AMPLYFI Intelligence Feeds
Elevating UX in AI-Powered Market Intelligence
Design Sprint
NLP & Boolean Interfaces
UX Strategy
Context
At AMPLYFI, I joined as the sole Product Designer with a key goal: improve user adoption and reduce dependency on manual onboarding from the Customer Success team. One of the biggest friction points was Intelligence Feeds, a feature designed to help users track the topics, trends, and signals that matter to them.
Although promising in theory, the existing experience was underused and frequently abandoned. My task was to figure out why, and fix it.
The Problem
Despite its importance, Intelligence Feeds suffered from poor engagement. From interviews and platform analytics, I identified three core issues:
Time to Value was Too High After creating a feed, users had to wait far too long to see results — creating uncertainty and frustration.
Opaque UX and Jargon Users didn’t understand the AMP-specific terminology required to configure a feed. There was no preview or feedback, leaving them disconnected from the system’s outputs.
Heavy Reliance on CS Support Sales calls showed strong interest in the feature, but without support from CS, most users avoided it altogether.
"I don’t know what the results will look like until I hit publish — and by then, I’m not even sure I set it up right"
In my first few weeks, I prioritised conversations with Sales and CS teams. I wanted to understand what they were doing behind the scenes to help users and why it was necessary. This gave me a clear view of the manual workarounds they’d built to patch gaps in the product. I also reviewed analytics and onboarding drop-off points, which validated the friction around set-up and system feedback.
Design Sprint & Collaboration
To quickly align the team and understand technical constraints, I ran a Design Sprint. It was early in my time at AMPLYFI, and this was a powerful way to:
Bring engineering, data science, and product together
Understand what was technically causing the delay
Frame the problem through a user-first lens
The Solution
We designed a dual-path experience:
Boolean Query Builder
For advanced users, this allowed full control with custom syntax.Natural Language Interface (NLP Assistant)
For everyone else, we introduced a conversational UI. Users could simply describe what they wanted to monitor, and AMPLYFI Explore would translate that into a boolean query in real time.
This meant users could now:
Get immediate feedback on what their query would return
Edit or refine their logic with guidance
Preview results before committing
Testing & Iteration
On day 4 of the sprint, I prototyped the new workflow. On day 5, I tested it with users.
The response was immediate:
“This is exactly what I thought the platform would be.”
“When can I use this in production?”
This energy translated directly into momentum and we moved straight into build. I worked alongside engineers to integrate user feedback from testing into the live product and ran continuous feedback loops throughout development.
Implementation & Rollout
To support this shift, we also needed to rework the back end. The team rewrote how data was processed in the lake to allow immediate, partial previews of matching content — giving users instant validation that their inputs were working.
I partnered with engineers throughout to:
Define progressive loading states
Ensure semantic mapping of NLP outputs
Retain full control for power users without overwhelming new users
Results & Impact
After launch, we saw significant improvements for this feature:
42% reduction in onboarding support tickets related to Intelligence Feeds
Time to first feed creation dropped by 60%
User satisfaction scores for the feature increased from 2.1 to 4.3
What i'm Proud of
Turning one of the platform’s most frustrating features into one of the most promising
Leading a sprint that aligned technical, business, and user needs in my first month
Introducing a dual-interface model that met both expert and novice users where they were
What's next
While Intelligence Feeds continue to evolve through user feedback and continuous optimisation, my focus has broadened to shaping AMPLYFI’s next wave of AI-powered features.
Currently exploring and designing:
Automated Reports
Creating a framework for scheduled, templated reports that translate research activity into digestible, shareable summaries, reducing cognitive load and manual synthesis for users.Generative Research Interfaces
Designing flexible UI for users to ask research questions and receive structured, traceable outputs. Blending human input with LLM-driven summarisation and clustering.Agentic Workflows
Prototyping new ways for users to delegate multi-step research tasks to autonomous agents, with clarity, context awareness, and checkpoints for human oversight.
These efforts build on the core principles established in Intelligence Feeds: make powerful tools feel approachable, explainable, and aligned with how users already work.
