3 Years of purposeful AI, dramatically improving brand perception

Summary

Security admins needed training to use our software to its fullest. Even routine tasks meant clicking through multiple screens. Meanwhile, AI was changing what people expected from software everywhere else. And everyone—from customers to the executive team—wanted to see what we could do with AI. We needed to figure out how AI could actually reduce friction—not just add flashy features to a deck.

As Design Director, I led Brivo's AI strategy from early experiments through production launches (2022–2025). This involved building organizational buy-in for emerging technology, establishing research processes to validate experimental features, and directing design execution across multiple teams simultaneously.

Team: 4 designers, 4 product managers, AI/ML engineering team, executive stakeholders

Impact at a Glance

Shipped 6 AI-powered features across web, mobile, and hardware platforms:

• Facial recognition
• Brivo Genius support AI (trained on installation materials)
• Smart search with natural language queries
• Brivo Genius mobile agentic AI
• Conversational Voice UI for unmanned locations
• Hardware health anomaly detection

Results:

• 80% reduction in time to complete administrative tasks
• 84.9% average ease of use rating across AI features
• 95% time savings in Voice of Customer analysis (days to minutes)
• Overwhelmingly positive user and community reaction, especially to agentic interactions

This gives us a clear mandate moving forward: deeper agentic capabilities, expanded voice, and AI-powered predictive security.

Bryan Neff | Director of Engineering @Brivo

“Steve holds a remarkably high bar for design quality; he doesn’t just aim for 'good,' he pushes for excellence. He consistently brings data to the table to validate assumptions, ensuring that every creative decision is rooted in user needs and business goals.”

The Challenge: Aligning Siloed Teams Around Emerging Tech

When I started exploring AI applications, the development team had already begun their own R&D—hiring specialized staff and running experiments completely separate from Product and Design.

This created several problems:

• Duplicated effort across teams exploring similar solutions
• No validation process to ensure technical experiments solved real user problems
• Unclear prioritization between multiple AI initiatives

I met with the CTO and proposed a collaborative approach: Product, Design, and Development would plan AI initiatives together. This coordination enabled us to create unified roadmaps and story maps, leading to significantly faster—and more user-centered—exploration.

How I Led This

Established Annual Strategic Planning Process

I facilitated annual AI brainstorming workshops, bringing together designers, product managers, and engineering to explore opportunities from multiple angles. One particularly productive session asked: "What problems do users experience that AI might solve?"

I brought Voice of Customer data, support analytics, and user research insights to ground the discussion in real pain points. This prevented us from chasing technically impressive but user-irrelevant features.

Built Cross-Functional Alignment

My coordination work went beyond Design. After securing agreement from the CTO, I:

• Created unified AI roadmaps spanning Product, Design, and Engineering
• Established story mapping processes to break down complex AI features
• Ran collaborative workshops to align on technical approach and user experience simultaneously

This structure allowed Dev to continue specialized R&D while ensuring work connected to validated user needs and product strategy.

Validated Ideas Before Committing Resources

For each AI initiative, I established research processes appropriate to the feature:

• Facial Recognition: Beta program with actual customers at real installations. I coordinated designers across email, web consent flows, administrator experiences, reseller tools, legal requirements, and hardware interactions.
• Voice UI: User research with admins, resellers, and Beta participants confirmed assumptions about unmanned location needs. Testing validated that conversational interaction was desirable before Product added it to the roadmap.
• Support AI (Brivo Genius): Leveraged opportunity to coordinate with technical writing, product management, and training teams to improve underlying support materials. Each iteration reduced support call volume.

Key Features Delivered

Support AI (Brivo Genius)

Our first generative AI was trained on support materials and embedded into the main application, addressing hardware installation pain points. I used this launch to coordinate cross-functional improvements—working with technical writing, product management, and training to roll out simplified installer-focused materials. Support call volume dropped with each iteration.

Facial Recognition with Privacy-First Design

The most significant design challenge wasn't the technology—it was consent and privacy flows. I brought together several designers to coordinate work across email consent, web tracking interfaces, administrator and reseller experiences, legal compliance, and physical hardware interactions. This required orchestrating multiple development teams simultaneously toward Beta customer installations.

Voice UI

User research validated Voice UI for unmanned locations, replacing traditional answering services.

Agentic Security Operations Management

The mobile agentic AI was particularly significant—enabling administrators to query security events ("How many people were in the office yesterday?") and immediately take action ("Lock down the office"), representing a fundamental shift from information retrieval to action execution.

The mobile agentic AI (Brivo Genius in Brivo Access mobile) generated the most user excitement. Administrators could now query security events ("How many people were in the office yesterday?") and immediately execute actions ("Lock down the office" or "Add a user")—capabilities we'd never offered before.

This represented a fundamental shift from information retrieval to action execution, achieving an 80% reduction in task completion time. We also explored using the reporting capabilities to generate ephemeral UI—having the AI create charts and apply them directly to dashboards, allowing dynamic page generation based on natural language queries.

User and community reaction was overwhelmingly positive, particularly to the agentic interactions. This validated our strategic direction and gave us a clear mandate: deeper agentic capabilities, expanded voice interfaces, and AI-powered predictive security.

My Leadership Approach

Growing Team Capabilities

I challenged designers to think bigger by:

• Analyzing competitor AI implementations
• Exploring inspirational sources beyond security software
• Using AI tools themselves as design exploration methods

Over time, both development team AI skills and design team ideation quality improved significantly. The iterative "brainstorm, try, brainstorm, try" approach created organizational learning.

Establishing AI Tools for Design Operations

Beyond customer-facing AI features, I implemented AI into our design practice:

Voice of Customer Analysis: Created a shareable Gemini Gem trained on VoC standards, reducing analysis time from days to minutes (95% time savings). Made it accessible to anyone in the organization.

Content Standards: Built a Gemini Gem for our technical writer trained on voice, tone, and style guidelines. His writing quality improved significantly, especially when he consistently used this tool.

Design Exploration: Began exploring Figma AI and Gemini Canvas for screen design, though this wasn't yet formalized into our process.

Organizational Impact

• Unified AI roadmap across Product, Design, and Engineering
• Established research-driven validation process for experimental features
• Built cross-functional collaboration framework that accelerated exploration
• Improved AI capabilities across both development and design teams
• Created reusable AI tools for operations (95% time savings in VoC analysis)

Leadership Lessons

This initiative reinforced several leadership principles:

1. Alignment before execution: Investing time to coordinate siloed teams dramatically accelerated progress.
2. Validate with users, not assumptions: Even technically impressive features needed user validation before committing resources.
3. Strategic rhythm: Annual visioning workshops created consistent opportunities to reassess direction as AI technology and organizational capabilities evolved.
4. Cross-functional coordination: Complex features like facial recognition required simultaneously coordinating multiple designers across email, web, mobile, legal, and hardware—demonstrating the importance of strong orchestration skills.
5. Build capabilities, not just products: Creating AI tools for design operations (VoC analysis, content standards) multiplied team effectiveness beyond individual feature launches.