Log in     Support     Status

The Support Automation Show: Episode 37

by | Oct 20, 2022

In this episode of The Support Automation Show, Justin Clegg, Founder and CEO of Allset discusses the benefits of SMS-based AI and how small businesses and direct-to-customer settings can utilize it efficiently.

Listen now!  

Justin Schmidt: Justin Clegg, welcome to the Support Automation Show.

Justin Clegg: Thanks for having me. Good to be here.

Schmidt: Where’s this podcast you find?

Clegg: We’re a Utah company and I am currently traveling to Chicago for a trade show event this week.

Schmidt: Love Chicago as St. Louis, and that’s our weekend getaway big city. I’ve spent many, many a day in my life in Chicago, Illinois. Great place.

Clegg: First time in the city. Excited to learn more and send some recommendations.

Schmidt: Yes, I will do it. Justin, you’re the founder and CEO of Allset. Why don’t you tell us a little bit about yourself, your journey to Allset, and what it is that Allset is out to do?

Clegg: Happy to. This is my fourth startup. I’m a repeat founder. I started my career in the bay area. I moved to Silicon Valley in 2013, where I worked for large tech companies leading in product and marketing capacities. From that point, I decided to make the jump into the wonderful world of startups and ventures. Joined a couple of companies that were backed by well-known accelerator programs like Y-Combinator, 500 Startups, and a few other programs. Learned by doing.

I’ve spent time, I put in the grind and the hours to understand the economics of startups and moved into a number of different roles and capacities from head of growth to head of product, and really dug my heels into conversational AI, which I think is what made it exciting for us to want to get together and chat today.

Schmidt: Yes, absolutely.

Clegg: We have been spending really the last decade researching and understanding natural language processing, machine learning, entity extraction, and understanding ways to make conversational chatbot and commerce experiences feel more like a human in the room.

Schmidt: There is definitely not a lot of pushback I’m going to give you on the benefits of conversational AI given our shared interest in the discipline and shared belief in the value that it drives for both the enterprise and you and I in our personal lives and just users in general. Allset itself is in this space. What is it that you guys are trying to do with Allset?

Clegg: We built a conversational texting platform that’s specifically focused on home service businesses. When we talk about home services, we define that in a few different segments, cleaning companies, so think window washers, pressure washing, even auto detailing your carpets, and then we talk to contractors, so HVAC, plumbing, roofers, painters, and then outdoor companies, so think landscaping, lawn care, pest control, and tree services.

It’s a really massive market. There’s about five million businesses in North America, and the opportunity for Allset is that you have this industry that’s completely overlooked and underserved from a Silicon Valley perspective, from a perspective of emerging tech and conversational commerce. We’re building tools to automate communications and payments to make the overall customer experience step less booked for homeowners, who are in constant need of these services post-pandemic, and for business owners who are operating, running, and trying to grow their business.

Schmidt: Awesome. One of the great things about the proliferation of conversational AI and chat and support and service automation is that we have products available now across the industry spectrum, across the size of the business landscape, everything from mom and pops all the way up to Fortune 100 companies have tools at their disposal, whether it’s products they buy on the open market or products they develop internally through available APIs and a lot of the DevOps tools available to bring these products to the masses. It’s always a pleasure for me to talk with somebody else who has a shared vision on this. My first question for you, Justin, is the same one I ask everybody at the start of this show. When I say the word support automation, what does that mean to you?

Clegg: It’s a good question. We think about supporting automation in a few different capacities and contexts. They’re situational, and it really comes down to how you handle things like customer service, how you handle a contact center, even from a sales perspective, how we think about qualifying and handing over a live agent once it’s been identified as a lead. Then we also think a lot about content, so content distribution.

I think support automation should be able to address all of those capacities to an extent, and in our earlier conversation, making it feel like there’s a human in the room without breaking down the experience. That’s probably how we would structure and how I would think about the world of support automation today.

Schmidt: I share a lot of that sentiment, and one of the areas where you guys have spent a lot of investment in the product, at least from my outsider looking in perspective, it’s correct to assume Allset’s, the primary modality of the engagement that your customers have with their customers is through SMS, correct?

Clegg: That’s right, yes. SMS is so ubiquitous, and whether you’re 16 and you’re getting your first phone or you’re 60, really in the US, at least, it’s this application that everybody is familiar with immediately. It’s UI-less to an extent. The biggest question we’ve always been asking over the last decade is, why is it so ubiquitous compared to some of these other ecosystems and platforms that are growing exponentially outside of the United States i.e. WhatsApp, WeChat, or some of these other platforms. Primarily SMS space and that’s mostly because of the high open rates and high read rates.

