Organising your life with AI & chatbots - an interview with Lisa Matthews

Host: Marco Oliver, Client Success Director at We Build Bots

Guest: Lisa Matthews, CEO at HellyHolly

Marco Oliver: Hi Lisa. Thank you for joining us today.

Lisa Matthews: Hi Marco.

MO: You have a really interesting background. You started out as a doctor in philosophy and then went on to become the founder and leader of an engineering consultancy. Furthermore you went on to be a co-founder and director of another consultancy providing expert witness and forensic services to the construction industry. Can you tell us a little bit more about your journey and how you’ve ended up where you are today?

LM: Yes, sure. I started by studying engineering and did my PHD in Bath. That was really about looking at problems, understanding how you go around solving problems and finding something that you could spend a good chunk of time researching and help push forward the state of knowledge in engineering. After that, I joined a consultancy, a global multi-disciplinary firm, and I worked in all sorts of different bits of the business in Dubai and in Birmingham and projects all over the world. It was great to see how all sorts of different disciplines need to work together to get stuff done.

Then we founded our consultancy coming up for six years ago and that was really to give my husband a bit of better work/life balance. He’s a structural engineer and we thought that we could do something that would give him flexibility and mean that he could see the kids more. So, we did that and that’s been successful so far.

MO: Great. And now you are the CEO of HellyHolly and that’s really your main company at the moment, isn’t it?

LM: Yes, that’s right.

MO: So, your flagship product is Our Canary and that’s what we’re here to talk around today, which is a chatbot essentially. I’ve had a little look at your website and it’s very minimalist.

LM: It is, yes.

MO: It says that Our Canary is a life engine and it’s there to help people juggle life in a single place for all the family. Can you give us maybe a rundown on what Our Canary does and how it works?

LM: Yes, sure. When we started out with HellyHolly, we didn’t know what kind of product we were going to build, so the company was founded to solve the problem of all the stress and conflict and guilt that you get in your life when you’re trying to juggle lots of things and stuff goes wrong. This was inspired by our own experiences. When we spent a bit of time trying to talk to other people and find out whether everyone else had all this sorted, were we  just missing out on some key piece of knowledge? We found that actually everyone’s struggling with this and we started learning about the impact it has on people’s lives. That’s when we realised that if we can do something valuable here it can really impact a lot of people and improve their lives, so that’s what we set out to do.

We didn’t know what we were going to build or even if it was a technology product or a tech solution. So, we started off just studying the problem and  looking at how people currently handle the problem. When things go wrong, why do they go wrong and how it impacts people. And that’s when we realised that actually, through technology we’ve got the opportunity to solve this for people in new ways that haven’t been available before.

MO: Okay, that’s interesting. So, what sort of problems were you focusing on as underpinning the solution?

LM: The key problems are that we’re living busier and busier lives. People are working more flexibly and they might have more than one job. The old model of one person goes out to work, one person stays at home, just doesn’t exist anymore. That’s not reality for the vast majority of people. And so, that means that you’re trying to share getting stuff done between multiple people. And it’s true at work as well. We’re collaborating more and sharing responsibility for achieving goals. The problem comes when you don’t have the tools that are really designed for that process. So, although we’re living and working much more collaboratively, most of the tools we use to organise ourselves and organise our lives are digital versions of analogue things we had in the sixties. If you think about a calendar, the design of a calendar has not changed from when it was a piece of paper on the wall to when you look at it on your iPhone , it’s essentially the same. It’s a database, so digitising it hasn’t actually delivered any better value to the user.

The tools don’t reflect how we’re working collaboratively and also, they don’t reflect this sharing of responsibility. Whilst we’ve got mobile technologies that help us work flexibly and remotely and in much more dynamic ways, it means that we’re just bombarded with information and notifications. We’ve got all this data that we’re trying to stitch together that’s all in different siloed channels and the only way we’re stitching it together at the moment and looking at the whole picture and trying to see if things are going to work is using brain power. When you’re busy, stressed and tired, the brain is not great at scenario planning, testing logistics, working out okay, actually is there a problem coming down the road which I’m not aware of. And this puts a base load of mental stress on people all the time to think, am I about to drop a ball?

MO: I completely understand that. I think that’s problems that we all face, whether there are two of you in a family or six of you in a family, I think you all have that kind of problem. We’ve got a shared calendar in my house so that’s a step in the right direction in terms of being able to be collaborative and put your events in but it’s not quite where it needs to be, is it?

