The Conversation Is The Interface: Experts In Your Pocket
Text-Messaging, Conversation Based Services
Text With A Lawyer, Text With A Doctor, Text with A Therapist.
These are examples of a new frontier of businesses that are springing up. These new types of businesses take the services we receive from in-person service providers (e.g. (dating, shopping, news, medicine, therapy, legal), and makes them available via mobile-first chat-based interfaces. In addition to expert services, we may see a second wave of online social networks & information sites (e.g. the first generation includes Yelp, Stack Overflow, Quora, Wikipedia, Google), that help us get to the information we’re looking for via chat-based interfaces and questions and answer style conversations.
I wanted to explore this space a bit, and think through which traditional businesses or services are good candidates for mobile-first chat-based interface approaches? How will these products evolve? Will this become a dominant paradigm or not? This post is a first stab at sharing some early thoughts on what these new text-based conversational interfaces might look like both in the near-future and further out.
Mobile has become the dominant medium, and we have shifted increasingly to text-messaging as a primary form of communication in our daily lives.
Regardless of how you feel about text-messaging’s impact on human society (e.g. is it good for us or bad for us — see footnote 1 at bottom for more on this), like any new medium, we should acknowledge the following points:
- It has its advantages and disadvantages
- It is now a core part of our lives
- Often times people prefer text-based, asynchronous, remote communication to voice based, synchronous, in-person communication.
- Text-messaging as a medium will drive a re-imagining and transformation of nearly all of the businesses and services that we interact with in our lives.
As an example, consider the following businesses that are already leveraging text to connect you to service providers. I’ve been using the following services myself to really understand what the text-messaging conversational interface feels like today:
First Opinion is texting with your Doctor
- There are many text with a doctor app, many of them verticalized — e.g. text with a dermatologist seems to be a popular one because you can snap a photo with your phone and send it for quick input from the doctor.
TalkSpace is texting with your Therapist
- I live in New York city and if you ride the MTA you’ll likely have seen many ads for TalkSpace. I gave their trial a spin and it was an interesting experience — not sure how effective it will be vs. in-person or audiovisual therapy, but definitely a great way to bring therapy more mainstream.
Coach.Me is texting with your Life Coach
- This is a goal-oriented, life-coaching service. My coach is texting with me to help me eat healthier, and I’m impressed by the quality of the service so far. It brings real value in terms of suggesting effective strategies for developing habits, but that may be because I got lucky and found a good coach on their service.
WunWun is text to shop/buy anything
- The other day I opened the WunWun app and just wrote that I want 4 lightbulbs from Crate & Barrel. I exchanged a few texts with the service provider on the other end and an hour later someone showed up at my house with 4 lightbulbs. This is like the Replicator from Star Trek that can make anything you want, but it just takes a little bit longer for it to appear at your house. WunWun’s interface feel pretty magical: just type whatever you want and then they’ll chat with you if they need further clarification.
What other professions/services can be delivered primarily through text-based, mobile, conversational interfaces? If you think about it, conversation is one of the most flexible interfaces that human beings have available to them. Conversation is how we already go about interacting with the services and people around us.
In some way, conversation can be considered the most versatile interface. Conversation has many aspects and functions to it:
- Give someone information about yourself or something you know. (Uploading Data)
- Here they can ask us questions about us so that we can give them the information that they need about us (Disambiguation, Requesting Specific Information)
- Ask someone questions about something, and learn from them (Search & Learn)
- Ask someone to do something that you want done (Action)
- Get to a point of disambiguation and clarity (Forming Understanding & Agreements)
Right now, most of the services we use in our daily lives: e.g. doctors, therapy, lawyers, involve talking primarily to humans. The first wave of chat-based apps just take services that are currently provide in person, and then provide them via mobile, text-messaging instead. This is the first wave of the text-messaging conversational interface: human-powered chat-based interfaces.
The second wave of chat-based interfaces revolves around artificial intelligence, which is a field that has been making incredible advancements in the past few years. Applications like Siri are an early example of how text/voice based, conversational interfaces, can let us learn about, and manipulate the world around us, not by talking to humans, but by talking to AI-powered conversational interfaces. I’m surprised that you can’t text with Siri yet — that seems like something Apple should offer since people may just not enjoy talking out loud.
Siri may still not be great for most use cases, but it’s getting better. And what we’re going to start seeing are verticalized versions of Siri. If we follow this trend forward, it becomes reasonable to expect that soon you won’t be talking to an actual doctor via text, but a Siri-like diagnostic algorithm that is having a conversation with you.
The next few years will usher in the first wave of text-based services. Many are already here, and they will likely still have us texting with humans on the other end e.g. text with an actual therapist, text with an actual doctor, text with an actual Fashion Consultant. These humans will have increasingly more sophisticated software on their end to make them very effective at their jobs. That model will last for a time.
But, as the film Her brought into mainstream awareness, eventually we’ll be talking to computers. Soon it will be: Text with an AI therapist, text with an AI doctor, text with an AI Fashion consultant. IBM just previewed Ross, their AI lawyer that you can pose questions to in natural language e.g. “What is the leading case in Ontario on an employee starting a competing business?” (example from the Ross Site). Other services such as Wit.ai let developers start to leverage machine understanding of language for their own applications (Natural Language Processing). Given these trends, for many verticals the time where we talk to AI instead of people may not be so far off. Her suggests that gradually we’ll be turning to AI based conversational interfaces even for something as fundamental as human companionship. But we’re not there yet, and we should reflect on whether it’s where we want to go.
