Agenda

Every AI platform now wants an MCP connector, but exposing thousands of Salesforce APIs as thousands of tools breaks agents before they start. Learn how Headless 360’s Discover → Describe → Dispatch model collapses that sprawl into four tools — and why we think of it less as “one server to rule them all” and more as the default on-ramp for agentic Salesforce work.

Abstract: Join us to design your Salesforce MCP strategy: which to use, for whom, and how. Hosted MCP for agents, the local DX MCP Server for developers, plus what to expose and how to govern it.

The idea of this session is to understand how to rollout Salesforce MCPs targeting different areas of your company. How developers will take advantage of the local DX MCP, and how business users could consume it externally from other systems, like ChatGPT, Claude, Slack and Teams. Also the session explains how to selectively expose different resources to different users within each server.

This session was submitted to Dreamforce.

As organizations push Data Cloud (Data 360) into real-time, high-volume use cases, the architecture starts to behave very differently than expected. What looks straightforward at smaller scale quickly turns into a series of tradeoffs between cost, performance, latency, and data freshness.

This session shares lessons from running Data Cloud at the edge of scale, where query volumes, ingestion patterns, and real-time expectations collide. We’ll walk through the decisions that mattered most, what to optimize, what to offload, and where Data Cloud delivers the most value when used intentionally.

Attendees will leave with a clear understanding of the hidden tradeoffs behind real-time architectures, and practical guidance on how to design a Data Cloud strategy that scales without surprises.

¿Crees que tu org de Salesforce está bien protegida? Tus integraciones podrían estar exponiendo datos sin que te des cuenta. En esta sesión veremos cómo Connected Apps, tokens y esas integraciones de “solo una más” pueden convertirse en puertas abiertas para la información sensible. Hablaremos de riesgos reales, errores comunes y formas prácticas de proteger tu entorno con mejores patrones, controles más inteligentes y soluciones fáciles de aplicar. Todo explicado de manera clara, directa y con uno que otra risa en el camino.

Sabemos que el uso de IA en Commerce acelera el trabajo de un ingeniero, pero ¿tenemos claro el estado real de su desempeño frente a las particularidades de la plataforma? ¿Contamos con un fingerprint claro de lo que la IA puede resolver con éxito y de dónde se desvía?

Esta sesión presenta un estudio experimental basado en un caso de uso real con lógica de negocio compleja dentro de Commerce Cloud. Evaluaremos el desempeño de distintos modelos (incluyendo soluciones de OpenAI y Claude, así como el uso de skills e instrucciones contextuales personalizadas), identificando tanto sus mayores fortalezas como sus limitaciones debido a la naturaleza de la plataforma.

Contrastaremos estos resultados empíricos con el criterio y las decisiones de un Senior Engineer y un Tech Lead, midiendo el impacto de las alucinaciones, el rendimiento, la seguridad y el apego a los patrones nativos de Commerce. Finalmente, ofreceremos un marco metodológico para determinar qué gaps pueden mitigarse mediante una mejor ingeniería de contexto y cuáles exigen de forma insustituible el juicio humano.

Every operations leader knows the pain. To answer one simple question — “Are we on track this quarter?” — someone has to open a BI dashboard, check three Slack channels, review the support queue, inspect the engineering backlog, and reconcile everything by hand. The answer is slow, often stale by the time it is shared, and usually trapped in one person’s head.

This session shows a production-grade pattern for solving that problem with Agentforce deployed as a Slack app, backed by Salesforce Data Cloud Zero Copy as a federated data layer across three common enterprise silos: ERP, support/CRM, and Jira.

The key idea is simple but powerful: the agent does not own or move the data. Data Cloud federates it in place through Zero Copy, allowing the agent to query a unified model while each source system remains the system of record. Adding a new data source becomes a connector decision, not a full integration project.

We will walk through the end-to-end architecture: Slack → Agentforce → Apex actions → Data Cloud SQL → federated enterprise sources. Then we will demo how an operations user can ask business questions in Slack and get grounded, actionable answers without manually stitching systems together.

We will close with the three lessons that mattered most in the real build: security is a permission chain you cannot skip, Zero Copy can reduce time-to-prototype from weeks to days, and the right unit of value is not “cool AI feature” — it is hours of human work recovered.

Attendees will leave with a reusable reference architecture they can map to their own Salesforce orgs and enterprise data landscape.

Marketing teams are being asked to do more with fewer resources while delivering faster, more personalized campaigns. This session shows how combining Marketing Cloud, Slack, and Agentforce creates a practical, modern operating model for campaign execution.