LM: Exactly. When we look at how collaboration is done at the moment, it still relies on you chucking data out of your silo into a shared environment so that the single player mode is still what we largely live in. If you want to delegate or share something you chuck it over the wall and then you never see it again and you don’t know what’s happened to it, it’s reliant on you remembering to do that and then remembering to check in and see what’s happening. You don’t really know if the  person has seen this information - Is it going to work? Are we creating a problem for ourselves? So, that basic logical checking of stuff is currently only done in your own head.

MO: Yes. No, that makes sense. There’s quite a host of possible options you could have probably gone for in terms of Our Canary. You could have maybe had a website, an app.

LM: Yes.

MO: Some people say that’s old tech now so maybe that’s part of the reason, but what was your reason for choosing a chatbot as the main interface? What is the background for that?

LM: I still remember the day when the realisation dawned on me that we’d been for about three months thinking that what everyone needed was some kind of better calendar, like an actual clever calendar. A calendar that’s not just a database, which is what most calendars are. All of this stuff sounds really obvious in hindsight  but I imagined trying to sell this and I just thought, how can you convince people to pick up all the data of their lives and transfer it to yet another new tool and put it down somewhere else and then convince everyone they want to work with to also do the same thing? I thought, do you know what? That’s just not going to happen. There’s too much friction in that. People don’t want another tool, they want a safety net. So, they want something that’s catching the problems before they impact them.

So, when we looked at what people are already doing in their lives for this problem, we found that mostly – because there’s no product that this substitutes -  people are obviously still managing their lives at the moment, albeit with the stress and the problems that happen. We looked at how they do that and what people tend to do is hack together a bunch of existing tools, but typically similar flavours. There’s one or two main communication channels and two or three main data sources like calendars, and that’s when we realised that people are communicating non-stop about this stuff. They’re on WhatsApp, they’re in Slack, they’re on emails and all this communication is happening  so we need to be right there where people are trying to solve the problems. Where they’re talking about their problems and coming up with solutions, we need to be right there so it’s as frictionless as possible for the service to deliver the value. So, that’s when we realised we’ve got to go to where people already are, and where people already are is in chat. That’s when we realised we’re running a chatbot.

MO: That sounds great. And you’re right, I think chatbots are becoming second nature to a lot of people. They’re still quite exciting because not everyone has seen them, but people are starting to get more familiar. They’re coming across them more. I mean, you’ve got banks who are doing it now.

So, one of the things that is an evolution of the chatbot is the move to voice chatbot - is that something you’re also planning?

LM: Yes, definitely. It’s definitely on our road map. Our first product we’ve built entirely in Facebook Messenger because we were looking for a platform that – and this was pre-Cambridge Analytica – was widely accessible. A mature developer environment so lots of support, lots of documentation, and that we could quickly get up and running with something that would do what we wanted it to do and deliver a service that we could then start learning from people how they use it.

Moving to voice is definitely on the road map. The ultimate goal is to be omnichannel. So, across whatever integration somebody wants, whatever tools you choose to use for you, for your family or at work, that you can access the service through any of those integrations. But obviously we’ve got a plan which ones to do in what order. Voice is very much one of the priority ones coming quite soon. I think there’s a real timing issue in terms of general adoption in the market of devices, but our plan is to be ready to capitalise on that. And when you think about how the decision to go to chat was because this is where people are already communicating through the day. What they do when they’re at home is, say, they’re cooking dinner or making the kids’ tea or whatever and they’re talking about the same issues at the same time, that’s when they might be updating each other on, by the way, tomorrow I’ve got to go to Bolton for a conference, or whatever it is.

So again, it’s about being right there where the conversation happens. To be honest, a speaker in the kitchen is an obvious next move. The challenge I think that will be largely for us will be about capturing data, so when people want to tell the service about something. I think discoverability of skills is a challenge but also, it’s a bit harder to do search through voice. That’s a bit more of a design challenge. Ultimately, where our data and AI side of things is going, it’s for people to be able to ask questions, like I need to go to Paris for a meeting in two weeks, which is the best day? And for the service to give them an answer which takes into account the full context of all the spheres of their lives. So, it’s not just what’s in your calendar over two weeks, but it’s also what else have you committed to that’s not in your calendar? Maybe you’ve committed to it on a text message or in WhatsApp. What are all the important people in your lives also doing, and if you go away how will that impact them? So, that full context decision making is the goal and actually being able to ask a question of a voice enabled device, it seems a natural fit for that.

MO: Yes. That’s really interesting, I mean, that’s a massive level of integration potentially, if you’re going to bring all the sources of your life into one place?

