Back to case studies

MarTech

Conversational AI Chatbot SaaS for Perceptional

Replaced static surveys with intelligent conversational bots - 4x deeper product insights, 48-hour time to useful data, concept to launch in 12 weeks.

We built Perceptional for a former Amazon Product Manager who was overwhelmed by static surveys and spreadsheets. The platform replaces one-size-fits-all questionnaires with AI-powered conversational bots that adapt in real time based on user input - delivering multi-intent support, ready-to-use summaries, and a customizable interface for product managers to set goals and tune bot behavior. Working prototype validated in under one week, full SaaS launched in 12 weeks.

Start a similar projectUpdated Mar 2026

Client

Perceptional

Industry

MarTech

Timeline

12 weeks

Team Size

5 engineers + 1 PM

Impact

Measurable results

12 weeks

Delivery timeline

4x deeper

Insight depth

48 hours

Time to insights

< 1 week

Prototype validation

Delivery timeline: From concept to production SaaS launch, including a working prototype validated in under one week.

Insight depth: Conversational bots surface four times more useful product insights than traditional static surveys.

Time to insights: Automated analysis delivers ready-to-use summaries within 48 hours of conversation completion.

Prototype validation: Working prototype built and validated with the founder in under one week to confirm product-market alignment.

I spent years at Amazon fighting static surveys. 1Raft built a working prototype in four days that already outperformed every survey tool I'd used. Twelve weeks later, we had a full SaaS that product teams actually want to use.

Founder

Perceptional (ex-Amazon PM)

The Challenge

What we were up against

The founder, a former Amazon Product Manager, spent years drowning in static surveys and spreadsheets. Traditional survey tools produced shallow, pre-structured responses that missed the specific feedback product teams actually need to make decisions.

Existing tools offered no multi-intent support - a single survey could only chase one line of questioning. When users surfaced unexpected insights, the rigid format had no way to follow up or probe deeper in real time.

High-volume feedback was impossible to analyze manually. Long unstructured responses piled up in spreadsheets, and extracting useful patterns from hundreds of open-text answers took days of analyst time - by which point the insights were stale.

Survey customization was limited to question order and branching logic. Product managers had no way to set conversational goals, adjust bot behavior based on product stage, or tailor the experience to different user segments without engineering support.

What We Built

Our approach

1
Step 1

Built AI-powered conversational bots that dynamically...

Built AI-powered conversational bots that dynamically adjust questions based on user input - probing deeper when responses signal valuable threads and pivoting when topics are exhausted, producing richer data than any static questionnaire.

2
Step 2

Implemented multi-intent support so a single conversation...

Implemented multi-intent support so a single conversation can explore product discovery, feature feedback, and user pain points simultaneously, following the natural flow of how people actually talk about products.

3
Step 3

Developed automated analysis that processes conversations...

Developed automated analysis that processes conversations into ready-to-use summaries within 48 hours - extracting themes, sentiment, and feature requests without manual spreadsheet work.

4
Step 4

Created a self-service interface where product managers set...

Created a self-service interface where product managers set conversational goals, adjust bot personality and probing depth, define user segments, and review results - no engineering tickets required.

Tech Stack

ReactJSAWS Serverless LambdaPostgreSQLClaude (Anthropic)AWS Bedrock

Related Work

Frequently asked questions about this project

Static surveys ask the same questions in the same order regardless of answers. Conversational bots adapt in real time - when a user mentions an unexpected pain point, the bot probes deeper. When a topic is exhausted, it pivots. This natural conversation flow surfaces insights that rigid questionnaires miss entirely, because users reveal more when they feel heard.

Next Step

Ready to build something similar?

One call with a founder. No sales team, no follow-up sequence. If we can't help, we'll say so.