Rather than focusing on theory, this session walks through real-world patterns for using Slack as the collaboration layer and Agentforce as the orchestration engine behind Marketing Cloud campaigns. Attendees will learn how teams can trigger actions, manage approvals, surface insights, and resolve issues in real time without leaving Slack. It demonstrates how Salesforce customers and partners can move beyond disconnected tools and manual processes to deliver measurable speed, alignment, and operational efficiency across marketing teams.

¿Qué pasaría si tu Slack pudiera gestionar todo tu programa comunitario —registro, participación diaria, reconocimiento y reportes— con casi ningún trabajo manual? Eso es exactamente lo que construyó la Comunidad Fitblazer.

Fitblazer comenzó como un sencillo espacio de trabajo en Slack para un reto de pasos previo a Dreamforce, TrailblazerDX y Connections, y ha crecido hasta reunir a más de 200 Trailblazers que han registrado más de 60 millones de pasos y creado conexiones reales a través de husos horarios y continentes, todo impulsado por Slack.

En esta sesión aprenderás a:
● Automatizar las tareas repetitivas — usar Workflows para capturar registros, pasos diarios y retos semanales sin esfuerzo manual
● Dar seguimiento al progreso a gran escala — conectar las Listas de Slack con Google Sheets y Tableau para paneles y visualizaciones en tiempo real
● Mantener viva la participación — usar Slackbot para motivar, celebrar y mantener a cientos de personas comprometidas en cada zona horaria

Ya sea que estés dirigiendo un reto de bienestar, una comunidad interna o un programa de participación de clientes, te llevarás ideas para convertir a Slack en el cerebro y el corazón de tu comunidad.

As AI automates more work, competitive advantage shifts to intentional human touchpoints. Learn how to identify, orchestrate, and measure moments that create trust, meaning, and value. The session explores a question I think will become increasingly important as Agentic AI evolves: Where should humans remain intentionally present in our workflows and experiences?

Rather than viewing humans as bottlenecks, the session positions human connection as a strategic differentiator. As AI handles more routine work, trust, empathy, creativity, coaching, celebration, and shared experiences become increasingly valuable. Organizations that deliberately design for these moments may create stronger customer loyalty, more engaged communities, and better business outcomes.

One of the examples I used was the idea that automation can handle much of the process, but intentional human moments create the memories people carry with them. Those moments become the “icing on the cake” that transforms efficiency into meaningful experiences.

Learning Objectives

Identify which stages of a workflow should be automated and which stages benefit most from intentional human interaction.
Apply a practical framework for designing human handoff points within agentic and AI-driven workflows.
Measure how human-centered moments can improve trust, engagement, loyalty, and business outcomes.

Session Outline

–The Automation Paradox: What happens to the human experience as AI automates more work?
–The Human Handoff Framework.
–Real-world examples from customer experience, hospitality, community building, and leadership.
–Why human moments matter: trust, recognition, belonging, celebration, and encouragement.
–Designing intentional touchpoints inside AI-driven workflows.
–Building agentic systems that augment people rather than replace them.
–Interactive audience exercise to map automation opportunities and human handoff opportunities.
–Future outlook on human-centered AI and competitive differentiation.

Learn what makes AI difficult to behave predictably and deterministically through the lens f a neurodivergent, and how Agent Studio and Agent Script give you the tools to make it work.

How do we make AI-assisted development secure, consistent, auditable, and good enough to pass an expert code review? Vibe coding gets you a working screen. It doesn’t get you past a reviewer who opens the pull request and finds a SOQL query with no sharing enforcement and tests that assert nothing.

In this interactive workshop, we build a complete Salesforce solution live with Claude Code. You’ll see how to extend it with skills, subagents, and custom tools into a repeatable process that plans, implements, tests, reviews, and validates, while keeping context tight and token consumption predictable.

We’ll add guardrails and automated checks that catch what AI-generated Apex and LWC most often get wrong: CRUD and FLS enforcement, sharing violations, SOQL injection, hardcoded secrets, and unguarded endpoints. Every artifact traces back to a requirement.
This interactive session is Salesforce developers and architects ready to stop treating AI as a code generator and start building secure, production read solutions for their orgs.

Join this interactive session designing agents with architectural discipline. Apply proven principles to diagram workflows, define boundaries, and build agents that scale for real-world use.

AI agents are quickly moving from experiments to production systems—but many fail due to missing structure, unclear boundaries, and lack of observability. This session introduces a practical design framework for building reliable AI agents by treating them as systems, not prompts. Designed for architects, developers, and technical leaders, this session helps attendees build agents that are easier to scale, debug, and trust. We will introduce our proven framework, called “The Agentic Seven” during this session.

The premise is that most career advice you find online is awful. So, I took common Salesforce concepts that admin/devs/architects use every day (Technical debt, Ask Why Before How, Measure Your Success, Close the Loop, etc) and showed how they apply to your career.

