Airkit.ai is positioned as a platform for building and customizing autonomous AI agents to support employees and customers around the clock, and includes full integration with the Salesforce ecosystem.
In terms of use, the platform addresses two major directions: customer service and employee service. Customer service scenarios include: resolving open cases, drafting follow-up emails, escalating to experts, routing incoming calls, retrieving customer history, logging interaction notes, scheduling technician visits, updating ticket status, confirming parts inventory, and more. Employee service scenarios include: handling leave requests, answering HR policy inquiries, onboarding new employees, screening inbound leads, updating opportunities, completing renewal deals, resolving help desk tickets, remediating system alerts, resetting user access permissions, and more.
In terms of main capabilities, the platform emphasizes the following points: First, "out of the box," it can launch agents for roles such as sales, service, and HR, equipped with role-specific subagents, prebuilt actions, and Salesforce context; second, "continuous work," agents can pursue goals across days and weeks, retain context, adjust plans, and complete multi-touch workflows under human oversight; third, "continuous improvement," built-in tuning partners can discover bottlenecks, understand performance, and provide actionable recommendations; fourth, "teach once, use everywhere," create reusable capabilities for AI teammates and bring them into existing work tools; fifth, "agent teams," can route work, share context, and coordinate specialized agents across complex customer and operational journeys.
Product modules include: Agentforce (formerly Einstein Copilot), which is a proactive, autonomous AI application that can answer questions, take actions, and improve productivity; Agent Builder is a low-code tool that configures agents through subagents, natural language instructions, and an action library, and within it you can observe agent action plans and test responses; Agent Script provides builders with precise control to create business-ready agents; Agentforce Voice brings AI voice capabilities to every channel; Agentforce Observability monitors, analyzes, and optimizes agent performance in near real time; Agentforce MCP Support enables secure open interoperability through verified MCP server support from AgentExchange partners; Multi-Agent Orchestration supports building collaborative AI agent teams.
Agent types cover service, sales, marketing, commerce, and more: Service Agent replaces traditional chatbots, without the need for preset scenScenes can handle a wide range of service issues; the Sales Development Representative (SDR) interacts with potential customers around the clock based on CRM and external data, answers questions, handles objections, and schedules meetings; the Sales Coach uses Salesforce data and generative AI to provide personalized role-playing; the Merchandiser assists with e-commerce operations; the Buyer Agent enhances the B2B procurement experience; the Personal Shopper acts as a digital concierge; and the Campaign Optimizer automates the full marketing campaign lifecycle.
In terms of how it works, autonomous agents use large language models (LLMs) to analyze and understand the full context of customer interactions or automatic triggers, autonomously reason about the next decision, generate responses that align with brand voice and guidelines based on trusted business data (including Salesforce CRM data, Data 360 external data, etc.), operate around the clock across platforms such as self-service portals and messaging channels within set guardrails, and can escalate complex issues beyond their scope to human agents. Agentforce's starting point is the Atlas Reasoning Engine, which breaks down the initial prompt into smaller tasks, progressively evaluates and recommends a plan of action, and continues until a complete answer or action is produced.
In terms of usage, enterprises can use existing Salesforce platform tools through Agent Builder to create standard and custom sub-agents and actions, and can use prompts based on trusted enterprise data, Flow automation, MuleSoft API connectors, and custom code and business logic in Apex and JavaScript.
Applicable scenarios include: providing round-the-clock customer-facing support, such as answering questions, resolving cases, managing orders, and troubleshooting, with benefits including shorter response times, round-the-clock service, and cost savings; employee support, providing each employee with a highly specialized AI agent dedicated to their tasks, with benefits including improved productivity and faster resolution; appointment scheduling, arranging appointments based on history and context, sending follow-ups, rescheduling, and reminder notifications; sales development, autonomously answering product questions around the clock, handling objections, and booking meetings for sales representatives; product recommendations, automating campaign information communication, booking sessions, and providing event logistics and surveys.
Regarding limitations and costs, the materials mention that every Salesforce customer can start using Agentforce for free through Salesforce Foundations, and pricing options include Flex Credits, Resolutions, or per-user licenses; specific cost details and deployment limitations are not yet provided.
In addition, the platform provides a library for builders to learn how agents are built and to find suitable Agentforce consulting partners.
