Why We're Focusing On AI-Powered Customer Success For SaaS At Locusive
Over the past year, I’ve been heavily focused on not making the same fatal startup mistake I’ve already made at least once in my life — building something that people don’t want.
I’ve always loved the idea of building a company from scratch. There’s a sense of fulfillment and satisfaction I get from trying to create something from the ground up that I don’t get in doing anything else, and it’s nice to imagine what might be if the whole thing grows like crazy.
But building a business has to start with solving a problem.
I didn’t always think that, and, especially as someone with a technical background, I believed that if you can just build something cool and figure out how to get it in front of people, you’ll grow.
Maybe that was true at one point, but I don’t think it is now.
These days, if you’re not doing something that people find truly valuable (and oftentimes, this means creating something that can save them time or money, or grow their income), then it’s going to be really tough to create a business that grows.
That’s why, this year I’ve spent a lot of time trying to find a problem that’s big enough, painful enough, and interesting enough (at least to me) that it’s worth me dedicating the rest of my life to solving (because really, when you’re building a business, you’ve got to take a long-term approach).
Fortunately, after around 10 months of work, I feel like I’ve gotten to a point where I understand a problem space, an industry, and a potential solution well enough that I can dive in head first to solve it.
Maximizing software value
The SaaS industry is huge. It’s valued somewhere between $1.7T and $3T USD, and projected to grow to $10T by 2030. Contrary to recent opinions, I don’t think the software industry is going anywhere. In fact, I think the recent boom in AI technology will lead to even more software in the world.
While it has become significantly easy to build software, building good software that adds value to users’ lives in a seamless manner is hard.
Over the past 20+ years of creating software, I’ve seen this myself. Folks will push out a new product, some customers will start using it, only a few will know how to really use it well enough to become power users, most will disappear completely, and nobody will use its full potential. And we all know what happens when users don’t find value from a product — customers will churn.
Companies that build software spend a lot of time and energy figuring out how to get their users to get the most out of their product, and many software companies even maintain teams of customer success and customer support personnel to hold the hands of their users through the process of them using their software.
One could argue that when you build really good, intuitive products (think Apple), users will just get it, eliminating the need for additional resources designed to help customers use the product. While that may be true, it’s significantly easier said than done to build software that’s easy enough for users to understand without them having to reach out for help, and this is doubly true for enterprise software.
So it makes sense for companies to bring on employees that are responsible for ensuring their customers find the most value from their products, but what I’ve found is that those employees are frequently spending time on low-value, tedious tasks that need to be done, but don’t add value to their employers.
This means that the very group of people who are supposed to be responsible for driving renewals and expansions are instead spinning their wheels with low-value tasks like trying to compile data or jumping on a 45-minute call to explain why a button is disabled on a dashboard.
CSM teams are overwhelmed even when SaaS customers aren’t feeling served
Over the past few months, my team has made a concerted effort to understand what’s happening at customer service and support teams at both small and large software providers.
When we first started exploring the idea of using AI as an “app assistant”, we knew that there may be an opportunity to help software developers create a new type of interface for their users to better use their products, but we weren’t sure if there was a real need or problem to be solved in doing so. But after approximately 60 interviews, we quickly found a large set of issues that caused customers to feel like they weren’t getting value while simultaneously leaving CSM teams feeling overwhelmed.
CSM teams are small and tasked with low-value but unavoidable work
At a large enterprise, the customer success team might generally be composed of two or three different divisions. In the first division, each CSM rep might be responsible for a handful of accounts, each of which represents one of the company’s largest customers. These reps are able to frequently communicate with the folks in their account, and are generally able to build a good rapport with them.
The second division might be composed of a pool of CSMs that handle all the other, smaller accounts, of which there might be several hundred or thousand. There’s no way for the company to maintain high-quality relationships with this many accounts, so the jobs of these reps is more transactional — handle new requests as they come in and then move on to the next one.
Some other companies might have an intermediary division between the first and second, where each rep might be responsible for a large (but manageable) number of mid-tier accounts.
Other companies, frequently startups and SMBs, might also roll in customer support/service into customer success and have one team that handles all functions.
But in nearly every interview we conducted, CEOs and Heads of Success told us that there was definitely a problem with their CSM team getting too in the weeds with low value tasks that didn’t move the needle.
It’s easy to see how this might happen, particularly in the CSM division that’s tasked with handling the myriad of smaller accounts. With so many accounts, it’s highly likely that CSMs will need to handle a large number of requests, and oftentimes these requests aren’t of the type that really impresses the customer. So when a rep completes a customer request, the impact on the customer is minimal, even if completing that request might have taken the rep 45 minutes or they had to pull in one of their engineers to help.
Even with reps that have fewer, higher-quality accounts, the number of low-value requests can add up. One of the most common requests we’ve seen is that users just don’t know where to find a particular piece of information in a complicated app, so their first move is to email (and oftentimes, call or text) their rep for help. Something like this might be a small interruption for that rep, but if it happens several times a day, that rep won’t be able to focus on more strategic work.