Salesforce impulsa alianzas estratégicas y programas de desarrollo laboral orientados a cerrar la brecha de talento digital mediante la capacitación en habilidades clave de Inteligencia Artificial.

Abstract: Discover how a personal SWOT analysis (Strengths, Weaknesses, Opportunities, Threats) can transform your Salesforce job search and career planning. I’ll share my own experience using SWOT to gain clarity and identify growth areas, then provide practical examples for each element. Attendees will leave inspired to map their own skills, uncover opportunities and take actionable steps toward their next career move.

Additional Info:
While working on my MBA, I learned about the concept of the SWOT Analysis. I thought it was a great idea for a business, but thought it would be an even better tool I could use to analyze myself and my previous experiences. I dug through everything in my past to understand what my skills were, where gaps existed and what I could do to improve. It helped to give me a lot of clarity, understand my strengths and find direction. It’s also a really great way to prepare for a job search or interview. I’ve continued to create personal SWOT analyses when I feel I’ve hit a roadblock or when I’m approaching a transition in life. I’ve found this to be helpful and am excited to share it with others in the community.

El Rock & Roll Hall of Fame recibe miles de preguntas de fans y visitantes: horarios, boletos, membresías, exhibiciones, y cada año una avalancha por el Fan Vote de la inducción. Durante años eso lo atendió un chatbot de un vendor externo. En esta sesión cuento, sin filtro, cómo estamos construyendo su reemplazo con Agentforce — desde el PoC hasta un piloto con capacidades agenticas reales: consulta de estatus de membresías, boletos de admisión general para el museo y conciertos vía Apex, información de eventos, artistas inducidos, y grounding contra nuestros datos de Salesforce.

No es una demo de slides de marketing. Vamos a ver lo que nadie te cuenta: cómo diseñar topics y actions que no se pisen entre sí, cómo configurar el Einstein Trust Layer y sus guardrails — incluyendo el bug de falso positivo que bloqueaba cualquier pregunta con la palabra “vote” (sí, en el museo del Fan Vote), cómo lo diagnosticamos y resolvimos —, cómo exponer el agente en canales con Messaging for In-App and Web (MIAW), y cómo estructuramos el caso de negocio: un decision gate con criterios medibles para decidir si el agente reemplaza al vendor o no.

Te llevas un framework replicable para pilotear Agentforce en tu organización: qué medir, qué guardrails configurar desde el día uno, los errores que ya cometimos por ti, y cómo venderle el proyecto a tu empresa con milestones concretos.

Durante mucho tiempo nos han enseñado a buscar empleo como ‘candidatos’ que esperan ser elegidos. Pero en un mercado tan competitivo como el actual, ese enfoque ya no funciona. En esta plática quiero compartir la mentalidad que cambió mi carrera: dejar de ver tu perfil como un CV y empezar a gestionarlo como una empresa de servicios.

Para que las mejores empresas quieran contratarnos —y para escalar dentro de ellas— necesitamos aplicar los mismos principios que Salesforce le vende a sus clientes: posicionamiento, generación de valor y diferenciación.

En esta sesión mostraré cómo ejecutar este modelo de negocio en 3 etapas clave:

1. Fase de ‘Marketing y Reconocimiento’ (Antes de entrar): Cómo dejar de mandar CVs genéricos y empezar a construir credibilidad real. Hablaremos de cómo comunicar lo que sabes, resolver problemas en público y hacer que las empresas te busquen a ti.

2. Propuesta de Valor (El Producto): Cómo pasar de decir ‘sé configurar Salesforce’ a construir un expertise claro en procesos de negocio específicos que las empresas mueren por resolver.

3. Ser un ‘Eslabón Crítico’ (Una vez adentro): Una vez que estás dentro de la empresa, tu trabajo no es solo cumplir tareas. Debes entender la cadena de valor de la compañía y posicionarte no como un costo operativo, sino como un generador de ingresos o eficiencia.

No importa si estás buscando tu primera oportunidad o si ya estás dentro y sientes que te estancaste; esta charla es para reprogramar la forma en que ves tu carrera y enseñarte a negociar tu valor en el mercado.

¿Alguna vez te has preguntado por qué una arquitectura que se veía perfecta en el papel termina fallando en producción? En esta sesión exploraremos los pilares que toda arquitectura sólida necesita como: seguridad, escalabilidad, integración y gobierno de datos, y los conectaremos con errores reales que se repiten una y otra vez en proyectos. Sales con ejemplos concretos y una checklist que puedes aplicar desde tu próximo proyecto.

Aprenderemos a configurar correctamente a los agentes y descubriremos cómo lograr una sincronización de metadatos perfecta para que las configuraciones se mantengan alineadas entre Salesforce y Amazon Connect. Es una oportunidad única para entender cómo esta herramienta innovadora potencia la eficiencia del equipo de soporte y la resolución de problemas.

