Fireside AI

Today we're joined by Sarah Coward from In the Room.

Sarah, welcome to the show.

Thanks very much.

It's great to be here.

Good to see you and thanks for having me on today.

Great to have you here.

So for people who meeting you for the very first time, tell us a little bit about yourself
and in the room.

So I've been in the conversational field for quite a long time now, since about 2015.

And in the room we incorporated in 2017.

In the room what we do is we enable organizations to make their best expert, their best
brand ambassador, their best talent kind of available to their customers or their

audiences 24-7, 365 days a year.

So we do that by integrating voice recognition, conversational AI approaches, and
authentic content predominantly.

and this is exactly the right time to be doing that technology is enabling a lot of
interesting things we can do with voice and conversational AI now.

So I know you have a really interesting origin story.

So how about take us back to the beginnings of the company and where the company came
from.

Yes, and origin story really helped, I think, to kind of ground what we do in authenticity
and to think about ethics, authenticity and human behaviour, which I think has been a real

benefit for us.

So in 2015, I was working in the cultural sector and particularly working at the National
Holocaust Centre Museum here in the UK and the Holocaust Museum

They run a really special education program where thousands of kids and adults come every
year.

They visit the center.

And as part of their sort of learning about the history of the Second World War and the
Holocaust, they have the opportunity at this really sort of special moment in their

education program.

The kids file into the hall, sit down, and a Holocaust survivor or a refugee from the
Holocaust comes in.

sits down in front of them and this person who's often in their 90s, early 90s, shares
their story and what happened to them during that period and it's during that moment when

a lot of learning takes place but also lot of meaning for the kids as they understand what
happened to that individual person and their families.

So in sort 2014, 2015 the museum was thinking

How are we going to preserve that moment for young people and the next generation when
these witnesses of history basically are no longer with us?

So we were looking at whether we could use new technologies to create an experience
whereby the kids would in the future be able to visit the museum, see a projection,

digital projection of

survivor and be able to interact with them and ask them questions.

But it was particularly important that those answers were given in a way that was
authentic, that was actually what they said.

So that's where it came from really.

We worked on a big project with 10 Holocaust survivors, some of whom have passed away
since we worked on that, and kids are still learning from those people and are able to

access those conversational experiences with those.

survivors online and in person.

And know we often think about AI as being this really sort of future forward looking
technology, but I love this example of connecting people with the past and creating this

emotional connection with our history so that they can really resonate with it.

completely, was almost like a kind of portal through time really, you know, that in this
case, the survivors, the witnesses, they had their part of the conversation then, you

know, years ago now than when we recorded them, but we can have a conversation with them
now.

So it kind of creates this permanent conversation that's available to different people at
different times, which I think kind of frees up that sense of

connection with humans and takes it out of things that are live and into something sort of
more beyond time almost, which can be applied in all sorts of different industries as

well.

So one of the reasons why we then started the company was really thinking about that power
of this moment of a conversation.

And when you start to think how frequently we all use conversations as we are today, you

in everyday life, that a lot of our businesses and organisations and social contact is all
based on human connection and human conversation.

So once you think about the technologies in that way, then it enabled us to really think,
you how could we scale up this technology to create something that could support a lot of

different organisations and companies doing lots of different things.

And so we're at this point where we have a lot of interesting technology that's able to
help with this and thinking about it from this perspective of authenticity and emotional

connection is I think quite different from what a lot of other companies are doing.

So tell us a bit about, your views on the technology and how it can really help with that
connection and that authenticity of interactions.

So we do, although we don't exclusively on our platform, you can use synthetic media.

Most of the applications we have focus on using authentic video content and authentic
audio content.

And one of the reasons for that is there's a big difference from what we've seen in
something that looks like a person and

something that actually has come from a person.

And that authenticity, particularly in certain applications, is extremely valuable and
meaningful to the end user and the audience.

And I think often that can get overlooked because people are focusing on what can be done
with technology rather than always thinking about what does this mean for the end.

user, the person who's having that experience, and what is meaningful to them, what's
important to them.

So we've seen different applications where some of those things are sort of more important
at different times.

So maybe tell me a little bit about some of those applications.

Where is it important to maintain this human connection?

Because I can see we're using technology in a lot of ways.

And sometimes it doesn't really matter what you're interacting with.

