August 2026 Summaries

Merge for Workforce is presented as an AI governance and cost-management platform that provisions approved models, MCP servers, and skills directly into employees’ AI tools. It integrates with SCIM-compatible identity providers so access policies align with organizational roles and update as teams change, while a lightweight client deployed through existing MDM systems enforces policies on employee devices and blocks unauthorized models or MCPs. Administrators can assign model tiers by group or individual, routing high-volume tasks to lower-cost options while reserving advanced models for more demanding work. The platform automatically configures employee AI environments, scans for unapproved AI usage, and attributes each call to a user, team, model, and associated cost, enabling organizations to monitor and manage spending continuously.

Merge has introduced Merge for Workforce, a platform intended to help enterprises manage rising AI spending without limiting employee access to AI tools. Deployed through existing mobile device management systems, its desktop client inventories AI tools, unapproved MCP servers, and user-installed skills across employee devices, giving administrators visibility into organizational AI usage. Administrators can set model-access policies by team, route tasks to approved and potentially less expensive models, require approval for additional access, and enforce these policies directly on devices while preserving existing tools such as Claude Code and Cursor. The platform also tracks spending by team, tool, and model, identifies concentrated usage and spending outside routing policies, and is positioned as a way for organizations to reduce token costs while maintaining AI-driven productivity.

Merge describes using its Agent Handler with Claude to automate weekly sales pipeline reviews by connecting CRM data, call-recording tools, and destinations such as Slack, Google Sheets, or analytics platforms. The workflow requires companies to define their own pipeline fields and qualification rules, then focuses on weekly deal movement—entries, advances, regressions, delayed close dates, and inactivity—rather than static pipeline totals. For stalled opportunities, the agent can retrieve and quote relevant call evidence, helping teams discuss specific risks and decisions with supporting context. Reviews should be posted shortly before meetings so participants begin with a common view of developments, while the agent remains read-only in the CRM and flags questionable data for human correction. Merge emphasizes that AI can reduce manual reporting work and reconcile information across systems, but that salespeople and managers retain responsibility for relationship-based judgment and forecasting decisions. The company also presents workforce management features including SCIM-based provisioning, audit logs, security controls, and model routing intended to manage employee AI access to customer systems.

Merge describes how its Agent Handler can connect Claude to sales-enablement tools such as enrichment platforms, CRMs, email systems, calendars, and conversation intelligence software to support AI-assisted SDR workflows. The proposed approach divides work among research and writing agents, emphasizes source attribution, stopping enrichment searches once a reliable match is found, and preserving unknown information rather than guessing. It recommends restricting research agents to read-only CRM access, requiring human approval for CRM updates and sequence enrollment, and having representatives review and send concise, personalized outreach drafts themselves. Merge argues that this structure reduces repetitive retrieval, tab-switching, duplicate lookups, and manual data entry while leaving sales judgment and final communication decisions with human reps. The company also highlights workforce governance features including identity-based access provisioning, data-loss-prevention scanning, tool-call logging, and model routing intended to manage security and token costs.

Merge offers a solution for developers to efficiently query user, group, and access data from Okta without leaving the terminal by using Merge Agent Handler's Okta MCP server. This integration allows users to manage Okta API credentials centrally, handles rate limits, and provides a control layer to scope operations and maintain an audit trail. By connecting Okta with Claude Code, developers can directly query Okta data, such as user status, group memberships, and security events, thus streamlining workflows, especially during time-sensitive situations like incidents or access reviews. Additionally, Merge provides a framework for organizations to govern and secure employee access to AI systems, with features like SCIM provisioning, DLP, policy enforcement, and audit logging, ensuring centralized control over identity data.