TDX 2026 Sets the Agentforce Agenda: What Changes for Your Lifecycle Stack This Quarter

Salesforce kicks off TDX 2026 in San Francisco on April 15–16 with 400+ sessions on the Agentic Enterprise and hands-on time with Agentforce 360 and Data 360. The real headline for marketing teams: agents, identity, and Slackbot move from demos to daily work in Q2 (Salesforce Newsroom live blog). Pair that with Gartner’s April 9 recognition of Salesforce (via Informatica) as a Leader in the 2026 MDM Magic Quadrant and you get a clear mandate—your next lift isn’t just building an agent; it’s making your data trustworthy enough for one (Gartner MQ summary via Salesforce).

What just happened

• Agentforce 360 is positioned as the way to build and govern agents across Salesforce clouds. Campaigns and customer service flows will share the same agent fabric.
• Data 360 appears alongside Agentforce, signaling identity resolution and context as first-class inputs for agent decisions—not just reporting.
• Slackbot adds desktop note-taking, reusable AI skills, and voice. Salesforce highlights customers like Engine and Asymbl saving “dozens of hours” per week pre-launch (Salesforce Newsroom).
• Independent operators are publishing how-tos. Salesforce Ben’s April 13 guide shows building an agent in Agentforce with Flow-like patterns—policies, tools, memory (Salesforce Ben).

Why this matters for SFMC, Braze, and Iterable teams

Agents aren’t another channel—they’re an orchestration layer above your messaging tools. If identity and governance are loose, agents will amplify the mess. Salesforce Ben’s companion post says it plainly: if your data is broken, Agentforce will multiply those problems (Salesforce Ben).

Here’s the practical impact:

1. Identity becomes a runtime dependency, not a dashboard KPI

• Agent decisions depend on stitched profiles (household, account, consent). MDM discipline moves from “nice-to-have” to “won’t ship” when agents execute updates, offers, and support actions.
• Expect pressure to unify SFMC Contact Key with CRM Person Accounts and product usage IDs. Data 360’s positioning suggests native patterns will harden quickly.

2. Slack becomes the control plane for content and approvals

• Note-taking and reusable AI skills in Slackbot lower friction to generate, summarize, and approve content. That pushes content ops closer to the event stream—good for speed, risky for brand/compliance unless you codify guardrails.

3. Journeys shift from choreography to agent decisions at the edge

• Braze/Iterable orchestration will trigger from agent outcomes (e.g., service resolution, eligibility) rather than time-based branches. You’ll need cleaner event contracts and fewer “catch-all” audiences.

4. Governance shifts left

• Agent policies, tool access, and PII scopes must be defined alongside segmentation and content rules. “Which tools can this agent call?” is the new “which DEs can this journey read?”

The stack-level implications

• Data tier: With Salesforce/Informatica back in the MDM Leader quadrant, expect stronger native options for identity and survivorship. If you’re stitching with CDP exports plus warehouse jobs, reassess what moves into governed pipelines.
• Orchestration tier: Keep SFMC, Braze, and Iterable for channel delivery and testing depth. Let Agentforce handle decisioning where context spans sales, service, and product telemetry.
• Collaboration tier: Slackbot skills + voice reclaim a chunk of “ticket ping-pong.” Document approval and model-use policies in channels and canvases—not buried in Confluence.

What good looks like this quarter

Operationalize three moves that de-risk the agent wave without stalling the roadmap:

• Define one identity contract

  • Choose the canonical key (Contact Key vs. CRMID vs. product UUID) and publish mapping rules, survivorship, and consent sources of truth. In Salesforce shops, align to Data 360’s profile schema to future-proof.

• Split decisioning from delivery

  • Move offer eligibility, suppression, and next-best-action into a decisions service (Agentforce or a rules API). Channel platforms ingest decisions; they don’t compute them.

• Put Slackbot on a lane, not the freeway

  • Pilot reusable skills in two flows only: content QA and incident/open-loop reduction. Bring a rubric: allowed sources, tone/style rules, PII constraints, and human checkpoints.

Watch-outs we’re seeing in audits

• Prompt sprawl in Slack causing inconsistent claims and compliance exposure. Fix with shared skills and versioned prompt libraries tied to brand/legal guidelines.
• “Agent can call anything” tool permissions. Enforce least privilege. Start with read-only tools plus simulated writes before touching production objects.
• Identity drift between SFMC DEs and CRM. Automate daily diff checks and auto-open tickets when Contact Key cardinality breaks 1:1 with Person/Lead.

What to do about it

• Pick 1–2 agent-backed decisions tied to revenue and service: warranty replacement eligibility or churn-save rules tied to support events. Instrument data lineage and access scopes before wiring channels.
• Consolidate suppression logic into one callable function or Agentforce skill. Measure lift from fewer misfires and lower manual overrides.
• Ship a Slackbot skill pack for content approvals with hard stops on claims and offer terms sourcing. Track cycle-time delta and redlines per asset.

For deeper patterns on observability and unified architecture as you introduce agents, see Agentic lifecycle marketing needs a unified architecture—or you’ll ship shadow AI and why context, not models, is your bottleneck in Context is the real GenAI bottleneck in lifecycle marketing (and how RevOps can fix it).

If your stack faces the same identity tension and Slackbot policy questions, that’s the migration and governance choreography we’ve shipped for multi-cloud teams. If your SFMC or Braze instance is bracing for Agentforce, let’s compare notes in a working session.