Deploy and Manage Apps Across Any Surface

Learning Objectives

After completing this unit, you’ll be able to:

Describe how the Headless Experience Layer delivers consistent experiences across AI surfaces.
Explain how to deploy apps confidently using integrated DevOps tools.
Summarize how embedded governance keeps your apps secure throughout the development lifecycle.

Deploy Consistent Experiences Across AI Surfaces

Your users talk to AI assistants everywhere—Slackbot, Claude, ChatGPT, Microsoft Teams, Agentforce. Each surface has different interface conventions. But your users expect the same rich, consistent, branded experience no matter which assistant they're using.

Northern Trail Outfitters (NTO) faced exactly this challenge with its influencer tracking agent. Sales reps wanted to check influencer performance in Slackbot. Marketing teams needed the same information during planning sessions in Microsoft Teams. Executives wanted to query metrics using ChatGPT or Claude on their preferred devices. This would require separate widgets for each AI surface and would mean inconsistent experiences, duplicated effort, and higher costs. Lucky for NTO, there was a better way.

Build Once, Deploy Everywhere with the Headless Experience Layer

The Headless Experience Layer (HXL), currently in beta, delivers consistent, rich user experiences across agentic surfaces without building separately for every surface. Define your UI widgets once, and they render consistently across Slackbot, Claude, ChatGPT, Microsoft Teams, and Agentforce.

Headless Experience Layer is a pilot or beta service that is subject to the Beta Services Terms at Agreements - Salesforce.com or a written Unified Pilot Agreement if executed by Customer, and applicable terms in the Product Terms Directory. Use of this pilot or beta service is at the Customer's sole discretion.

Here's how Sofia and John use HXL.

Sofia designs the UI widgets once in HXL Playground—an influencer list card and a detail card with metrics, trends, and action buttons.
Sofia previews across surfaces in HXL Playground to ensure consistent rendering.
John deploys the widgets to the influencer agent.
Users get the same rich experience everywhere, no matter which surface they use.

Consistent Experiences in Practice

With HXL, the Northern Trail Outfitters (NTO) influencer agent delivers unified experiences across every AI surface.

Sales reps in Slackbot: Ask “Get me top influencers” and get a rich card with engagement rates and reach right in their conversation.
Marketing teams in Microsoft Teams: Request “Provide me campaign performance” and get the same card rendered as an adaptive card within Teams.
Executives using ChatGPT or Claude: Ask “Get details for influencer ID 12345” and get the same detailed widget with metrics and recommendations on their preferred device.
Service agents in Agentforce: Help partners with influencer questions and get the same consistent widget on desktop or mobile.

Same business logic. Same UI widgets. Same branding. Different AI surfaces. Zero duplication. Users get consistent, rich experiences no matter which AI assistant they prefer.

Deploy with Confidence

Now that Sofia and John built their multi-surface app, they need to get it into production safely. Traditional deployments often involve manual checklists, late-night emergency fixes, and stressful release events. But with integrated DevOps tools, deployment becomes predictable and routine.

Integrated Release Management with DX Inspector and DevOps Center

Sofia and John use DevOps Center to orchestrate their deployment pipeline. Instead of manually tracking what changed and coordinating with multiple people, DevOps Center gives them complete visibility and automation.

Here’s how they deploy the influencer tracking app.

Track Changes from Development to Deployment

As Sofia and John work in their dev sandboxes, DX Inspector tracks every metadata change they make in real time. When ready to commit, they use DX Inspector to select their changes and commit them to a DevOps Center work item with clicks, not code. Sofia also creates a Pull Request directly from DX Inspector for team review. Then DevOps Center tracks each work item’s status through the pipeline. Sofia’s agent logic moves from In Progress to In Review to Ready to Promote. Everyone can easily check the status. And no more Slack messages asking, “Did you deploy your changes yet?”

Follow the Governed Deployment Process

When ready, Sofia promotes her changes to the next pipeline stage. DevOps Center automatically packages all the changes, tests them, and deploys everything in the correct order. What used to take 4 hours and require three people now takes 30 minutes with one click.

