Understanding sentiment to make super service bots

October 26, 2018

Bots are in their infancy, but they are poised for explosive growth provided they can figure out what we want and understand our requirements in intent-rich moments when we need their assistance most. Our President Dave Westin chats with Susana Duran, Director of Engineering Mobile & Bots at Sage, to get her views on the ways industry will need to design train and teach bots to grasp our sentiment to deliver assistance that is extremely helpful and clever, and always sensitive to our needs.Dave: This week's podcast, we're interviewing Susana Duran from Sage and we're gonna be talking about bots and AI. Susana, why don't you just give us a little background on yourself as well as what Sage does?

Susana: Yeah. Thanks, Dave. Yes, I work for Sage. We are a business cloud company and what we do is building cloud solutions to help business in their accounting and finance. And we do this through cloud-based solutions, mobile apps, bots, and other products. I'm managing a team that's based in Barcelona and we had to focus it on emerging technologies. So mobile, bots, AI, IOT. But we deal also with other colleagues based across all states, so in different locations.

Dave: So I really wanted to do a dive into bots and how you guys got into it, what your experience has been, and how you utilize bots, and, you know, what type of AI you've put in place.

Susana: Well, our team started working back in mobile apps, so for our different products. And then we realized that some users, they really want to just use mobile apps in a different way. So we start creating bots, bots for, well, different features, different products. And as soon as we started doing this, what we discovered was that the most part of those users wanted to use bots to get information, to have proactive bots. So what we started doing was to start building customer care, customer service bots. So super bots for our products.

And we are getting quite a lot of success with this kind of bots, just keeping our customers, just asking using natural language, whatever they want to do related to our products.

Dave: Got it. You know, what insight did you guys look into in terms of developing the AI for the bots, you know, in terms of what functions or what interactions they would have, and how those interactions would go, you know?

Susana: At first, we need to define what is AI because AI can be, well, can be many things right now. Almost everything that is a bit smart can be considered as AI. What we use is NLP. So we allow our bots to use models that will ask everyone, "Do you use natural language?" to just interact with our bots. But we also use training and models that are allowed in our bots and others products to be smarter enough.

So we create models, base it on big data and we train them to get better insights from our products, and just share those insights through our bots. So what we do is just make sure that the data we use to train those models is good enough and it's safe enough just to share everything that we're getting, and it's relevant for our users in the best way, so through bots in this case or even through the mobile apps or the web products.

Dave: So, you know, there's been a lot of talk about ethics and principles and best practices. Can you talk to some of that and how you guys have incorporated that?

Susana: Sure, sure. This is something that has been discussed, as you said, tons of times and the big companies had also quite, well, aware of the dangers of developing AI without any kind of responsibility. And what we think is that with this great power, what comes next is a great responsibility as well. So what we are pushing is a kind of core principles just to make sure that everyone is aware of what can happen if we don't develop with enough...being conscious of what can happen. And this means that what we want is to have AI reflecting the diversity of data, the diversity of the society.

So, make sure that we are filtering bias and negative sentiment of the data we use to train our AI so that we will make sure that AI is not perpetuating stereotypes. Also, we would want to make sure that technology, it's not allowed to become too clever, to be accountable for its decisions and actions. So if we don't allow human experts to behave in…well, for some behaviors, we shouldn't allow at technology to behave like that. So this is something that we need to always take into account.

Also, AI is going to create. Well, AI is also creating and replacing. That means that we'll be replacing with AI jobs, but we also need to ensure that we are replacing jobs that will be because…we'll be creating better jobs for us or better positions. These are the kind of things that we want to share with everyone and even making sure that with AI, we level the playing field just to using voice and our accessible solutions, to make sure that we are broadening the talent pool. So we are adding other people that right now is not playing at the same level than us.

So using this AI, we can add them. So that's the kind of things that we want to share with everyone and we are pushing hard to share with the community and understand and make them understand that AI is not just a technology thing. It's something that we need to be responsible for.

MGS: If you haven't already, be sure to check out the Mobile Growth Global Event Series with thousands of the world's top growth marketing experts come together to discover the latest in growth marketing innovation. We run events all around the world in cities like San Francisco, Los Angeles, New York, Berlin, Montreal, and more. And our candid panel discussions feature industry leaders including Facebook, Google, Uber, EA, and loads of other top-tier publishers.

