Agent.ai has raised $2.7 million to give customer service representatives an artificial intelligence copilot to craft responses for customers with complaints.
The Sunnyvale, California-based startup recognizes that messaging is growing rapidly as a channel for customer service interactions. While messaging can be efficient and asynchronous, most apps and websites provide a poor customer service experience and cannot easily or economically integrate messaging or bots into their existing infrastructure. Agent.ai is building a platform to help address that.
The funding comes from a number of individual angels, including David Zhao (Evie Labs, ZumoDrive), Ethan Davidoff (RiskIQ, Mixrank), Dave Liu (Bowers & Wilkins, Jefferies), and Agent.ai cofounder and CEO Fred Hsu (Manage.com, Oversee.net). The seed round will accelerate Agent.ai’s platform development, as well as expand its onboarding and sales teams to meet growing customer demand in 2017.
Current bot-based messaging solutions often fail to provide humanlike, customized responses and are built with a one-size-fits-all mindset.
Agent.ai can be used either as a full-stack customer relationship management or as a layer on top of existing channels and systems such as Salesforce, Zendesk, Facebook Messenger, or Slack.
The Agent.ai platform uses event-based analytics to extract customer and session context, bringing the most relevant information to agents’ fingertips during customer conversations, resulting in efficient ticket resolution. Then, Agent.ai’s automation engine introduces AI gradually, learning from the most accurate source possible: a company’s own agents.
Agent.ai’s automation increases over three stages: The first stage consists of the AI engine passively monitoring agent interactions and company logs, as it learns to suggest and mimic the best agent responses.
The second stage is called copilot. Copilot suggests the appropriate response to customer questions; these are based on customer and agent interactions, support ticket resolutions, company knowledge bases, FAQs, and data from custom selected app events. Human agents can accept, edit, or reject a suggestion, which further aids the system in learning faster.
As the system grows more confident in its responses, agents can turn on autopilot mode. Autopilot handles routine questions so agents can focus on more complex inquiries. If a problem cannot be easily solved, the ticket will be automatically and seamlessly escalated to a human agent. But the autopilot isn’t meant to replace humans, according to Hsu. Rather, the focus is on complementing humans.
The company, publicly launched in October 2016, already has hundreds of customers and continues to add features to its full-stack, AI-enhanced customer support system.
“This funding will allow us to continue scaling our team and deploy the Agent.ai platform to our robust pipeline of customers,” said Hsu, in a statement. “With messaging-based customer service coming into the mainstream, Agent.ai brings together the best of CRM systems and machine learning bots. The system is built to aggregate customer and event data in one place, resulting in a more powerful and accurate offering for companies. Meanwhile our bot-based automation learns from, and mimics, humans to add a massive efficiency multiplier to customer service teams.”
Agent.ai’s software platform is available. Companies can send their first 5,000 monthly messages for free. Thereafter, companies can pay as they go, with each message costing 5 cents across any customer channel. This usage-based model increases flexibility and savings for companies. Enterprise plans are also available.