How far does this conversation as interface paradigm take us & what industries will be impacted as a result? I’ve put together a list of existing companies/verticals that my friends and I have thought about as being ripe for re-imagining as a primarily text-based, conversational interface.
Here’s a list of companies I’m aware of that leverage chat-interfaces, and other markets that might be interesting to explore:
Verticals where people pay a lot for the expertise of the service person, and nothing is fundamentally lost by text messaging alone seem like the best candidates for tackling via text. Text-based messaging is asynchronous, so it has certain aspects that make it an effective way to connect people — e.g. you can answer your service provider when it’s a good moment to do so. On Coach.Me, I find I write to my life coach once or twice a week when I have a free moment.
Texting also allows you to handle way more customers, and work from home as a service provider (i.e. no real-estate costs).
The verticals above are a few that quickly come to mind. In an upcoming post I’ll go more in-depth on one conversational interface I think is particularly interesting: Programming (Software) Through Conversation — what if you could create software applications just by having conversations with Siri?
Here are a few preliminary thoughts for companies that are trying to build interfaces around these text-based chat interfaces:
Low cost to test ideas in this space.
If you have an idea in this space, e.g. provide legal advice via text, it’s dirt cheap to test — just throw up an SMS number via Twilio and put a service provider on the other end (e.g. a lawyer), and you’re in business.
Business Models & Cost Structures
- There are many different business models. e.g. charging for speed of response, charging for additional functionality like sending pictures (this is how many medical text apps work). What other variants will we see here? Will people charge different prices for different quality of providers — e.g. some life coaches cost more than others? Unlike Uber where you expect the ride to be roughly comparable regardless of the driver, with expert services there could be a wide range of service quality.
- One important question is do the cost structures for these types of businesses make sense, and in which verticals are they already practical? It may be the case that these companies don’t make much profit early on but continue to become more profitable as more and more of the conversation can be automated. In the markets I’ve considered this is often a big concern — are people willing to pay such a high price that it covers the costs of the person texting with them.
How many people can one service provider text with at once? I heard from a friend that some therapists on these types of services may be talking to around 50+ patients simultaneously! But they also get to work from home and make comparable income to working at a practice — if not more.
Conversation as Search
How does the conversational interface replace existing search paradigms? Siri is a good example of this but there will be far more.
Shifting from People to AI
For now, these conversational interfaces are powered by people. Where is it important to text with a person rather than with an AI? How does the transition from having conversations with people to having conversations with AI look concretely? What tools & data make people much effective at their jobs?
Turning Conversation Into Content & Data
Stack Overflow, Yelp, and similar sites are expert opinion networks where the questions and explicit reviews serve as the content. It’s interesting to think how in chat based interfaces we’re actually creating content within our conversations. For example, a lawyer might answer someone’s question that many later clients also have as a question — the first response has the potential to be turned into more-evergreen content & reference material for lawyers using the service when they encounter the question again. This can bring the cost of providing the service down over tie.
The digital storage of the conversation we have with our service providers is a new phenomenon. Today when people have in-person communication with their service providers much of the conversation is lost. When the conversations are happening digitally, it allows for the possibility of really large scale data collection & analysis that could enable these providers to offer higher quality service. Storing and sutdying all of these conversations may have it’s benefits, but it also scares me a little bit — for example, how long does TalkSpace store the conversations that patients are having with their therapists? It may make people more reluctant to have conversations that before they were having ‘off-the-record’ — so we may need to develop features that let us take things off the record.
Texting is definitely easy, it feels non-committal — and so that seems to be one of the big strengths of these types of businesses.
But, one thing to consider about conversation through text or voice is that it’s only one of the limited ways we interact as humans. As Brett Victor discusses in-depth in his recent talk The Humane Representation of Thought, conversation can be about much more than just text or voice. There are futures that we can strive for where we aren’t just texting and reading with our interfaces but having much-richer conversational interactions that leverage all of our senses — seeing & drawing, listening & talking, feeling & being guided — all at the same time. I’m not sure how smell & taste interact yet — but this TED talk suggest we should think about it: Design for all 5 Senses.
Today, our conversations with human beings in person are rich, we can draw pictures together, we can be looking at the same things and pointing things out. These are synchronous interactions, that have their own benefits currently. One of the major benefits of texting may be that that it is easy and asynchronous, on your own time and terms. This might lead us to to ask: how can we make even our asynchronous conversations/interactions more multi-sensory and humane? Is there a way to make rich-interactions feel just as easy as text?
Hope you’ve enjoyed this exploration of Conversation as Interface & Experts In Your Pocket. Looking forward to hearing your thoughts. The next post will be about Programming through Conversation (creating software just by chatting). Stay Tuned.
- In some ways, text messaging is both natural and unnatural, comfortable and uncomfortable. We find ourselves occasionally confused by what someone actually meant when they sent their message, and we may lose additional cues, e.g. body language, tone, etc. that convey huge amounts of additional clarifying information. This article is not about thinking through the pros and cons of text-messaging — if you’re looking for a discussion of that topic, MIT Professor Sherry Turkle has thought extensively about what the implications of texting are on society, what we gain through asynchronous text-based communication, and what we lose in terms of not having as frequent in-person, real-time conversations. I’ve heard her speak, in person, and she’s a profound thinker whose books and talks are worth checking out. Here’s a NY Times piece about some of her recent work: http://www.nytimes.com/2012/04/22/opinion/sunday/the-flight-from-conversation.html?pagewanted=all&_r=0