Scaling Grayscale’s UX for Enterprise Personas

Context

Grayscale is a B2B hiring platform that integrates with customers’ ATS to automate messaging workflows for high-volume recruiting. As a designer, I supported Grayscale’s move upmarket by scaling the UX to serve large, data-heavy customers with multilevel org structures. As the sole designer, I actively pushed against working in a silo by anchoring my decisions in:

Product strategy for each quarter
Cross-functional insight
How different enterprise personas use Grayscale

Balancing UX for Enterprise & MDM Customers

Even though Grayscale was scaling for enterprise adoption, I couldn’t let mid-market or SMB customers fall by the wayside. I had to develop a deep understanding of their nuanced use cases and needs.

Grayscale's Personas

Persona	Behavior	UX Implication	Case Study
Candidate	Apply to jobs through AI Chat, WhatsApp, or SMS.	Interfaces must preserve conversational flow. In-chat UI or lightweight mobile web tasks prevent context-switching.	Secure EEOC Widget
End User	Use Grayscale alongside their ATS to complete hiring tasks.	Clean, explainable UI for quick task completion with minimal training.	Automations Setup (Case Study Will Become Available After GA)
Enterprise Admin	Own and maintain the system for end users; often in close contact with Support.	Configuration-heavy flows prioritize efficiency over simplicity. Admins are heavily trained and expect dense, flexible controls.	Self-Service AI Assistant Configuration

Scaling UX Across Org Sizes

	Enterprise	SMB/MDM	UX Balance
Data Volume	Tables may contain tens of thousands of records across multiple brands	Typically dozens to hundreds	Tables support large datasets (filtering, segmentation, and configurable columns)
Filtering Needs	Require dynamic, multi-select filters and saved segments	Static filters with single-select options are often sufficient	Filters must be powerful, yet easily manageable. They must have intuitive logic and give users control over saved views (like segments)
Permissions & Visibility	Strict role-based access and data visibility controls	Access is often shared across roles	Clearly differentiate user roles in the UI and make ownership (private vs. shared) visually obvious

The Challenge

The challenge was ensuring that UX patterns scaled gracefully across both SMB and enterprise organizations. The so-called "enterprise user" wasn’t a single persona but a mix of candidates, recruiters, and admins who all had very different needs. I needed to balance simplicity for some users with power and flexibility for others, while preserving a holistic brand experience.

Feature #1

Talent Pool

Target User Personas: End Users & Enterprise Admins

Problem

Customers wanted a centralized database of every lead. This not only included past applicants, but anyone who interacted with their chatbot or marked a job as interesting. Since many customers were hiring for high-volume roles, the number of entries in this database was likely to be astronomical, making discoverability critical.

For end users, Grayscale functioned primarily as a messaging add-on to their ATS, not a replacement for it. Recruiters needed to dip into the Talent Pool to find the right candidates without breaking their flow in core ATS tasks. Meanwhile, enterprise admins expected more advanced functionality: configuration, segmentation, and permissions that respected the complexity of large org structures.

Design Goal

Create a Talent Pool that scaled to enterprise data volume while still feeling lightweight and usable for recruiters in the flow of daily work.

Before

Tables crammed multiple data points into single cells, making them difficult to scan or export cleanly.
There was no search or sorting capability, so recruiters had to either use CTRL+F or look for Talent Pool entries nestled inside paginated data.
Filtering used an unusual, nonstandard pattern to build complex queries, but it wasn’t filtering on fields users would find useful.
Candidates could be added to lists for easier organization, but the list feature was inaccessible from anywhere else in the application.

After

Structured, export-friendly tables: Each data point was separated into its own cell, improving readability and enabling clean CSV exports that matched what users saw in Grayscale.
Configurable data columns: Users could choose which fields mattered most to them and hide the rest. This reduced overwhelm and gave different personas the power to surface only the fields they found most important without us having to design for edge cases based on persona or org size.
Instant searchability: A dedicated search bar allowed recruiters to find candidates by name, email, or keyword in seconds.
Advanced filtering: Searchable, multi-select filters with include/exclude logic and intuitive presentation replaced the old filters, making it easy to refine thousands of candidates into relevant subsets.
Segmentation & tagging tools: Recruiters could now save groups, apply dynamic tags, and build lists to power automations such as outreach campaigns.

Impact

Whether they were searching for a single candidate or building a segment of targeted candidates to run automations against, recruiters could find them within seconds using their Talent Pool.

Feature #2

Self-Service AI Assistant Configuration

Target User Personas: Enterprise Admins

Context

For context, Gracie is the AI personification of Grayscale’s automated messaging product.

Enterprise admins relied on customer support for simple updates to their instance of Gracie, like changing their name, adjusting brand colors, updating FAQs, or tweaking tone. This request was escalated to engineers, pulling them and the support team away from their daily tasks. This revealed an opportunity to reduce the support tickets by giving admins more power over their AI Assistants.

Problem

Admins shouldn’t have to wait for quick, low-risk changes to their AI Assistant.
Support’s time is better spent on strategic guidance, such as onboarding and solving complex challenges.
Engineering should be protected to focus on shipping new features rather than small customer requests.

Design Considerations

Clear Navigation: Each data point was separated into its own cell, improving readability and enabling clean CSV exports that matched what users saw in Grayscale.
Guided UX Copy: Since there were so many setup configurations, it would be easy for even the most knowledgeable admin to get lost in a long form. These settings required a clear menu that mapped each step to a part of Gracie’s behavior or appearance, so users could quickly find what they needed.
Familiar Flow: Inline explanations clarified what each option controlled and the downstream effect of a change, reducing uncertainty.
Available Tech: A preview → save → deploy pattern mirrored common enterprise tools, lowering the learning curve for busy recruiters.
Segmentation & tagging tools: It wasn’t difficult for engineers to set permissions in such a way that admin level users could preview and publish changes and admins could act in minutes instead of waiting days.

Solution

I designed a self-service configuration flow that was easy to navigate and gave customers direct control over Gracie’s everyday settings while preserving white-glove support for complex needs. My guiding principle was to empower admins while maintaining Grayscale’s white-glove support model.

Impact

Reduced repetitive support tickets, enabling Support to focus on onboarding, optimization, and troubleshooting complex workflows.
Allowed engineering to reclaim time from manual updates and invest it in building new product capabilities.
Strengthened customer confidence by giving admins quick, transparent control while keeping Grayscale’s support team available for higher-touch needs.