Tribal's New York Workshop Puts Agentforce in Your Sandbox

Tribal is asking Salesforce teams to arrive with a problem they have been working around. Its upcoming Build an Agent with Tribal workshop is scheduled for December 2, 2026, from 3–6 PM EST in Midtown East, New York. The organizer plans to help participants build and deploy an Agentforce agent in their own Salesforce sandbox in under 2 hours, followed by demonstrations and happy hour.

The organizer's registration page describes a hands-on session for Salesforce platform owners, business users, admins and developers. Registration is free and requires approval; the precise address is shared with guests. Participants are asked to connect a Salesforce sandbox to Tribal before arriving and can bring colleagues who understand the problem they want to solve.

That invitation gives the workshop its business relevance. The person who requested an automation, the admin who understands the org and the developer reviewing its implementation can examine the same build together. For teams considering Agentforce, the useful opportunity is to make the conversation about requirements and behavior concrete while there is still room to change both.

A Salesforce workshop built around the participant's org

The planned sequence starts with choosing an agent worth building. Participants can ask Tribal for use-case suggestions based on their org or begin with a template such as the Sales Discovery Assistant or Lead Nurture Agent. The organizer then describes working through requirements, architecture and components before deploying the agent into a sandbox.

Tribal says its workflow translates a plain-language request into requirements, user stories and acceptance criteria, then develops architecture, Apex, tests and agent scripts for team review. Those are the company's descriptions of its product and workshop plan. The event offers a chance to inspect that process in the context of a participant's own Salesforce data model, where assumptions have somewhere specific to show up.

The scheduled event runs for 3 hours, while the organizer's build objective is under 2 hours. Demos, a competition for swag and happy hour are also on the agenda. The distinction matters for anyone planning the afternoon: the build is one part of the session, and the registration page should remain the reference for current timing and access details.

The business user has work to do, too

A request for an agent can sound straightforward until somebody has to specify what a useful result looks like. An admin may need the request narrowed to a particular workflow. A developer may need clarity about the generated component. The business user can explain which part of the current process consumes attention, and which exception would make the proposed fix unhelpful.

Bringing those people together creates an opportunity to resolve questions while the requirements are taking shape. Acceptance criteria give a team something it can revisit during the demo: does the agent handle the problem that brought the team into the room? That is an editorial reason to pay attention to this workshop's format, rather than a promise about the performance of any resulting agent.

The organizer explicitly welcomes people attending alone as well as teams. Platform owners can use the session to examine a potential Agentforce use case in their own environment; admins can bring a request that has remained in the development backlog; developers are invited to inspect the Apex that Tribal generates. Business users have a place in that review because the meaning of the task belongs to them as much as its implementation belongs to the technical team.

A sandbox build creates the next set of questions

Salesforce's agent-testing guidance explains why evaluation remains important as teams develop agents. Instructions, actions, underlying data and permissions can all influence behavior, and an agent needs to be examined across a variety of scenarios. Salesforce also advises sandbox testing because tests can modify CRM data.

For attendees, that context suggests a useful way to approach the afternoon. Choose a bounded task and be prepared to explain how the output will be judged. During the build, ask how the requirements connect to the implementation and which parts the team can review. Afterward, the sandbox agent can support further testing against the use case rather than remaining an abstract proposal in a queue.

Those are operator questions to bring into the room, not additional activities the organizer has promised to teach. The event's under-2-hour build goal is also an organizer claim, not an independently verified measure of how every attendee will progress. A participant's production decision will require its own evaluation; the workshop gives that participant a particular build to take back into the discussion.

The hosts and the room beyond the build

Daniel Goldstein, whose event bio identifies him as Head of Solutions at Tribal, is among the listed hosts. Nikolai Avrutov and Gilad Shoham also appear on the host list. The registration page does not publish a separate speaker lineup, so the useful expectation is a practical build session with hosts and other participants, rather than an assumed executive panel.

Demos create a natural reason to compare approaches. Seeing another team describe its problem, explain the choices it made and show a sandbox result can open a more specific conversation than exchanging broad opinions about AI adoption. Happy hour is planned afterward, but there is no verified attendee roster or basis for promising introductions to particular buyers, investors or executives.

Tribal's Salesforce product page places Agentforce architecture within its broader enterprise-app development offering. The December workshop concentrates that positioning into a problem an attendee can bring to New York. Before registering, confirm the current event details and the sandbox prerequisite; before arriving, agree with the colleague who requested the fix on what is worth building together. That shared brief will still matter when the demo ends and the team opens its own org again.