• ~70% adoption of all orgs with Agentforce activated, since the Spring ’26 rollout.
• >50% weekly active of adopted accounts, a sticky core weekly routine.
• Unlocked exec funding traction that resourced the broader Marketing Goals Agent · patent pending.

Modern marketers are drowning in manual configuration. When teams are stuck fighting legacy tools, the customer experience degrades into fragmented, irrelevant noise. Teams become burnout-rich and insight-poor.

The mission wasn’t just to build a faster text generator. We needed to simplify the end-to-end campaign lifecycle, empowering marketers to build at the pace of an idea.

To pull this off, we needed an engine built on two non-negotiable design pillars. Every downstream decision was anchored here:

• Robust Grounding: Anchoring AI in brand guidelines, historical data, and audience insights so the agent generates safe, relevant content.
• Human-in-the-Loop (HITL) Control: Giving the marketer a persistent, editable brief. The AI holds the pen, but the human steers the ship.

Building ahead of the curve meant we had to continuously adapt our approach. To ensure we were always delivering the best possible user experience, we evolved the product architecture through four distinct phases, carefully balancing platform constraints, user-prioritized requirements, and rapidly expanding technical capabilities.

Draft With Einstein

The Spark · Spring 2024

We started with a focused, standalone experience to quickly validate the behavioral shift. Users loved the speed but gave us a clear mandate: We want the AI, but we need the steering wheel.

Einstein Copilot

The Scale · Summer 2024

To give users that steering wheel, we integrated into the platform’s native conversational assistant. But we hit a massive UX wall: the platform restricted us to a 300px utility panel. Users felt they were “orchestrating in the dark,” unable to visually connect their prompts to the campaigns they were building.

Campaign Designer

The Clarity · Spring 2025

To solve the real-estate crisis, we broke free from the standard side-panel constraints. We custom-built a wide-canvas modal, finally giving marketers the expansive space and inline trust citations they needed.

Campaign Creation Agent

The Current Evolution · Spring 2026

As the core platform’s capabilities advanced, we were able to fully converge our vision natively. We brought the expansive visual canvas and the conversational agent together, culminating in the unified Agentforce Workspace.

‹ ›

Phase 01 / 04

The evolution above is the what. This is the how: the hard problems I had to solve to make it real, all rooted in the second pillar: the AI holds the pen, the human steers the ship. Give marketers a surface they can steer, output they can trust, and control that doesn’t cost a prompt every time.

300px Just 300px wide: barely room to read one message’s copy, let alone see the full set we were drafting or steer what the AI generated.

Phase 2 forced the sharpest decision of the project. The platform’s Copilot gave me reach, but it capped every AI feature at a 300-pixel utility panel, and marketers were orchestrating campaigns they couldn’t actually see. I scoped two very different ways out:

I chose the modal even though, at the time, Salesforce modals couldn’t yet host the conversational panel. The clarity win outweighed the constraint, and Phase 4 later closed that gap.

Then I tested it.

Before we wrote production code, I put the wide-canvas modal in marketers’ hands. About 90% said they’d adopt it on paper, but the score wasn’t the real signal, the behavior was. Watching real use surfaced two fixes I made before we shipped.

Before

→

After

The polished preview looked editable, but marketers couldn’t touch the details they wanted to change. Stripping each message back to raw copy gave them a clear space to edit the words.

Because that same testing surfaced something deeper: two problems no amount of canvas space could fix. Marketers couldn’t always tell which words were theirs and which were the agent’s, and they were worn down re-typing prompts for small tweaks. Those two signals, trust and refinement fatigue, defined the next stretch of work, and ultimately converged in the native Phase 4 agent.

The first of those Phase 3 signals, “who wrote what?”, was fundamentally a trust problem. An agent that drafts marketing content is only useful if the marketer trusts it. Three failure modes threatened that trust, and each one became a design problem I had to solve before the experience could ship:

The second Phase 3 signal, refinement fatigue, is what pushed the design past the chatbox entirely. In its fully realized state, the Campaign Creation Agent bridges two distinct interaction models: conversational speed and visual control. When a marketer drops a prompt, the agent instantly drafts the brief and maps out a full multi-channel campaign on the expansive canvas. They get the big-picture view while keeping the conversational assistant docked for macro-level orchestration.

However, to combat “refinement fatigue” during micro-edits, we realized we needed to move beyond the chatbox entirely. We embedded quick actions directly onto the generated campaign nodes as intent-based direct manipulation controls. Now, instead of constantly typing prompts to make minor tweaks, users can express their intent through contextual actions: nudge tone, shorten a subject line, or regenerate a single step with a single click right where the content lives. We successfully blended the generative speed of AI with the precision of intent-based direct-manipulation UI.

Generative AI doesn't replace marketing craft; it collapses the busywork that surrounds it. The design job here was less about the model and more about the scaffolding: giving marketers an obvious next step at every stage, and trustworthy ways to review, edit, and approve AI-drafted work before it ships.

Furthermore, this project underscored a massive shift in our craft. The center of gravity is moving from producing individual, static screens to architecting underlying systems. While my core focus here was system scaffolding, I have actively integrated AI-assisted coding tools like Claude and Cursor into my daily workflow. Redefining the traditional design-to-engineering handoff through functional, systems-oriented prototypes is the necessary next step for AI-native design.