This project focused on redesigning the “New Sequence” experience for a B2B sales platform that helps startups automate outbound outreach using AI.
The core workflow allows sales teams to create multi-step email and LinkedIn sequences. The challenge was to make the creation process faster, clearer, and more aligned with an AI-first product vision.
My Role
Product Designer, UX/UI
End-to-end UX (problem framing → flow → high-fidelity UI)
Context
Design Challenge based on existing screen.
Timeline
1 day
The Problem
Sales teams rely heavily on outbound sequences, but creating them often feels overwhelming.
The existing experience:
• Presented too many decisions upfront
• Required manual setup before seeing value
• Positioned AI as an optional assistant instead of a core collaborator
This slowed down activation and increased friction in a daily-use workflow.
​
How might we help sales teams create and launch a high-quality outbound sequence in minutes, while keeping AI transparent and controllable?
The current “Create New Sequence” screen is functionally complete but cognitively heavy. It relies too much on manual input, treats AI as an optional assistant, and lacks clear structure and confidence-building cues. As a result, it slows down sequence creation and weakens the first value proposition.
The Audit
A review of the existing flow surfaced three primary issues:
-
Cognitive overload at the start
-
AI embedded as a side feature rather than a guiding system
-
No clear validation before launch
While functionally complete, the flow lacked clarity and confidence-building cues.
What works
-
The screen covers all required inputs in a single place (name, goal, recipients, steps, settings), reducing navigation overhead.
-
Familiar form-based structure lowers the learning curve for first-time users.
-
Inline personalization tokens hint at best practices for outbound messaging.
-
Clear primary CTA (“Create & Start Sequence”) reinforces the end goal.
Key Issues & Gaps
-
Overloaded First Impression
The page presents too many decisions at once (goal, recipients, steps, settings).-
Users must think about everything before seeing any meaningful output.
-
High cognitive load, especially for first-time or occasional users (founders).
-
-
AI Assistant Is Passive and Detached
The AI Assistant lives in a side panel, disconnected from the core workflow.-
It requires users to proactively ask for help instead of guiding them.
-
No clear relationship between user inputs (goal, recipients) and AI behavior.
-
-
Lack of Clear Sequence Structure
“Sequence Steps” start as an empty manual editor.-
Users don’t get an upfront sense of what a “good” sequence looks like (number of steps, spacing, channels).
-
AI does not proactively propose a structure or timeline.
-
-
Manual-First, AI-Second Flow
Users must fill fields before AI meaningfully contributes.-
This contradicts Flamey’s value proposition as an agentic, AI-first tool.
-
AI feels like a helper, not a creative partner.
-
-
Recipients Selection Too Early
Selecting recipients before the sequence exists forces premature commitment.-
Users may want to draft and iterate before attaching a list.
-
This slows exploration and experimentation.
-
-
Editing Experience Is Rigid
Each step is fully expanded, making the page long and visually dense.-
No quick overview of steps (channel, delay, goal) at a glance.
-
Hard to scan, reorder, or reason about the sequence flow.
-
-
No Confidence or Readiness Signals
There’s no validation or guidance before launching.-
Users are left guessing if the sequence is “good enough” or missing key elements.
-
Risk of launching low-quality or incomplete sequences.
-
Strategy
Rather than redesigning individual UI elements, I focused on restructuring the experience.
Guiding principles:
-
Reduce upfront complexity
-
Make AI proactive, not reactive
-
Shift user effort from creation to refinement
The goal was to help users move quickly from intent to a structured draft.
Key Solution Features
Success Criteria
Users can generate a first usable sequence draft in under 5 minutes.
AI-generated steps feel predictable, editable, and easy to refine.
Users feel confident launching a sequence without extensive manual rework.
The interface was built using an existing design system (Chakra UI) to ensure feasibility and fast implementation.
Prototype
The layout emphasizes:
• Clear hierarchy between sequence steps
• Scanability of timing and channels
• Inline editing
• Embedded AI controls without clutter
Given the short timeline and mature design system, I moved directly to high-fidelity screens to validate interaction and hierarchy under realistic constraints.
Design Decisions
The redesigned flow reduces cognitive load and accelerates time-to-value. Users reach a usable draft quickly and retain full controll over structure and messaging. AI acts as a collaborator, not a black box.
Conclusions
This redesign reframes sequence creation as a guided, high-leverage workflow rather than a configuration-heavy task. By prioritizing speed to first draft and embedding AI directly into the core interaction, the experience reduces friction at the most critical activation moment.
The focus was not just on improving UI clarity, but on strengthening the product’s value loop: define intent → generate structure → refine → launch with confidence. The result is a flow that accelerates time-to-value while maintaining user control.
From a product perspective, this solution establishes a strong foundation that can evolve incrementally.
This redesign reframes sequence creation as a guided, high-leverage workflow rather than a configuration-heavy task. By prioritizing speed to first draft and embedding AI directly into the core interaction, the experience reduces friction at the most critical activation moment.
The focus was not just on improving UI clarity, but on strengthening the product’s value loop: define intent → generate structure → refine → launch with confidence. The result is a flow that accelerates time-to-value while maintaining user control.
From a product perspective, this solution establishes a strong foundation that can evolve incrementally.
Next Steps / Future Improvements
The next iterations would focus on strengthening quality, confidence, and scalability:
-
Recipient validation
Add a pre-launch check to confirm recipients are correctly selected and meet basic quality criteria (e.g. missing email, duplicates).
-
Post-launch success feedback
Provide clear confirmation after launch, reinforcing momentum and guiding users toward performance monitoring.
-
Accidental exit prevention
Prevent data loss by prompting users to save the sequence as a draft when attempting to exit mid-flow.
-
Readiness indicators
Surface lightweight signals (e.g. missing personalization tokens, empty steps) to clarify whether the sequence is safe to launch.
-
Step-level AI tuning
Allow quick, contextual AI adjustments per step (e.g. “shorter”, “more direct”, “less salesy”) without regenerating the entire sequence.
-
Draft reuse
Enable saving high-performing sequences as reusable templates to support scale and consistency across teams.