LM: Yes.

MO: But if you can achieve that, I think that has got massive potential, hasn’t it?

LM: It has and actually there’s key technology trends being able to do this. You couldn’t have done this five years ago when everybody’s work calendars were sat on servers in the building, for example.

As more and more data migrates to cloud and as more and more of those services have public APIs or subscription APIs, that’s what’s unlocking our ability to get the data in in the first place. And then being present in chat and voice means that currently the lowest friction way of getting the data which is in conversations, what people talk about, the offline stuff that they do.

MO: Yes. That’s quite interesting actually then because if you’ve got information on my work life, you’ve got information on potentially events I’ve got with friends booked in in my personal calendar. And then are you looking at also maybe integrating external factors? So, if I said I wanted to go to Rome tomorrow, would you also then look at the flight timetables for Rome and give me information on that?

LM: The potential, once we’ve got all that data, is really massive. And what we’re being very conscious to do is driving the development roadmap based on usage and user feedback and data that we get of how people are interacting with the product. There are all sorts of ways we could offer additional services. So, if somebody knows okay, I know I want to do this thing, then there are all sorts of affiliate sales or things you could bring in at that point. We haven’t currently planned on – at least in our financial projections – generating revenue like that because we don’t yet know how sensitive people are going to be to the privacy around their data.

So, when you think about the type of data people are talking about, it’s the whereabouts of your kids, who’s going in to see an elderly parent this day, the next day, whatever. So, we take an approach to data which is comparable to a fintech company. It’s that level of sensitivity and that level of privacy that people expect. So, we have no plan to monetise the data through re-targeting advertising kind of stuff because it just doesn’t sit, I don’t think, with the type of problem that we’re solving. But being able to offer people the opportunity to achieve a goal – I mean, we started off really trying to be a painkiller for these real hair on fire problems. And I would love to move it into actual ways of helping nudge people where they want to improve their habits or behaviours they want to change. If they want to find opportunities to do new things, that would be great to be able to move into actually enhancing what people are doing with their time rather than only solving the big problems, which is valuable in itself  but to be able to offer somebody for example; you want to do Pilates? Well, here’s a bunch of classes that fit in and we know your babysitter’s free and we know all this stuff works.

MO: Yes. It really is becoming like a life engine, isn’t it?

LM: Yes.

MO: It’s covering all aspects of your life for you. That’s brilliant actually. So, in terms of coming back to that data point because it’s always an area of contention. Even if you anonymise data and you send it out because things are getting cleverer now including artificial intelligence. It can be used in good ways. That’s what most of us are trying to do but you could use that then to work out who somebody is if you passed in enough information.

LM: Yes.

MO: So, I think maybe given the amount of information you’ll potentially be collecting you’re going the right way but in terms of holding that data, obviously we’ll come on to how people can start working and then interacting with the bot. But if they want to go into your chatbot knowing that it’s secure, what kind of protection have you put in place for Our Canary? What are the data protocols and things?

LM: Obviously, we picked a platform because it was going to be efficient for us to build a first version in, but the public perception of Facebook and data has changed quite a lot of the landscape over the last year. How we’ve built the product within Facebook means that somebody interacts through the conversation and then when they want to tell us anything they open a web view. That way,  whatever they say to us in the message text obviously Facebook see but whatever goes in our web form in the web view is direct communication with us, so Facebook don’t see that.

MO: So, it’s a secure API effectively?

LM: Yes.

MO: That’s great. And the data is held – I’m assuming it’s encrypted at rest and in transit?

LM: Yes. All that good stuff.

MO: Good. That’s brilliant. I’m quite excited by that. So, for example, if I wanted to get involved, how do I go about signing up?

LM: Our very minimal website you mentioned at the beginning is our landing page.

MO: Minimal but interesting I should say. It is quite intriguing.

LM: So, there’s a sign up form on the website. We’re taking sign ups of our first cohort of beta testers. We’ve got the product in private testing at the moment. We’re building out a couple more features around the logic to do with looking at things that clash between different people. So, you can go to the website at ourcanary.com, fill in the form which is two fields, and sign up. And then we’ll let you know when it’s open for beta testing.

MO: Excellent. Okay, that’s great. I’ll definitely be doing that.

LM: Great.

MO: One of our previous guests, Zara Nanu, who I think you actually know?

LM: I do, yes.

MO: Yes. So, we talked about the gender pay gap and that’s quite a hot topic at the moment. And I know you are quite an active member in the Bristol Women’s Tech Hub.

