LoginSign Up
← Back to Support
Weekly · Support · Week of September 7, 2026

Customer support AI gains governance layers as model provider flexibility expands and feedback tools open to ChatGPT.

ai-governance-in-supportmodel-provider-flexibilityfeedback-ai-accessself-service-provisioninglive-product-data-for-support-ai
O
Generated 20m agoDrawn from 6 products

The week in customer-support

The defining pattern this week was AI governance becoming a shipping priority, not just a compliance checkbox. Xurrent added guardrail reviews and an audit log to Sera AI — the infrastructure of AI accountability that enterprises require before moving AI support agents to production. Sleekplan connected its feedback board to ChatGPT, letting teams triage and manage customer feedback from inside an AI assistant, while also shipping anonymized public board views so feedback can be shared externally without exposing personal data.

The second shift was model provider flexibility. LiveAgent added Gemini and Anthropic as first-class AI providers for answer suggestions and AI agents, alongside the existing options. Support platforms that tie customers to one AI provider are becoming an exception rather than the rule.

Leaders

Sleekplan shipped two sparks in the same window. Workspaces can now connect to ChatGPT, letting teams triage, respond to, and manage the feedback board from inside an AI assistant — without needing to switch to the Sleekplan UI. The anonymized public board feature makes it practical to share feedback publicly with customers without exposing personal data. Sleekplan is building toward a feedback management workflow that an AI can participate in as a primary interface.

Hatz AI shipped Activate, a dedicated MSP command center for managing AI adoption across client organizations — Shadow AI scanning, PSA sync, and deployment governance in one surface. In the same window, phone agents became self-service for tenant admins: submit a request, the MSP approves, the number is provisioned — removing the MSP admin as a required step for each new phone agent deployment. Hatz AI is building the operational layer that MSPs need to manage AI support infrastructure at client scale.

LiveAgent 5.68 added Gemini and Anthropic as first-class AI providers for answer suggestions and AI agents, alongside voice transcription for calls. Both OpenAI GPT models and now Google and Anthropic models are available within the same support platform. LiveAgent also began its Svelte 5 migration — an infrastructure update that will take several releases but signals a front-end modernization commitment.

Xurrent added guardrail reviews and an audit log to Sera AI, its AI support layer. Guardrail reviews examine AI output before it reaches end users; the audit log makes every AI action inspectable after the fact. Combined, these shift Sera from a productivity feature to something enterprise security teams can accept in production. Xurrent also shipped an iPaaS Connector SDK, letting teams build integrations with enterprise platforms directly.

Wildcards

Chatway's AI agent can now read a merchant's WordPress site and WooCommerce store directly — published pages, posts, products, and store data — and reference that information in support conversations without a human configuring a knowledge base. For e-commerce support use cases, the AI that can read the actual store data is meaningfully more accurate than one that works from a manually curated FAQ.

Themes that compounded

  • AI governance infrastructure is shipping: Xurrent guardrail reviews and audit log and Hatz AI Shadow AI scanning both add the accountability layer that enterprise AI support requires.
  • Model provider flexibility is now table stakes: LiveAgent Gemini and Anthropic addition means support platforms that lock to one AI provider are now an exception.
  • Feedback management is opening to AI agent access: Sleekplan ChatGPT connection is the first customer feedback tool to make the full feedback board manageable from inside an AI assistant.
  • Self-service provisioning is reducing MSP admin overhead: Hatz AI tenant self-service phone agent deployment removes a human bottleneck from each new support capability deployment.
  • AI agents are gaining live product data access: Chatway WordPress/WooCommerce read access and Respond.io GPT-5.6 upgrade both improve the AI's context quality.

Watch this week

Xurrent guardrail reviews and audit log are the features that enterprise procurement will ask about — watch whether the guardrail configuration is expressive enough for specific industry compliance requirements (HIPAA, SOC 2) or whether it requires additional policy engineering. Hatz AI Activate is an MSP-targeted product; adoption will be visible through partner announcements and MSP community discussions. Sleekplan ChatGPT integration is early — the key question is whether the AI can correctly distinguish actionable product feedback from general customer noise when triaging at scale.