OBJECTIVES

Integrate AI insight generation seamlessly into existing reporting workflows
Ensure generated summaries are accurate, editable, and aligned with user intent
Reduce manual effort in performance interpretation

01.

02.

03.

The opportunity was to make summary creation faster without disrupting how teams already worked — combining automation with review, editing, and control so the experience could support both efficiency and trust.

An Embedded Intelligence Layer

To deliver on that opportunity, I introduced intelligence across three focused layers already reflected in the solution.

Native Workflow Integration

AI generation was integrated into the existing text widget, making adoption more natural and removing the need to introduce a separate reporting tool.

Context-Aware Analysis

The system was designed to analyze the active report context, including data and date range, and generate structured summaries based on configurable focus areas such as wins, issues, or opportunities.

Controlled Automation

Preview, editing, tone and length adjustments, language selection, and optional auto-updates helped balance automation with user trust and practical control.

gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building

01.

02.

03.

A guided start made recurring reporting tasks faster to complete and easier to scale
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building

Instead of writing performance narratives from scratch, users could begin with prompt-based generation tied to the report context and quickly create a first draft that was already structured around the right type of insight.

To reduce effort in the most common reporting scenarios, I introduced a more guided starting point for summary generation.

Automation alone would not be enough for reporting teams. The experience also needed to support review, steering, and refinement before a summary became final.
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building

I designed the workflow so users could preview results, adjust the focus of the output, change tone, length, or language, and turn generated content into fully editable text. This kept users in control while still reducing repetitive effort.

Beyond faster generation, the experience also introduced a more dynamic way to keep reporting narratives aligned with changing data.
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
gray concrete wall inside building
Because summaries were tied to the report’s date range, the system could update content as reporting periods changed. Reducing the need to manually rewrite recurring insights and creating a stronger foundation for ongoing intelligent assistance.
gray concrete wall inside building
gray concrete wall inside building

AI SUMMARY INTEGRATION

AI Summary integrates directly into the existing text widget, enabling users to generate structured performance insights in seconds. With a built-in preview, teams can review results before applying them, refine prompts based on focus, adjust tone, length, or language, and convert the output into fully editable text.

Because the summaries are tied to the report’s date range, they can also update as reporting periods change — reducing repetitive manual rewrites while maintaining quality and consistency at scale.

Turning reporting data into clearer insight, with speed, control, and consistency built in.

OVERALL IMPACT

Reduced insight creation time from up to 15 minutes per report to seconds, saving teams an estimated 6+ hours monthly. The solution also improved consistency across reporting, reduced repetitive manual effort, and helped position the product as a more intelligent decision-support tool.

6+ hours

62%

From 15 Minutes to Seconds per Report

Reduced manual effort required to generate performance summaries, turning a repetitive task into an instant workflow.

Summaries Used by Active Users

More than half of active users engaged with AI-powered capabilities, validating strong product adoption and impact.

Let's talk. I'd love to hear from you.

Contact

Ovidius Avizienus Product Design Portfolio

2026