Schmidt: There’s no user training that you need to do to teach somebody how to text on their phone. There’s no new app to install, there’s no new UX paradigm to use, so you really do have this, to use a computer science term, you’re writing to the metal, so to speak, of the way people interact with each other, which is extremely cool.

In your customers, and this is something we deal with too, I’m just a little inside baseball with a couple of conversational AI nerds chatting here. The NLP entity-matching, when to pull a human in the loop, all that stuff aside, one of the really interesting things to look at when you’re onboarding users into a conversational AI-driven environment in support automation is just some principles of conversational design.

I look at some of the larger chat players in the space, Intercom, Drift, Talkdesk, for example, all these things, LivePerson. One of the really interesting things that I see a lot is where and when customers want to infuse their own brand and personality into the conversational design of the chat they use, and where they seem to lean most into out-of-the-box formatting and templates or even maybe biasing less towards brand and personality and more towards what can I do to maximize the conversion, whatever the conversion goal of the particular chat that we’ve deployed is.

How does Allset work with the service businesses that you guys use on this question of conversational design? Do you mostly out-of-the-box or do you guide through semi-personalization based on the business and the industry that they’re in?

Clegg: It’s a great question. The experience today is relatively simple. When a service appointment is completed in someone’s home, the business is going to mark that as completed in their field service CRM. When that moment takes place, that’s when Allset takes over from a conversational AI perspective. An event is triggered, we’re going to post and essentially send a text to the homeowner letting them know that the service was completed. We’re going to prompt them to do a number of different interactions post-service. Those interactions can range from leaving a tip, to leaving a review, to creating a referral, and referring a family or friend, even to completing a payment.

You have all of these different contexts. What we found, I think the biggest discovery was that we are very much a guided experience where we prompt and we’re big believers and focus on a very specific intent. If the context and the intent is too broad, end-users will get lost, and then they end up just ignoring it or moving on to the next experience. Also, it’s very much a user flow that starts from capturing a payment, to then leaving a tip, to then leaving a positive review because you’ve already opened up the customer’s wallet. From there, then prompting referrals and going from there.

We think about conversational design a lot. To your question and point on branding, we have to be very intentional about things like the length, the time of day, the cadence of each message, the personality, the branding. Do these messages say powered by Allset? At what point does the consumer feel safe to leave a payment through a random text? Those are things that we have to think about when we do get to move the user to a UI from the initial SMS experience.

Schmidt: Your point about intent is something that I think all support leaders that listen to this show, or if this part of the conversation gets clipped for sharing on social media, this is something that I think is really important, is that focusing on the intent of what you’re trying to do. A lot of the times when there’s frustration from the customer with automation, the paradigmatic example of this is sitting on the phone and saying, “Speak to an agent, speak to an agent. I want to talk to an” The IVR recursive loop of death when you just want to talk to somebody.

A lot of times, that can be mitigated by giving very clear instruction and the sort of voice menu, or in our case, in our being as a conversational AI, understanding, if you’re going to guide, to make sure that you build that conversation with a very specific intent to avoid the frustration, gain trust, do all these other things. One of the things that comes up on the show, and I’m going to see if I can land the plane gracefully into my next question to you is, there is sometimes pushback inside businesses to bring automation into the business.

Whether it’s in a large call center, I don’t want to install some tool that’s going to make all my agents think I’m trying to replace them. Or I don’t want to create customer confusion and frustration with the recursive loop of IVR hell, or whatever it is. I’m curious to hear your take on this because you deal a lot with service-oriented businesses, I assume a lot of small businesses too just given the nature of that space, where the customer relationship in a service business is literally everything. All businesses say their customers are the most important thing, and that’s true.

It’s really true in a service business where if the relationship doesn’t go well the customer can go to Google reviews, or Facebook, or whatever it is, and you can create a death spiral for yourself from bad interactions. I’m curious to hear from your point of view, what are some of the challenges that you, or maybe not you specifically, but just some of the trepidation of bringing automation into the service space and what leaders and business owners there can do to mitigate some of those challenges?

Clegg: I love the question. Number one, we have to talk about engendering trust into the conversation, and how quickly can you build trust with a phone number? Another way to think about this is, what is the relationship that the retail customer, in this case, a homeowner, should have with a phone number? Who is the phone number from? What is the associated area code? Is it a long code with your standard 10 digits? Is it a shortcode? We’ve found a lot of research, a shortcode, phone numbers are actually perceived to be spammy, so you get higher risk or higher concern that there is that spam.