Maintain Version Control

Every change gets tracked in Git. When the CIO asks about what changed between last week and this week, they can tell her exactly which features were added, who made them, and when.

Automate Quality Checks

Every change gets automatically tested before moving forward. When Sofia submits her work for review, tests run in the background to catch bugs early. Before promoting to staging or production, the system validates that everything meets their quality standards. If something fails, the deployment is blocked until the issue is fixed, which prevents bad code from reaching customers. As changes move closer to production, the quality bar gets higher, which ensures only the most thoroughly tested features go live.

Use Natural Language to Manage Deployments

During a sprint review, their product manager wants to know what’s different in this release compared to what the team tested last month.

Before, Sofia would have spent 30 minutes digging through Git commits and deployment logs. Now, she opens Agentforce Vibes in DevOps Center and asks, “Show me what changed in the Q2 feature branch.”

Vibes immediately provides a summary. Three new agent actions, two updated React components, one modified permission set, and five test classes. The product manager gets the answer in seconds.

You can also use natural language to resolve blockers.

When Sofia encounters a merge conflict mid-deployment, the release comes to a halt. Rather than spending days untangling it manually, she turns to Agentforce Vibes integrated with DevOps Center. Vibes analyzes the conflict, surfaces the root cause, and guides her through a fix, and resolves the issue in hours instead of days.

Observe and Optimize

Shipping your app isn’t the finish line, it’s the starting line. Real success comes when you understand how your apps and agents perform in the real world, and use that data to continuously improve them.

For example, NTO monitors their influencer tracking app in real time. They discover that users in Slack engage with quick action buttons 3x more than traditional forms, so they adjust their interface design accordingly. The observability data feeds directly back into their next development cycle.

Governance Is Built In, Not Bolted On

As AI coding agents accelerate development, governance becomes more important than ever. How do you ensure agent-generated code meets your standards? How do you maintain security and quality when agents are writing code alongside humans?

Salesforce builds governance into the entire development process. Whether you’re using Agentforce Vibes, Builder Central, Multi-Framework with React, or DevOps Center, security and compliance are embedded from the start, not added as an afterthought.

Here's how governance worked for NTO.

Agents Respect Your Org’s Governance Rules

When Vibes generates code for John's dashboard component, it automatically respects the org's governance rules and validation logic. When Sofia builds agent logic in Builder Central, the same governance standards apply. The Einstein Trust Layer ensures agent-generated code follows your existing governance model.

IT Controls Which Tools and Skills Agents Can Access

When John tries to add a third-party integration skill, DevOps Center checks it against the approved tools list. Unapproved tools are blocked, which ensures agents only access trusted, allowed capabilities.

Quality Gates Enforce Standards

DevOps Center uses quality gate rules to ensure only code that meets your standards moves forward. Before Sofia can promote her work item to the next pipeline stage, DevOps Testing automatically runs assigned test suites and evaluates quality gate rules. Code Analyzer scans for issues. Tests verifies functionality. When her code passes all quality gates, it moves to staging. If something fails, DevOps Center blocks the deployment, which catches problems early when fixes are quick and easy.

Teams Refine Governance

When Vibes flags that John’s component needs additional test coverage, he and Sofia reviewed the requirements together. They add the tests, and the code moves forward. Governance acts as a collaborative checkpoint that improves code quality.

The result? NTO moves faster with AI coding agents while maintaining enterprise standards. Governance isn’t a blocker, it’s embedded in every phase of the development process.

Your team moves fast while governance keeps everything secure.

The Complete Development Journey

You now understand the complete agentic app development journey—from ideating and planning with your team, to building with AI-powered tools, testing with realistic data, deploying across multiple surfaces, and continuously observing and optimizing your solutions. You've learned how governance is embedded at every stage, which keeps your apps secure while you move fast.

予想時間

トピック

ヘルプをお探しですか?

Agentforce のリソース