You could be chatting with ChatGPT and it doesn't matter that it's a synthetic answer and
it hasn't come from anybody.

So where really are those applications where it matters that we're talking and hearing
from a real person?

Yeah, no, perfect.

one of the frameworks we use for our experience is that tend to be three types of reasons
why people want to talk to a specific person.

Either because of who they are.

So they're a talent, they're a celebrity and actually kind of

getting to know that person in an authentic way and knowing that they have said those
specific things and that they've acted in that way is quite important.

Secondly, what that specific person knows, so their level of expertise.

So somebody introducing themselves as, I'm Dr.

Smith, I have worked for 30 years in this particular medical field is

very different in the context of building trust with somebody than a kind of generic type
application.

And thirdly, what people have experienced or what they've been through.

So for example, almost taking inspiration as well from the Holocaust survivor testimony,
those are individual human experiences that people have gone through.

And similarly with people who might have experienced

heart disease or cancer or have run a marathon, people talking about things that they as a
human have experienced, it has meaning when it's authentic in a way that, know, chat GPT

talking about...

their experience of running the London Marathon isn't meaningful in any way that a human
would connect with because everybody knows that one thing we know that chat GPT hasn't

done is run a marathon.

So that makes a big difference.

But I would like to give maybe one example where it's really important.

So one application we have for our technology is a school that wanted to give parents the
opportunity to

connect with and sort of meet the head teacher and ask questions to the head teacher.

Now, if you're thinking about it purely from a content perspective, you could have a text
chat bot answering questions about the school.

You know, what are the sporting facilities like?

How far is it to the bus stop?

Whatever sort of questions there are.

Has it got a swimming pool?

lots of examples of people trying to use that technology for question answering in chat
text chatbots right now.

So it is one of the ways that people lean on straight away, I think.

Completely.

But I think the interesting thing is, for example, parents aren't actually wanting to meet
the headteacher to ask them those sorts of questions.

They're wanting to meet the headteacher because they want to know whether they trust the
headteacher with their child and whether they like the ethos of the school.

So people kind of think it's an FAQ, but actually it's more about understanding.

the people and the culture and things like that, which you really get from a level of
authenticity.

And you touched a little bit there on trust and trust of people, but also trust of
technology, which I think is a topic people are thinking about now as we bring AI more and

more into the fold.

So obviously you're talking about videos of people and you mentioned text interactions as
well on chatbots.

What's your view or what have you seen about how people feel when they're thinking about
interacting with a video or

with text or with voice?

Do these modalities make a difference to people's trust in what they're hearing and seeing

So I think naturally humans were social creatures for thousands and thousands of years as
we've evolved.

have predominantly honed all our skills in assessing trust.

This is such a fundamental kind of biological instinct.

Whether you trust people or not.

and your connection with other people from a social point of view is fundamental to
survival.

That's why humans are so good at it.

video adds an additional layer of information that we as human animals use to pick up on
cues that are important to us that we don't even know that we're assessing, but we are

assessing.

So I think it's not video.

video isn't always necessary, but I think it adds a layer that helps to develop that trust
when it's needed.

Yeah, and I feel like maybe if you're seeing someone and especially if you know that that
person has said those things, maybe that perhaps makes you more likely to trust what

you're hearing in response.

And maybe that gives you a better sense of the person that you're talking to and how they
behave and how they present themselves, which you couldn't get from a text-based medium.

Exactly, exactly right.

I think there are things around voice which have a different type of connection and as a
number of people have studied, the ability for us to detect the uncanny valley aspect in

voice is kind of easier to get over those issues in voice than it is necessarily with
visual input.

So I think that's quite interesting and I think having this sort of multimodal
conversations, video, voice or text or a combination of all of those things will be really

interesting as things go forward.

But also the transparency requirement I think will be really, really important as well.

we talked about a lot of cases where you can build trust by being authentic and using
technology to help foster that authentic connection rather than a synthetic one.

what other, maybe on the flip side of that, the use cases or the places you've seen where
it doesn't really matter, people don't mind that much, they don't care so much who they're

talking to and whether it's a person or not.

Have you come across examples like that?

Yeah, yeah, so we kind of have a framework that we use which kind of has different aspects
to it which kind of give us clues as to how important that authenticity element is and so

the further out on this framework the more important that human connection is and then the
closer to the middle the less important it becomes.