LM: Yes.

MO: So, Lucy, our researcher, has put some stats together and I just wanted to get your thoughts on some of these stats. So, for example, 3% of females say a career in technology is their first choice. Only 3% which is very small. Why do you think the percentage is so low?

LM: It’s a really good question. I was at an awards ceremony recently where the speaker was talking about getting women and girls into STEM and talking about there needing to be more role models and we need to do things in schools and encourage this more in education. I think a lot of it really misses the point because I don’t think children are born innately preferring or discounting science and technology subjects. I think that’s a learned behaviour that happens once they come into contact with society and adults basically. I think there are all sorts of ways that we subliminally message to girls and boys about behaviour that’s expected of them and therefore what career paths are open to them, and I think it goes across all sorts of stuff.

I have one girl and one boy myself and  it’s been really interesting to notice how I feel differently about gender issues - I think it goes both ways. We send difficult and harmful messages to girls and we send harmful messages to boys as well. I feel like I’m constantly trying to protect both of them from that but there’s all sorts of campaigns around, for example, gender advertising of toys, around gendered slogans on clothing, you know, it goes through all these different factors that kids get impacted in. I think then that builds the perception in people that this particular type of problem solving isn’t for me or I wouldn’t be good at it. And then you’ve got to think about well, why aren’t there role models in the industry? And it’s fine to say some people say women don’t negotiate hard enough for promotions or pay rises etc, but I would say don’t try and fix the women, you’ve got to fix the system that is providing unequal opportunity. Anybody that starts to talk to me about gender issues and doesn’t start from a cultural standpoint that we need to look at culture in business, that’s really hard for me to swallow because I think you’ve got to fix the culture first.

MO: You have, yes. That’s a really good point. You mentioned not many role models in tech really. I mean, you’ve got people like Elon Musk, for example, who’s very prominent at the moment. And then you’ve got all the Apple people, you have Steve Jobs and people like that. Has there been any kind of women in tech that you’ve had as a role model?

LM: It’s a really good question. Yes, there are. There was a survey done last year asking people to name famous women in technology, and the top answer was Alexa and the second answer was Siri.

MO: Oh, really? Wow.

LM: And you’re like, okay, not even Sheryl Sandberg? You’re struggling to reach the top of the list. So yes, it is a problem but there are! Katrina Lake is somebody that I follow a lot, she’s the Stitch Fix founder and CEO. She was the youngest person to take a company in the States public and the only woman last year to IPO. So, I think that there are people around, but I think that if you only look to the media you get one particular version of what success looks like. It’s hard to keep reminding yourself though you can do things your own way and be successful if you don’t look like a 23-year-old white male Stanford graduate.

MO: Some of the people I’ve been interviewing, yourself included, you’re doing really well in the tech industry. Who knows, someone might be getting interviewed in five years’ time and saying your name.

LM: Well, there you go. You never know.

MO: That’s the way to look at it. Well, thank you very much for joining us. It’s been great.

LM: It’s been great to be here.

MO: We do ask all our guests one final question and that question is, if you had to live without any of your apps on your phone or on your laptop apart from one, which one would you choose and why?

LM: I did prepare for this question. I’ve forgotten what my answer was. I think the thing I use the most, apart from our own app obviously which we do use a lot and my co-founders are tweeting ‘don’t you just love it when you use your own app and it really improves your life’.  Anyway, probably my podcast app because I just find that there’s so much content. If you don’t have time to read very much the way to consume stuff is listening to podcasts and it’s a way to spend time with people you want to be more like that you might not otherwise have access to. It’s a way to up your consumption of new ideas, different perspectives, you know? I think if you really want to be able to think creatively and innovatively you’ve got to expose yourself to as broad a range of thinking as possible. So, yes, podcasts.

MO: Do you know what? I think I’d actually agree with you there. I’ve been asking this question a few times now and I haven’t been able to decide what app. But I think I would go with that actually. I’m an avid listener of podcasts, especially the We Build Bots, The Botcast podcast that we do (shameless plug)

LM: Yes, exactly.

MO: No, in seriousness I listen to podcasts every day and I think it’s a good way to get current information fed to you rather than having to go through the effort of searching on the web. It obviously gives you a bit more insight as well in terms of the topic. It’s not just a headline or a tag line that you’re reading.

LM: Definitely.

MO: There’s a real in-depth discussion about that topic. Yes, that’s a good answer. I like that answer.

LM: Good.

MO: Well, thank you very much again and I wish you all the luck with Our Canary.

LM: Thank you very much.