I turned a manual consulting service into an AI-assisted platform, with one developer.

I turned a manual consulting service into an AI-assisted platform, with one developer.

Alioth’s organizational health surveys had 70 clients and an NPS of 100, all delivered by hand. I designed the platform to scale it to hundreds.

Alioth’s organizational health surveys had 70 clients and an NPS of 100, all delivered by hand. I designed the platform to scale it to hundreds.

OrgDx results dashboard with filters and demographic comparisons

Company

Alioth

Alioth

Team

VP of Product × 1

Product Manager × 1

Full-stack Engineer × 1

My Role

Founding Designer

Founding Designer

Responsibilities

Product Design, User Research, Design Systems

Product Design, User Research, Design Systems

Tools

Figma

Figma

Platforms

Web

Web

Year

2019–2024

2019–2024

Situation

Alioth’s OrgDx consulting had 70 clients and an NPS of 100, but it ran on manual work: SurveyMonkey surveys, spreadsheet analysis, and reports built in InDesign. Each engagement took weeks.

Task

Design the platform that would scale OrgDx from 70 clients to hundreds, automating analysis without losing the expert quality clients valued. One developer, limited runway.

Result

  • Internal analyst tool shipped and was adopted immediately

  • AI ran sentiment analysis and flagged statistically significant demographic differences

  • Customer platform fully designed; the company closed before launch

Strategy

  • Phased the delivery: internal analyst tool, then customer self-service, then AI observations with analyst approval.

  • Automated the mechanical work first, freeing analysts for the hard part: synthesizing open-text responses.

  • Kept what already worked. Survey authoring stayed in SurveyMonkey rather than being rebuilt.

Lo-fi sketches of survey score roll-ups

Lo-fi exploration of how survey scores could roll up before we committed to the build.

Design

  • Designed AI to support experts. Analysts could edit, approve, or reject AI-generated insights before clients saw them.

  • Made HR leaders feel like data scientists with filtering that compares any segment against the company.

  • Built a design system so a one-developer team could iterate quickly.

Analysis complete view with AI-generated observations
Roll-up view of cultural attributes
Survey running view with response progress

Final designs across the survey lifecycle: results, roll-ups, and a survey in progress.

Learnings

Augment experts, don’t replace them. Automating the mechanical work first kept insight quality high, and the analyst-approval pattern for AI still holds up. With three people running two products, tighter focus would have gotten us to market sooner.

More case studies

I gave recruiters hours back by turning dry job descriptions into pitches passive candidates would actually read.

HR Tech

0 to 1 Design

B2B

Web

Read more

I gave recruiters hours back by turning dry job descriptions into pitches passive candidates would actually read.

HR Tech

0 to 1 Design

B2B

Web

Read more

Nursegrid colleague schedule screen

I let research scope Nursegrid’s most-requested feature, and daily downloads grew 67%.

Healthcare

Feature Design

B2C

iOS/Android

Read more

Nursegrid colleague schedule screen

I let research scope Nursegrid’s most-requested feature, and daily downloads grew 67%.

Healthcare

Feature Design

B2C

iOS/Android

Read more