Why 'Flatbots' Are Failing and What CX Leaders Should Deploy Instead

The End of the Keyword-Matching Chatbot

For years, customer service organizations deployed chatbots that could retrieve help articles, recognize keywords, and route customers to forms or phone numbers. When customers arrived with exceptions, missing orders, or complex account issues, these bots hit the same wall: "I can't help with that."

At Salesforce's Dreamforce 2026, a different approach emerged from conversations with brands including F1, Live Nation, Crocs, and SPECS. These organizations are building AI systems that connect to the data, workflows, and actions required to actually resolve customer problems—not just respond to them.

The distinction between digital response and digital resolution is reshaping how CX leaders think about automation.

Three Deployment Paths to Resolution

The brands observed at Dreamforce took notably different technical approaches to the same goal. Some deployed prebuilt Agentforce interfaces and agents. Others used Salesforce's headless capabilities to build custom front ends on top of existing data and workflows. A third group used AI-assisted development to create new experiences across customer, frontline, manager, and leadership touchpoints.

Despite different methods, all shared a common objective: personalize experiences, resolve more issues digitally, reduce avoidable escalations, and create commercial value from better interactions.

Fin, Salesforce's recent acquisition, reported that its customers achieve an average 76% resolution rate through digital interactions. While vendor-provided figures warrant scrutiny—leaders should examine what counts as "resolved" and track repeat contacts—the metric represents a meaningful shift from measuring whether a bot simply responded.

Why it matters

CX leaders now face a strategic choice that extends far beyond chatbot deployment. AI agents that can take action require connected systems, clear data governance, defined permissions, and well-designed human handoff protocols. Organizations that treat this as a transformation program rather than a technology purchase will separate themselves from competitors still measuring containment rates instead of resolution outcomes.

The Infrastructure Reality Behind AI Agents

Salesforce's broader AIforce strategy positions the CRM as a governed layer beneath customer experiences—housing data, permissions, business rules, workflows, security controls, and approved actions. This architecture enables customer-facing agents to operate with the same trusted context as service representatives, and allows employees to access information through Slack, Claude, or custom interfaces rather than switching between applications.

But better AI models will not fix fragmented customer data, ambiguous policies, disconnected workflows, or poorly designed escalation paths. AI may simply expose these operational problems more quickly and at larger scale.

Five Questions Before Deployment

CX leaders should answer these questions before selecting an approach:

Do you want easy deployment? Start with a prebuilt agent on a well-defined, high-volume journey like returns processing, account updates, or order status.

Do you need customization? Build around the workflows, terminology, channels, and decision boundaries that make your customer experience distinctive.

Do you want rapid development? Ensure development speed doesn't outrun security, accessibility, data quality, or operational accountability.

Do you want model choice? Define which jobs require which models, what customer data they can access, and how you'll evaluate performance beyond conversational fluency.

Do you want full integration? Treat it as transformation—connecting systems, clarifying ownership, setting action permissions, designing handoffs, and giving CX a seat in AI governance.

From Deflection to Resolution

Customers have outgrown chatbots that politely direct them elsewhere. They expect their problems understood, their context recognized, and appropriate action taken. Vendors are rapidly providing CX teams with more sophisticated tools to build that future.

The harder question is whether organizations are prepared to do the operational work required to make it real.

These observations were first reported by Automation Watch following Dreamforce 2026.