So, attend our lineup of intimate workshops and gain the in-demand skills you need to stay on top of your growth marketing game. Expand your professional network, and build meaningful relationships in a fun, friendly, publisher-first environment. When and where can you get in on the action? Well, just visit mobilegrowthsummit.com for a complete list of upcoming shows.

And if you should decide to join us, and I hope that you will, then be sure to use our special promo code, MGSPODCAST30, for an additional 30% off your ticket order. Once again, that's MGSPODCAST30, all caps, no spaces, for 30% off of your order. We hope to see you there. I certainly hope to see you there and please, enjoy the rest of our episode.

Dave: How do you track interaction with your bots and, you know, to see how the interactions were successful? Is that something you built internally? Do you use an external solution to track, like a dashboard or something to that effect?

Susana: Yeah. Well, since the beginning we wanted to track and we wanted to understand if our bots were successful or not. So we start using traditional metrics, mobile or web metrics. But we discovered that they were not applicable or they made no sense. So we start working with other companies. In this case, with Microsoft, to create better dashboards, to measure the user behavior, and just to get the relevant metrics for bots. In this case, we established some general metrics like distinct users, messages in, out, and even retention.

But what we found that was really, really interesting and useful for us was to measure how the users are interacting with the conversation. So, how many times we hit the, in terms like, the responses like, "I don't get that," or just measuring the sentiment of the conversation like, well, if someone is talking about our company in a negative way, we need to detect this. So these kind of things are very relevant and hard, kind of metrics make no sense in the mobile app, make no sense in the web app, but you need to make sure that you are measuring in a bot.

Dave: What were the steps that you took to implement a bot strategy? You know, you started from ground zero. You know, what steps did you take to make sure it was a right fit for you, number one, but, number two, the bots were used in the fashion that you wanted it?

Susana: Yeah. Well, of course, technology was an important first step. But then what we discovered and one of the things that we understood since the beginning was that conversation is the core of any bot. So, you need to design the right conversation, not just focus it on the happy path. It's just trying to have everything on a map, and making sure that the user will get the right response. Trying to avoid frustration which is something that right now many users are just frustrated after a few conversations, a few dialogues with a bot because they are getting the typical, "I don't get what you mean."

So it's just about having the right skills, the right person working on the conversation, and then getting right workflow, making sure that it's coherent. And it's showing the values of your company, and it's aligned with the rest of the branding, the rest of the features you want the show to those users.

Dave: Who handled that strategy? Was it IT-driven? Was it marketing driven? Who is the stakeholders that drove, you know, the bot strategy?

Susan: Currently, we have at Sage a kind of triangle where product management, product engineering, and also the experience design team is working together as owners of all the products. So we are trying to design those strategies in collaboration. So it's about technology. It's about what you want to add the strategy of the product, but it's also about, and this is very important, about the user experience. So how to design those bots, those mobile apps, in collaboration. We try to do this kind of triangle to make sure that everything works as expected.

Dave: Got it. You know, are there any other tips or advice that you would give to the listeners?

Susana: Well, if you want to create a bot, just make sure that the use case is relevant enough for your users to use it through a conversation. Anything can't be just created through a bot and that's all. Make sure they need the bot.

Dave: Got it. Are there good apps that use bots or are there good resources for people to research bots and AI? Do you recommend any resources that they should look at?

Susan: There are many right now, but if you go to Microsoft sites, Facebook sites, and the big ones, they will give you a lot of hints because we mainly use those platforms. So Messenger, Skype, Slack, and others. And there are many tips. I would say that Facebook has quite a lot of good tips that you can just follow through and they are very interesting. So it's a good way to find this. And, of course, at Sage, we are posting from time to time blogs and articles about this. So we are talking about bots, AI, and other stuff that can be very interesting.

Dave: Got it. Thank you. If there's nothing else, thank you so much for your time, Susana. We appreciate it and hope you had a great Mobile Growth Europe.

Susan: Thank you, Dave.

MGS: And that's a wrap of today's show. Thanks for listening to this episode of Mobile Growth. And a quick reminder to visit mobilegrowthsummit.com for a complete listing of our upcoming events. And don't forget to use the very special promo code, MGSPODCAST30, for 30% off your next order. And, remember, no matter what kind of app or business you have, this is the destination for everything you need to move the needle on growth. So until next time. Make it real. Make it matter. And we'll talk to you soon.



Leave a Comment

MGS Custom Ad Spot

Want to be featured here? Submit a request below and we'll contact you.