Worse still, a lot of these customers will ask for larger, more custom work from these reps. It’s not uncommon for them to ask for a custom data file of stats or metrics specific to their account on the software platform, or for them to ask questions that might require the rep to chase down lots of unknowns over the period of a couple of days.
These types of requests have to be addressed in order to maintain a good relationship with the customer (and maximize the chance that they’ll renew once their subscription comes due), but completing these requests likely don’t add any additional value to the lives of the customer. For example, many times, the person requesting something that requires extra effort isn’t even the one that’s using the results, they might be reaching out because their boss or board or customer wanted to know something that they were decided to simply offload to the CSM.
All this goes to say that CSM teams are spending a lot of time on tasks that, even when completed, won’t increase the probability that a customer will renew or expand. And despite the efforts of these overwhelmed CSMs, most customers still don’t feel like they’re getting what they need.
Users want to achieve their main objective as easily (and lazily) as possible
When you’re just one of a thousand customers for a large software company, chances are, you’re not going to get the white glove support you wish you could from your software tools.
If you need something taken care of, you’ll likely have to try to figure it out yourself using documentation in a help center. You might be able to use an automated, rules-based chatbot that, 99% of the time, won’t be able to solve your problem. If you’re willing to wait an hour or more, maybe you can call in and get a real person to help you out (this frequently feels like the case for me when I need to dial in to customer support for Quickbooks!).
If this happens to a crucial software tool you need to run your business, you might be willing to put up with the hassle of not getting what you need for some time. But if you’re trying out a new product, or if you’ve got alternatives, it’s highly likely you’ll simply go sign up for one of those alternatives and use them instead, causing you to never even think about the offending software application (that is, of course, until they send you some irrelevant marketing email).
Forget about taking advantage of the advanced features that the software tool might have helped you save hours of your time, you’re not even able to complete the most basic functions of the product, leaving you frustrated and annoyed.
Your experience as a user of that software application has been forever tainted. You were just trying to get something done, but ultimately found your time wasted and you may never have even gotten a resolution.
These problems are widespread and becoming more painful
Businesses have always had to ensure their customers are happy; staffing customer support/success teams and having them handle inbound requests is nothing new.
But I believe that as more software tools come around, each of which has an increasing number of features, companies are going to find it harder and harder to scale the level of personalized support they can give to their customers. I bet you most CSMs today are simply reacting to inbound customer requests, rather than proactively going out and finding new ways to bring their customers new value.
My customer discovery calls have already illuminated the extent to which this is a problem across industry, and while it may not be an existential problem for most companies today, it is quickly becoming a major time and cost concern, which means it will become a hair-on-fire problem in the near future (if it isn’t already, today) — just the type of problem I was looking for as the basis for a new startup.
The stars are aligning to attack this problem
We’re now in a Goldilocks moment where the world is using a ton of software, that software isn’t always straightforward to use (particularly if it’s enterprise-focused), companies making this software need to ensure they provide excellent customer experiences so that they can drive renewals, and AI is now in a position where it can automate much of the tedious tasks that have traditionally required some non-zero amount of human intelligence.
On top of all this, much of my career has been spent working with data — building data systems, creating machine learning applications, implementing data pipelines, forecasting and predicting the future with data — and the AI-enabled assistants that I just know in my bones will power our lives in the future will all need to be connected to external data systems and tools in order to function properly.
I love the idea of being able to create semi, potentially even fully, autonomous systems that are capable of accomplishing complicated tasks. And I’ve seen this problem firsthand myself — it’s really hard to ensure that new software users get the most out of the software I’ve created in the past.
On top of that, it’s easy to imagine how an AI-enabled assistant that handles user requests might provide new insights to any software company that leverages it for support or success.
The right opportunity seems to be presenting itself using a big neon sign right in front of my face, and I intend to pursue it.
Customer development partners are helping to lay the groundwork
I’m now working with a handful of early customer development partners to build out a solution to these issues. They’re companies that have experienced many of the problems I described above and are interested in pioneering new automated solutions to provide high-quality customer experiences to their valuable customers.
I’m working with them to understand the most common types of requests they get, integrate the data sources and systems they use into a functional minimum viable product, create an optimized UX, and identify areas of optimization for the new, AI-enabled assistant I’m creating at Locusive (which I’ve named the Locusive Concierge).
In exchange, they’re getting a product that’s designed to solve their most pressing customer success needs — one that comes at a large discount to what we’ll be charging when product is generally available.
While I know I’ll be facing some formidable technical challenges for the Concierge, I know this will be possible, since I’ve already spent a lot of time over the past year building out our chatbot for Slack, which incorporates a lot of the same autonomous agent, data integration, and Generative AI technologies that I intend to use for the Concierge.
I’ve also got some really world-class folks interested in working with me to build this thing out and go to market, so I’m geared up and ready to go.
If you’re interested in becoming a customer development partner, I’d love to chat more. Feel free to reach out to me at shanif@locusive.com or book time on a calendar I’ve created specifically for our early adopters.
