TORONTO – A year ago at Sage Summit in Chicago, Sage introduced Pegg, their chatbot. Now, having taken an advanced position in the application of machine learning technology to their software through the bots, Sage is expanding their use throughout their portfolio. They have also called on the industry to adopt a common set of principles in the application of these AI tools to business.
“Sage is a leader in the AI space right now,” said Kriti Sharma, Sage’s VP of Bots and AI, in the opening keynote at Sage Summit in Toronto on Tuesday. “The principle behind bots is that you chat like you are messaging a friend.” That provides a more user-friendly interface for interacting with the software, and getting insights and information from it.
Bots are especially valuable to the kind of small business customers that make up a large part of Sage’s customer base.
“In the small business segment, 80 per cent don’t use proper accounting software,” said Nick Goode, Sage’s EVP of Product. “85 per cent of apps that people in this segment download, they don’t use. We think that doing things through a familiar chat interface would make it easier for them, and will encourage them to get more use out of their software.”
Goode indicated that Sage has been following a multi-pronged strategy in this area since Pegg was introduced last year. He hired Sharma to her newly-created position to create a champion and generate managerial oversight. Sage has also partnered with companies like Facebook and Slack to further develop their chat strategy.
“This message is so important, and goes beyond Pegg,” Goode said.
Sage announced in the keynote that the Pegg chatbot has been integrated into Sage One, their startup-level cloud offering for small business. It will be available in Canada next month.
“The integration into Sage One is very real,” Goode said. “It does things like tell you that your data on invoice payments this month was better than last month, and rewards you with a badge.”
That’s only the beginning, however.
“Bots aren’t just about Pegg,” Goode said. “We have built a bot API and we are able to add these conversation APIs into other products. It will be in Sage 300 and Sage X3 and Sage People in the future. With Sage People, we have already done a proof of concept.”
Goode also said that Pegg has come a long way in the last year.
“At the beginning, it just stored data and could give a little bit of feedback,” he said. “Now we have built a big framework around it. That’s the real product we are building – the bot platform itself. It lets us build bots into all our platforms, and lets automated data be pushed out to users.”
Goode said that the commercialization of Pegg was always a three-stage process.
“First we opened up the bot framework with Pegg,” he indicated. “Then we began building out this framework. The third stage is to open it all up to partners.”
Sage is further along the road with bots than some of their competitors, something that Goode said they intend to develop further.
“If you don’t innovate, you die,” he said. “That’s the rule of software. We’ve spent a long time apologizing for the fact that we were late to cloud. We don’t want to be catching up everyone. We know some people see this as a gimmick, but we see it as changing the way you interact with your software.”
Zack Thomas, VP, Global HCM Product Marketing at Sage, said their use of bots is a natural extension of what’s being done in the consumer space.
Sage also called on the technology industry to ensure that AI applications for business are developed responsibly. They articulated what they call the ‘Ethics of Code,’ a set of values they developed while building Pegg.
1.AI should reflect the diversity of the users it serves, by developing effective mechanisms to filter bias as well as negative sentiment in the data that the AI learns from.
2.AI must be held to account – and so must users. Because users start to trust the AI early, the AI needs to be held accountable for its actions and decisions, just like humans.
3.Reward AI for ‘showing its workings.’ Most AI today has no cognition of what it is saying. Only broad listening and learning from diverse data sets will stop this from becoming a problem, and a reward mechanism when training AI is one way to do this.
4.AI should level the playing field. IT needs to push voice technology and social robots, which provide newly accessible solutions, specifically to people disadvantaged by sight problems, dyslexia and limited mobility.
5.AI will replace, but it must also create. New opportunities will be created by the robotification of tasks, and humans need to be trained for these prospects.
While Sage’s progress in voice was not demonstrated at the Toronto event, Kelly said they have made great strides there.
“We actually demonstrated this at our event in Atlanta a month ago,” he said. “The best UI is no UI. It’s the human voice. We demonstrated voice activation on Alexa, and how it can give an opportunity to people who have been previously disadvantaged. If Canada had been in the near-term launch plans for voice, we would have done that here. But we are working on this strenuously in the labs. Voice is now over 95 per cent accurate – up from 70 per cent five years ago – and it will be 99 per cent in a few years.”