If you’ve ever tried to call an airline or almost any service business of a certain size, you’re probably familiar with the problem: After navigating a seemingly endless set of options, you’re asked by an automated message to share information, but it fails to understand your response, and no amount of e-nun-ci-a-tion helps. Before long, you’re hollering into a void of inhumanity, asking to speak with a human.

All this can leave us feeling that we’re being held captive in a robot-hostage situation, says Alex Sambvani (MBA 2017), CEO of the startup Slang.ai. There are plenty of factors that explain how we have arrived at this point, he says: Businesses everywhere are short staffed, and call volumes are at record levels, so the human-to-call ratio doesn’t work in the customer’s favor. Plus, with wages as high as they are, staffing the phones has never been more expensive. But this isn’t a fate we have to accept, Sambvani explains.

Better voice technology and AI-based tools already exist, and some large organizations are deploying them in modern, innovative systems that don’t routinely confound their customers. “But most mid-market businesses can’t afford to spend $10 million for a custom voice system, which means there’s a whole range of small- and mid-sized businesses that are underserved,” Sambvani says. With Slang.ai, he’s offering mid-market companies access to these more sophisticated tools that can immediately route urgent calls to a human being and offer analytics about what exactly customers are asking for, and when.

Slang.ai’s offerings are also customizable, with a portfolio of digital voices from which to choose and background sound elements that can be branded. “We’re focusing on not just the transaction but also on the audio experience, which is all backed by research. We know that the quality of the voice is a big driver in how much the end user enjoys the experience,” he says.

The solution started to take shape during Sambvani’s tenure at Spotify, where he worked as a data scientist just after attending HBS. There, he teamed up with coworker Gabriel Duncan for the company’s Hack Week competition, in which employees are invited to spend a week pursuing a passion project. Their idea for a personality-forward voice assistant won an award and, with it, some funding. “Seeing that Spotify was interested in commercializing our idea was a real trigger for us to start this,” Sambvani says. Now, a year after the launch of their product, Sambvani and Duncan are running a seed-stage company with 10 staffers and more than 100 customers, all of whom were found without investing in marketing.

The company’s long-term goal is to build a big business by focusing on the small- and mid-sized market across all industries. The question is how to get there. Slang.ai is currently serving clients in three verticals (restaurants and hospitality, ecommerce, and brick-and-mortar retail and services), but the founders believe that the product has potential for almost any service business that connects with customers over the phone. “That market is too broad for a startup,” Sambvani concedes, so the team spends a lot of time debating which of the following three ways forward makes the most sense.

1. Go all in on one. They could choose the most promising of the three existing verticals and align their resources on one major target.

2. Dig deeper. The team could stay focused on the three existing verticals and delve deeper into each one, investing in vertical-specific features and functionality—like an end-to-end food-ordering system for restaurants. “There are whole startups focused on that, and it’s a very difficult problem to solve,” Sambvani says.

3. Go wider. They could expand into entirely new markets, such as financial services, travel, or health care. This approach would mean identifying the features that are universally appealing, then boosting the company’s sales and marketing capacity and acquisition tactics to suit each industry. The biggest challenge with this approach, Sambvani explains, is having enough resources for the sales and marketing needs of multiple industries.

Sambvani’s instinct is to focus on a few verticals while the sales and marketing organization expands, but there might be a happy medium—going just slightly deeper, for example, instead of taking on a whole new set of problems, end-to-end. Which approach do you think the founders should take: deeper, wider, or narrower?

Market selection is potentially the most critical question a founder can answer during the scaling stage of a startup. As is well understood by VCs (but less so by founders!), small companies rarely do more than one thing well; efforts to hedge across industry targets typically mean a diluted product road map and messy go-to-market motion. At the same time, zeroing in on a single market too early can mean missing the most durable and compelling long-term segment.

My guidance is to follow Sutton’s law: “I rob banks because that’s where the money is.” Pursue the segments with the highest unhappy call volumes and the customer experience and net promoter score (NPS) that’s most contingent on happy outcomes for well-handled, urgent calls. For me, this is choice three, going wider into financial services, travel, and health care. Episodic NPS (such as customers’ experience hanging in the balance during a key moment) has gained traction, and we’ve all had terrible call experiences at the hands of a credit card issuer, airline, or hospital. Pitch those segments on the criticality of properly handling angry and/or time-critical customer cases, and there should be more willingness to pay than in the existing target markets. Great enterprise selling happens at the intersection of true customer pain and a credible solution. Play to win the game!

—Rob Biederman (MBA 2014) is a managing partner of Asymmetric VC, the chairman and cofounder of Catalant Technologies, and an Executive Fellow at HBS.

Slang.ai can be a big business if they get it right.Two years out of HBS, I cofounded the first company to develop the original automated attendant systems for small businesses and sold them through AT&T (the phone company at the time). We sold more than 100,000 of these “Merlin Mail” systems over the next decade, so I guess if you want to blame the system, blame me!

The problem may be more nuanced. In 1984, I met with a senior manager at AT&T who said they had done market research and concluded, “Small business owners will never want their customers to talk with a machine.” Pretty funny now. They changed their tune and soon dominated the market. Turns out, it’s much more about understanding the workflow in a small business, and computers do only what you tell them to do. And, many times, it’s really the business that has broken customer service, not the machine. Small businesses have much more in common than they have that’s different, so “go wider.” If you get it right, Alex, you’ll be very successful.

—Eric Giler (MBA 1982) is the founder and co-CEO of Omnizare Imaging. He is the author of more than 30 patents and previously led several other tech companies, including Brooktrout Inc., a provider of telecom software and hardware platforms, from 1984 until its acquisition in 2005.

Got a case? To take part in a future “Case Study,” send an outline of your company’s challenge to bulletin@hbs.edu

Post a Comment