Number one is, how do you build trust between the phone number that’s being powered by Allset, owned by the business, and then viewed by the consumer? We have to be very upfront and intentional about setting expectations in that context, “Hey, you’re going to be receiving a phone number, or a text message from this phone number. This is not spam, you can opt-out anytime. This is something that we endorse as a business.”

It’s helping professionalize the business’s representation of themselves through this number. A challenge that we run into quite a bit is that there’s a lot of software that can provision a phone number from a Twilio, Plivo, and Nexmo. How do you think about handling and routing a variety of phone numbers and even things like taking an existing phone number that a business owns and making it textable? Which are things that we think about quite a bit. That’s the first step, is how do you establish trust with a phone number, and get the consumer feeling comfortable?

The second is then, how do you think about derailing the conversation? We talk a lot about user stories in developing their product. We have to create almost conversation stories where you have scenario mapping that’s beyond just the traditional decision tree. The decision tree is the reason why most people hate chatbots and why they became this shiny toy in 2016, and then crashed and burned because people realized, “Oh, it’s a feature, not a business.” Decision trees are pretty simple.

We ingest a lot of data, we take a lot of unstructured data. Our biggest surprise is that people have thoughts and they want to share them, and they text those phone numbers back like their friend or family member. What that means for us is we have a lot of unstructured data, a lot of feedback, and that could be derailed in a number of use cases, “Hey, Justin, I want to reschedule.” “Hey, your team did an amazing job, I want to compliment you guys.” “Hey, I have a complaint, you missed a spot. Come back. I hate you guys.”

Or it’s things like random questions or random thoughts. Now you’ve got, does it fall into the complaint or compliment category? As we think about those different contexts, the number one goal for us is to prevent conversations from being derailed, and providing useful and insightful responses so that we can add value to both the end-user without upsetting them, and also let the business know that these insights are coming into the conversation.

Schmidt: The ability to own as many communication channels with your customer as possible is universally a good thing. It’s good to be able to reach your customer on their terms and be able to parse their language to give them what they’re looking for in the easiest and simplest way possible, and to avoid the multi-channel handoff hell that can happen if this isn’t all well done. I like the theme that you’ve brought up several times in this conversation already of intention, and how leaders need to really build their support automation initiatives in whatever form that takes with intention.

Whether that’s conversational design or the workflow automation that may go on in your ticketing system on how to resolve tickets or whatever it is. One thing that I’m curious about is your take on given the limitations of SMS and the potential future of something like RCS being widely adopted. Do you see in a world where, let’s say RCS is the new– SMS is gone, RCS– Apple acquiesce and everyone gets along and RCS at least becomes an option on every single phone.

Do you still see most conversation design being text-based and trying to keep it in as much of a box as possible, or do you see more vendors and technology providers branch out and start using, the more– I don’t want to say, infinite canvas, but a much wider canvas of the types of formatting and buttons, and replies and all the other stuff that you can do with something like iMessage, or RCS?

Clegg: There’s no doubt that RCS having access to rich media within an interface is going to provide more value in terms of selection, carouselling, imagery, photography, we do it a lot with MMS today. MMS, we can build Netflix-like experiences within our product, where we have thumbnails or faces of the contractors or the technician. You’ve got logos, and people are drawn to people. When you can display those things, there’s always an advantage in terms of building that trust and that experience. I think, one, having an omnichannel strategy, I completely agree with your approach that consumers want to be able to communicate with a business in a variety of channels.

It really just becomes a case-by-case personal preference. We have customers who are in between 35 to 55 demographic. They just need to get on the phone and know that everything’s working okay. They don’t need access to a chatbot. They don’t need some ticketing support system. We have folks on the other end too who are very self-service. They don’t want to make a phone call. They don’t want to chat with anybody. Obviously, our number is– our support line is textable. They’ll send a text in. They’ll ask a few questions.

We rely heavily on MMS, on screenshots, and custom link sharing and building so that we can get them to the support tools and experience that they need on the spot. Then they can continue on with their day. It’s really people who have different jobs that they want to get done when reaching out to a business. The goal is to be proactive and to number one, deflect as many easy questions or softballs as we might call them. Then the other aspect is the containment and just making sure that we can give them access to information in a seamless way.