This session outlines how to shift from a reactive “builder tendency” (optimizing for short-term speed) to a proactive “architect practice” (designing for future system scale and managing technical debt). It explains how everyday project constraints like budgets, timelines, and limits act as the ultimate repetition-driven training ground to build architectural judgment

Learn how to connect an external CCaaS platform to Salesforce Service Cloud using BYOC APIs, Platform Events, and Agentforce — with real architecture patterns, packaging tips, and lessons learned.

La inteligencia artificial está dejando de ser una herramienta para convertirse en un actor dentro de las organizaciones. En esta sesión exploraremos qué significa construir una Empresa Agéntica: una organización donde agentes de IA colaboran con personas, sistemas y procesos para tomar decisiones y ejecutar tareas de forma autónoma. Veremos cómo tecnologías como Salesforce Agentforce, el MuleSoft Agent Fabric (la malla de conectividad e interoperabilidad entre agentes) e Informatica (para calidad de datos) se combinan para dar vida a esta arquitectura en el mundo real.

API-led connectivity is a powerful pattern in MuleSoft, but it is often over applied. Many teams introduce System, Process, and Experience APIs even when reuse is limited and simpler approaches would be more effective.
This session explores how overengineering MuleSoft integrations can increase cost, complexity, and resource consumption. We will walk through common scenarios where MuleSoft is overused, when Salesforce-native options like Flows with MuleSoft Connectors, external callouts, or Data Cloud may be a better fit, and how to right-size your architecture based on real business needs.

Attendees will leave with a set of questions to ask before defaulting to API-led so you can tell the difference between integrations that deserve three layers and ones that don’t.

As organisations evolve in their data strategies, the challenge often lies in optimising multiple specialised platforms. Salesforce Data 360 revolutionises real-time customer data management and the creation of unified customer profiles, while Snowflake and Databricks excel in large-scale data warehousing, complex data transformations, and machine learning.

This session fills a crucial gap in the community’s knowledge by showing how these platforms can co-exist and complement each other within a company’s data ecosystem.

Attendees will discover how combining the real-time capabilities of Salesforce Data 360 with the deep analytics and AI power of Snowflake and Databricks can lead to enhanced strategic advantages. Looking toward 2027, the integration of these technologies is more critical than ever, as organisations increasingly leverage AI and machine learning to stay competitive.

This session will provide hands-on insights, practical examples in various areas, and a clear integration roadmap, empowering Salesforce users to maximise their data-driven potential.

By embracing this synergy, organisations can push their data strategies further and deliver superior, personalised customer experiences. This session will offer unique knowledge and actionable strategies.

SOC Agéntico es una demo en vivo de un Centro de Operaciones de Seguridad operado por agentes de IA, construidos y orquestados sobre MuleSoft. Tres agentes especializados —Triage, Investigador y Responder—colaboran de forma autónoma para procesar alertas de seguridad de principio a fin, coordinados por Agent Fabric, que gestiona el ciclo de vida de los agentes y el paso de contexto entre etapas de la investigación.

Todo el tráfico de APIs y agentes queda gobernado por Omni Gateway, que aplica políticas de seguridad, control y observabilidad sobre cada invocación. Los agentes exponen y consumen capacidades —como el enriquecimiento de indicadores de compromiso o la creación de tickets— a través de servidores MCP, mientras que Anypoint Platform conecta todo el pipeline con el ecosistema real del SOC: fuentes de alertas, enriquecimiento de amenazas, ticketing y mensajería. El flujo es además agnóstico de LLM, de modo que el motor de razonamiento puede intercambiarse sin tocar la capa de integración construida en MuleSoft.

En la demo, una alerta cruda se transforma en un ticket de Asana y una notificación de WhatsApp al analista en apenas 50 segundos, usando fuentes de inteligencia de amenazas reales (VirusTotal, AbuseIPDB) para un enriquecimiento verídico, no simulado. El mensaje central para la audiencia de Salesforce es claro: MuleSoft no es solo integración de sistemas — es la capa que hace posible operar IA agéntica de forma segura, gobernada y conectada al negocio.

Bring your laptop and build an AI-powered Salesforce app from scratch. You’ll use the Agentforce Vibes IDE to generate and deploy Apex and Lightning Web Components using natural language, and connect your org to AI tools through Salesforce Hosted and Custom MCP Servers. Along the way you’ll build a headless experience with the Headless Toolkit and work with Plan and Act modes, Skills, and the CLI, so you leave with patterns you can use in your own projects.

Realizar una Demostración de lo que ya pueda hacer la inteligencia artificial conectando Aircall (Telefonía) con Salesforce, haciendo un seguimiento vía telefonica con un agente virtual y logrando conectar con un gestor de siniestros.