Capitalise.ai
AGAGAdam Goddenyu

Intent-based trading: driving growth and performance through behavioral design.

Redesigned the trading automation wizard by aligning syntax logic with user intent, resolving critical UX friction and driving activation and growth.

Intent-based trading: driving growth and performance through behavioral design.
Company
Capitalise.ai icon

Capitalise.ai

A B2B2C, no-code, text-based trading automation & analysis platform, founded to democratize trading automation and made accessible for traders without coding skills (years before the LLM boom).

Capitalise.ai trading automation platform dashboard
RoleSole Product Designer
IndustryFintech / B2B2C
TeamCPO / PM / 3 BE / 1 FE
Customers
Context

Redesigning the core automation wizard to drive trader activation.

The product strategy was designed to secure top-tier broker partnerships, a goal it successfully achieved. While this resource allocation was understandable, it came at the expense of the end-user experience.

2% 31%16x Increase

Increased core user flow task success rate.

0.12% 5.88%49x Increase

Raised average trader profit.

~$10k $400K+40x Increase

Total trading volume.

The Problem

The focus was on the brokers, not the traders.

Trading volume stalled at ~$10K against a ~$200K expectancy. The product's core flow of running live automated strategies was flawed. Traders weren't actually trading.

The trading automation wizard: type a trading strategy in plain English and run. Designed to feel effortless.
The trading automation wizard — entry strategy input with Buy, Sell, If, At tokens
UX Process

Isolating the friction point, where the wizard is failing.

Mixpanel setup was misconfigured, I audited and implemented relevant tracking paths, following up with in-person usability testings.

Core user flow event tracking

Mixpanel~1 month
Mixpanel funnel — 90.5% drop-off at Execute live strategy step

Usability testing - Main learnings

In-person

5/5 Participants:

Were impressed by the concept of simplicity.
Struggled to express their ideas using the suggested commands, trying to type, deleting, and retrying.
Began motivated but ended frustrated: 4 didn't finish, only 1 executed a live strategy.
Never automated a trading strategy before.
Hypothesis

It was English, but it wasn't natural language.

Providing clear, actionable contextual guidance will help traders confidently execute live automated strategies, increasing trading volume and reducing churn.

Solution

Aligning the syntax logic with user-intent.

Transformed the suggestion system to contextual, full sentences paired with secondary helper text to guide traders as they write. Added indicative & adaptive helper text for all user states.

01

Auto suggestion, completion & prediction

Guiding the traders from the very first encounter with the interface.

Auto suggestion dropdown showing contextual full-sentence completions
02

Smart textual indicator

Adaptive helper text, that clarifies to the trader how to progress.

Smart textual indicator — adaptive helper text guiding the trader
03

Informative error handling

Allowing to quickly fix mistakes and teaching the appropriate syntax.

Informative error handling — symbol correction with BTC, BTC/USD, BTC/ETH suggestions
04

Improved edit mode

Adding suggestions to replace, and better experience on selecting an asset to edit.

Improved edit mode — BTC/USD replacement suggestions panel

Personal note.

This project was a definitive moment in my career. It's one thing to design a good product, it's another to see your strategy become a primary engine for the business.

Capitalise.ai app screenshot