We’re already quite familiar with artificial-intelligence apps like Siri, Alexa, Google and Cortana at the consumer level. And we know that AI and machine learning are slipping into more and more enterprise apps—at the industrial, value-chain and customer-facing levels, for starters.
Now there’s an example of an internal enterprise app that uses AI for another purpose: Spoke, a young (1½-year-old) San Francisco-based startup which launched on March 28 a cloud-based tool to manage internal requests for midrange and smaller companies.
Conventional ticketing systems at larger companies, such as Atlassian’s JIRA Service Desk or ServiceNow, are designed for duty that usually includes project management or customer service. Spoke fills this need for IT, HR and operations departments at midsize companies (50 to 500 employees).
By using AI to tee up and streamline internal-request management, Spoke portends to provide a new way for line-of-business employees to manage and search for information at work and help teams be more productive. And get this: Employees can use any channel that’s handy for filing a ticket request.
Built Atop a Knowledge Base
“We’ve built an internal ticketing system on a knowledge base,” CEO Jay Srinavasan told eWEEK. “We collect requests from multiple sources like Slack, email, messaging, all in one interface, and we use machine learning to learn the answers to common questions and automatically handle self-service requests for the agents directly in the product.
“Thirdly, we’ve built Spoke so you don’t have to have a separate ticketing system for your IT, HR or ops teams—it’s one interface for all these functions at once.”
Here’s how Spoke works:
- Employees can ask Spoke for anything they need in the places where they already work using Slack, email, SMS or the web;
- if Spoke’s AI finds an answer in the included knowledge base, it automatically provides a friendly response;
- if Spoke doesn’t (yet) know the answer, it intelligently assigns a ticket to the right team in the organization; and
- as people respond to tickets, Spoke gets even smarter about responding to and routing requests.
Spoke’s secret sauce is in its use of AI, natural-language processing and empathic design; it uses a machine learning algorithm and system that learns from responses in real time, Srinavasan said. This is in stark contrast with standard machine learning setups that accumulate a batch of training examples and train over time, with the end users not seeing the benefits in real time.
Learns Responses Quickly
“We’re literally the only company doing this that has built machine learning into the product from Day 1,” Srinavasan said. “It will only take one or two tries for Spoke to learn something and handle it for you from then on.”
For example, if an employee calls HR and wants to make a change to his 401K, that’s not necessarily something an HR manager would do, Srinavasan said. “That would be a self-service-type answer; Spoke will find that answer for you and return it,” he said.
Spoke already has onboarded multiple hundreds of customers in beta trials with a waiting list of more than 5,000 companies, ranging from oil and gas companies to tech startups. About 25 percent of the beta testers have rolled out Spoke to multiple departments, Srinavasan said.
Pricing is based on total number of users plus number of teams (different departments / job functions) using it:
- Option 1 (Starter): 1 team, $1 per invited user
- Option 2 (Standard): up to 4 teams, $2 per invited user
- Option 3 (Plus): 5+ teams, $3 per invited user
The company was founded in 2016 by ex-Googlers Srinivasan, Pratyus Patnaik and David Kaneda, and it raised $28 million from investors that include Accel, Greylock, Felicis Ventures and Webb Investment Network.