So for example if you're wanting to inquire about your banking account

or something like that.

People are less, it's a different, it's more transactional.

So some of those transactional information-based applications don't require as much, in my
opinion, human-based trust.

You need trust in the institution that's providing the chatbot or the conversational
interface, but you're not seeking to build up information or trust with a particular

person or gaining

information that is kind of importantly based on that sort of trust.

that makes sense.

And coming back to what you said about the cases where you see it's really helpful, sort
of the personality, the expertise and the lived experience of somebody, you can really see

that building up the connection with the person there is more important, whereas the bank
or customer service may be very ad hoc, you don't need to stick around and build a

relationship with them, you just need to get what you want and get away.

That's right.

then at the same time, I think there are some areas of product.

So we do have applications where our conversational technologies are used within the
context of product sales.

But I think that's particularly powerful in product sales or service sales, where there is
an element of sort of jeopardy or trust or relationship required.

So give an example we have

a company that does tattooing and piercing that uses our technology.

And when people are shopping for those type of services, actually, if you're going to go
and get a tattoo somewhere, you want to know that those people are well qualified,

trusted, they look like they know what they're talking about.

So that is a transactional conversation.

But what you're purchasing or what you're booking

kind of has a high level of jeopardy.

Same thing with healthcare products, vitamins, anything that you would consume or put
yourself at risk with, really requires people to build up that trust online in quite a

robust way.

when there's a risk of something going wrong maybe or something harming you maybe you care
about where you're buying your services from and who's on the other end.

Or financial advice.

If you're going to be working with a financial advisor for a few years perhaps, then being
able to meet them or see whether you get on with them before booking an appointment would

probably help you vet people that you didn't want to work with quite quickly as well as
the people that you did.

Great.

I really love this focus on using this technology for building some of these human
relationships and helping foster empathy and connection in a way that other companies

aren't necessarily thinking about.

So as you you built up this technology over the past decade coming from the Holocaust
Museum all the way through to In the Room Today, what are the things you've learned about

how people interact with technology and the wisdom you could share with people?

So I think one kind of fundamental thing is that we kind of see conversational experiences
can be seen in two kind of modes almost.

There are conversations where the organisation and the conversational experience is
designed for people to sort of dwell in an experience.

The conversation itself

is the output.

So fan experiences, celebrity work, where people want to spend time with that individual
and ask them questions as if they are lucky enough to have that moment.

And it's kind of experiential conversation.

And then on the other side, there's a kind of directional conversation.

where the organisation using the technology is aiming for that conversation to do
something, which is a lot of the applications that we see sort of in commercial spaces

where you're supporting product purchase or ticket booking or meeting booking.

And obviously the way in which you think about each of those is quite different.

The other thing is

Again, going back to what we were talking about earlier, is whether people see that
conversation as being from an AI or from an individual or authentic or non-authentic is

really important for the end user to understand.

know, have they been handed off to a live agent?

is the information they're getting directly from the individual they think it's from, is
fundamental to how people will feel about that experience after they've left.

So all of the elements around trust and understanding of what you are experiencing in the
conversation is fundamental to how they'll feel afterwards and whether they feel exploited

or whether they enjoyed it.

to what extent have they built a good relationship with the company.

And I feel like some of those aspects are worth thinking about whatever sort of
conversational experience you're designing.

You have to think about what you're doing and who's going to be on the other end and how
they're going to feel at the end of it.

Definitely, definitely.

And I think that comes down to sort of thinking about the legal aspects as well as more
regulations come in.

And obviously some of those might stay, some might go, they might be more restricted, they
might be less restricted.

But from a moral, ethical and brand perspective, companies have to think through how are
they being perceived by people?

when they're employing these types of conversation technologies and is it supporting trust
in their own brand as well as trust in the experience itself or not.

So a lot to think about and a lot for companies I think to take away from from your
experience in building this sort of Interactions and to finish this up for the day.

Where can we find out more about you and in the room?

So you can come and see the website which is intheroom.global and that tells you a little
bit about what we do and also explains a little bit about sort case studies and

organisations that using our technology.

and I'll put the link to that in the show notes so people can click through and go
straight and see what you're doing over there.

Perfect.

Great.

Thanks so much.

much for your time, Sarah.

Great to chat with you today.

Yeah, great to see you Catherine.

Thanks so much for having me on.