Driving 45% weekly active usage with AI Performance Insights

Valuelink Software · Shipped · 2025

0%

Weekly Active Usage

0%

Increase in Company Revenue

200K+

Orders Processed through Cogent

Problem

Appraisal Management Companies struggle to quickly analyze their performance and translate data into actionable decisions. Most users jumped between disconnected dashboards and rarely explored trends beyond top-level stats, leading to time wasted and missed opportunities.

Solution

I designed an AI Performance Analytics experience called Cogent which was embedded into our Core app, to turn overwhelming appraisal data into clear, actionable decisions.

Role

Lead Product Designer

Responsibilities

Design product from 0 → 1

Collaborators

1 Product Manager, 1 Product Designer (Me), Software Engineers, QAs, Marketing

Timeline

December 2025

Things worth knowing about this project

How I contributed to the codebase

I shipped features end-to-end, opened PRs that went through engineer review before merging, and wrote production code directly. I used Claude Code as part of my everyday workflow.

The constraints we were designing within

I was working in a high-constraint embedded environment. Our UI had to live inside two external platforms, so I had to deeply understand their individual limitations and make sure what I built felt seamless within both.

My Role

From confusion to clarity with AI-driven insights

Core is a platform used by AMCs to manage appraisal workflows. To understand their performance, users had to dig through long, detailed pages across the website to piece together bits of information, which they weren't equipped to analyze.

I designed an AI Performance Analytics dashboard embedded within an existing AMC platform called Core by Valuelink, giving AMCs real-time visibility into their financial and operational health. I built it across three analytical lenses with three separate dashboards: Revenue, Turn Time, and Revision. This case study will focus on Revenue only.

Business goals

  • Grow revenue by 30%.
  • Enable measurable weekly active usage.

Problem

Cogent risked becoming a data-heavy product without clarity on which insights actually mattered to AMCs.

Solution

I worked directly with AMCs to identify the decisions they make daily and the bottlenecks they face. I focused on shipping quickly and iterating fast on feedback.

Research

Turning user pain points into validated product metrics

I led stakeholder meetings and workshops with our users to identify user pain points, define the most meaningful metrics, and prioritize what users want to see. These points surfaced:

User behavior

  • Jumped between dashboards
  • Rarely explored trends beyond top-level stats
  • Took notes manually

User pain points

  • Data was overwhelming and hard to interpret
  • Insights weren't actionable. Users did not know what to do next
  • Users lacked the confidence to make data-driven decisions
Org chart from validation workshop

AMC Organizational Chart

User personas

User Pain Points and Motivations

Competitors

Competitive analysis

Understanding competitor capabilities helped identify gaps in advanced performance intelligence and actionable analytics across the appraisal lifecycle.

Reggora

Insights primarily center around process performance rather than deeper benchmarking or predictive business intelligence.

ValuTrac

Analytics are largely operational and rule-based rather than delivering strategic, cross-workflow performance insights.

Anow

Analytics focus mainly on workflow visibility and productivity monitoring rather than holistic performance benchmarking or root-cause driven insights.

Reggora dashboard
ValuTrac dashboard
Anow dashboard

Launch

Advocating for an iterative launch

I disagreed with my PM on launching the entire feature set (Revenue, Turn time, Revision) at once. I advocated for an incremental rollout instead, prioritizing validation, faster feedback loops, and continuous improvement.

We settled on releasing the Revenue dashboard first, monitoring for feedback before expanding further. The incremental approach proved its value early, surfacing issues that shaped the design decisions covered below.

Design

Rejected designs

These are the designs I explored and rejected.

AI Insights light mode rejected design
AI Insights old design rejected

The final design flow

Cogent AI Performance Dashboard.

Final Design - Dark Mode

Dark Mode

Cogent Preview 1

Revenue Dashboard

Final Design - Light Mode

Light Mode

Cogent Preview 2

AI Insights

Final Design

Final Design

Problem

Users weren't accepting terms and conditions

...which led to another problem: PM insisted on making T&C obscure

Product Manager

Initially we had very low T&C acceptance. My PM insisted on loopholes such as making the CTA vague or making the link to terms unobvious.

Me

I insisted that we need to protect ourselves from a legal standpoint if we are to suggest decisions to users. My goal was to find out why users were not accepting, while also increase the number of users. Based on my industry knowledge, its possible that not all users have the authority to accept T&Cs and need a decision maker to do so, its also possible that the reason was entirely something else. My goal was to find out what the issue was but I did not want to wait on one or the other and cause delays. So I built two solutions in parallel:

1. Give decision makers the authority to accept conditions for their organization, and Cogent becomes available immediately for everyone.

2. If decision maker has not interacted with the T&C, user can still see the conditions on their own screen when they visit Cogent and accept it if they are allowed, if not, they can provide a reason for rejecting.

Onboarding

Performance dashboards in Cogent surface sensitive competitive data including operational metrics, fees, and margin analysis. T&C acceptance was required to establish explicit consent around data ownership, usage rights, and how comparative insights are generated across both AMC and lender user bases.

Final flow

  1. Since Cogent is embedded into Core, all users are already onboarded to Core.
  2. I identified organizational decision-makers already active on Core.
  3. I built a targeted consent flow with custom T&C modal for decision-makers only.
  4. Single acceptance unlocked dashboard access for entire organization.
  5. If decision-maker hadn't arrived at T&Cs yet, individual users can still accept on their end.

Terms & conditions according to roles

T&C for Decision Makers

T&C for Decision Makers

T&C for Other Users

T&C for Other Users

Terms & conditions acceptance journey

Role and Uses

Role and Uses

Guided Tour

Guided Tour

Terms & conditions decline journey

Reason for Declining

Reason for Declining

Get In Touch

Get In Touch

Results

Results

>0%

Weekly Active Usage

Achieved >45% weekly active usage across 200+ lenders within 2 months by leading the design of Cogent's AI-driven Revenue dashboard.

0%

Increase in Company Revenue

Contributed to a 30% increase in company revenue by launching data visualization features that improved client retention and engagement.

200K

Orders Processed

Total orders processed through Cogent

Feedback from our largest clients

Absolute Lending Logistics

"You know, as for the revenue increase and decrease, you know, we track that pretty much hourly anyways on what we're doing. You know, obviously with a small business like us, it's very important to maintain margin. So we're always on top of that. You know, your map of appraisal fee and the client fee graph, those are phenomenal. You know, so definitely, definitely appreciate everything that you guys are doing here."

1st Signature Landing

"It does all my monthly work for me."

LVS (Ascribe)

"Can dashboards be used as HomePage for internal users?"

Single Source

"Our users really like it."

Growing interest in Turn time and Revision dashboard after we shipped Revenue dashboard

"You know, because of this, I'd be excited about the turn time and the revisions. The revisions is going to be pretty huge. That's something that we track internally for our staff appraisers. We keep track on a broad spectrum, not as detailed for our contractors. So I'll be interested to see what you guys have coming out for that."

Absolute Lending Logistics

Tradeoffs

Tradeoffs and how I plan to tackle them

Why I chose a modal over a side panel

I chose a modal for the AI Insights view because the embedded platform environment imposed strict layout constraints. There was limited horizontal space to introduce a persistent side panel without disrupting the existing Direct interface. A modal allowed me to deliver a rich insight experience without touching the surrounding platform layout. The tradeoff is that the modal obscures the dashboard behind it.

View more work