Schmidt: Yes, the impetus to deflect and the desire to contain is absolutely a North star for automated support using conversational AI, or even if it’s not conversational-driven contextual menus and buttons to click and actions to take to resolve the issue as fast as the laws of physics allow. This has been a great conversation, Justin. I can talk to you all day about this. I want to end with the same question I ask everyone at the end of the episode and that is, when you look at the future, say the next 10 years of automation and the support world and the conversation we’ve had today, what is it that excites you the most?

Clegg: I would say– a few thoughts. One, support becomes anticipatory. You can start to build up predictive modeling to answer support questions before even the end-user or the consumer of the product knows that they need support. I think that’s interesting. We see powerful tools like Capacity and others that are out there, that can present intelligent knowledge bases that, again, they know exactly what the customer is going to need based on where they’re at in that customer journey. Then the second thing that personally I get excited about is just seeing more ecosystems that are built around support.

For example, when we look at WeChat, in China, a consumer can go into WeChat, and they can have a variety of experiences. They can order a cab, they can order a meal, send money, pay friends, do social media, and communicate and chat, which are those core values of human connection. When I think about the world of support, you’ve got this merge between virtual assistants and just overall efficiency. It’s there a one-off catchall platform that handles all support needs across all platforms.

That, obviously, I think, becomes the holy grail. It’s not just specific to our particular companies or products, but it’s said, it’s a one-stop-shop to be able to get support across billing, across new technology at scale. I think there’s a world where that ecosystem becomes available and true and consolidates all of my support needs across all of my tools and things that I’m invested in. I’m interested in that world. Sign me up.

Schmidt: Yes, the concept of the super-app. It’s interesting that we don’t have one of those necessarily in the United States, but like WeChat’s the ultimate example of it. Justin, to end our show, I’m going to do our world-famous quickfire round, much like the roadside diner has world-famous fried chicken, or whatever. One of these days, I’m going to come up with a brand for this. I joke sometimes that the brand for the quickfire round is me telling the guests that one of these days I’m going to come up for a brand, so maybe I’m just going to call it the unbranded quickfire round. What’s the book you most often recommend to people?

Clegg: Professionally, Atomic Habits. Personally, The Alchemist. Both great books to really focus on, aligning yourself, building intentional focus habits, and finding your personal treasure. Those are, I think, important things in doing business, and in living your life.

Schmidt: I couldn’t agree more. In terms of the various hacks, automations, workflows, et cetera, that you’ve set up for yourself to be as productive as possible and to be able to manage the startup and keep your sanity, what’s the productivity practice that you’ve brought into your life that has stuck with you and the one you recommend the most to people?

Clegg: Be consistent. Be the hardest-working person in the room. I think, if you can combine those two, you can become dangerous.

Schmidt: Love it. Is there a website, blogs, Slack community, LinkedIn group, et cetera, for support leaders or business owners looking to increase their ability to deliver world-class support in a semi-automated or scalable way? What would you recommend?

Clegg: I am not a part of the ton, and I’d love to be a part of more. We’re constantly on the lookout, but the first place I would go is likely a Facebook community or LinkedIn group to meet with some of those like-minded professionals.

Schmidt: Yes, there’s some good ones out there. I’ll shoot you a note when we get done recording here of–

Clegg: Thanks to you. Thank you.

Schmidt: -Yes, all the ones that guests had recommended. There’s definitely a lot. If there is one person in the business world that you could take out for lunch, or coffee, or a cocktail just to pick their brain, who would it be?

Clegg: I wouldn’t be able to because he recently passed, but Clayton Christensen has been, I think, a really powerful example in terms of somebody who understands strong business theory and principles, articulates it in a way that is easy to understand, but also lives his life and conducts his life in a way that is something that I’d like to aspire to be more like, so definitely–

Schmidt: He’s the one who wrote The Innovator’s Dilemma, right?

Clegg: That’s right.

Schmidt: Yes.

Clegg: Disruptive Innovation and Innovator’s Dilemma. Yes, a very underrated business leader in the community over the last 10, 20 years.

Schmidt: Excellent. Justin, if people wanted to find out more information about yourself or Allset, where could they go to find you?

Clegg: Start with tryallset.com. That’s our website. Allset is a Utah-based company. We’re growing very quickly. That’s where I would point people to learn more about our company.

Schmidt: Awesome. Justin, thank you so much for joining me on The Support Automation Show, and you have a wonderful day.

Clegg: Thanks, Justin. Bye-bye.