Gil Dibner recently wrote a great post about „Systems of Intelligence“, continuing the 2017 thoughts of Jerry Chen, a Greylock partner. I really like Gil’s ideas regarding barriers to entry as they apply to Systems of Intelligence. I highly recommend reading both posts as they are very relevant to anyone in the tech field.
Image borrowed from Jerry Chen’s post
To summarise: verses previous ways of creating value in software, you now need new and different barriers to entry as a player in Systems of Intelligence — technology alone or network effects won‘t suffice.
Gil’s post is aimed at venture investors but I felt it applies to receeve as we are running a business that is a system of intelligence. His discussion around new barriers to entry resonated with me, as it reflects how we strive to position ourselves to remain competitive. I felt it would be interesting to go through his list of barriers to entry and relate them to what we are doing. This may help illustrate how we think about these things.
This is an easy one in regards to receeve. We have been focusing on the collections industry for years and previously built a successful startup in this segment. We‘ve also had the ability to see the inner workings of multiple incumbents and their interactions with customers. To drive receeve we hired experts from the field into our team. It is safe to say that we know the domain and have built up extensive expertise in the vertical. Having been an investor in the past I know that sometimes it is better to not know what you are doing in a segment and perceive things with fresh eyes that allow you to see disruptive solutions. In collections, you need to understand the processes and work within regulatory constraints. Without domain expertise, you will struggle to create any barriers to entry or to differentiate yourself from competitors.
receeve derives advantage from the inherent complexity our customers operate in. Large enterprises are not easy to navigate. For collections teams at large banks, there are extensive workflows and a ton of data. Edge cases are where receeve’s software can shine. We can see patterns in the data not currently visible via manual workflows.
Additionally, we can inject behavioural analysis into these workflows and combined with the data, significantly increase recovery rates. We can leverage channels that weren’t evident in the past and enrich the customer experience. There are numerous layers in working a claim, and adding AI on top garners a certain level of lock-in once up and running. We keep competitors at bay by aggregating our ever increasing understanding of the complexity and continually productising it into our software — growing automation of this complexity keeps others from entering the segment.
We are selling initially into existing operations and teams. Our technology is digitising workflows that are extensively manual at the moment. Yet, we cannot completely rule out human interaction. Our system with AI implemented, will automate as much as possible but you will still need human input. Sometimes with specific collections strategies a machine cannot beat a person in determining the proper steps and in certain cases, person to person interaction is essential to achieving the best outcome. We have no choice but to remain hybrid and see this as an opportunity for optimal performance. Regardless of how much you automate and optimise, there will simply be situations where a person talking to another person adds a layer of trust or comfort. This benefits both parties in coming to a resolution and drives customer retention.
This is where our strategy is headed. Here Gil writes that you can keep competitors away by having a network effect across customers of a system of intelligence. As mentioned above when it comes to complexity, our enterprise customers are all difficult to navigate. Even when the industries differ from one another, you can still get network effects sharing intelligence across customers. We first create lock-in at our customers by automating their processes and thereafter via the data we collect. In a best case scenario, the network effects from our portfolio kick in. The AI and what it learns from our customer network keeps competitors away. Switching means that our customers would have to start the machine learning from scratch with any new implementation. Conversely, as we grow, the direct benefit to our individual customers grows.
We may in the future add features which today probably don‘t even exist. As Gil states, this is a tough one, because so much tech is becoming commoditised. It won‘t be the AI that differentiates receeve nor the SaaS model. We differentiate via the systems of intelligence we put in place. One of our challenges will remain to create a barrier of entry with hard tech. Being a startup allows us to quickly try out new things. We are continually assessing technology to see where we can integrate it to increase the value our solution brings to customers. Thankfully we see multiple avenues of technology evolving in parallel to our solution. These new technologies could in the future easily be used to further differentiate our offering and benefit our customers. In this case, we are happy to hold back reporting on certain capabilities we are integrating to keep ahead of the competition.
I agree with Gil that data is probably not going to create the moat one wishes for. It will be everywhere, it will be had by all. Customer data may gain increased privacy protection but customers will be incentivised to share it. The sharing of data will eventually benefit everyone in the industry but will also decrease the opportunity to use it as a barrier to entry. It remains to be seen how things evolve here. Nevertheless, change happens slowly sometimes. I believe we still have time ahead of us where data remains relevant as a barrier to entry. I personally thought as a society we’d be farther along and data would be shared far more freely than it currently is. Privacy issues have held back evolution here. One needs only to look towards China and the West’s reaction to their use of data to wonder how open we will be to sharing everything.
Ultimately we are trying to create „the System“ to which Gil alludes. We realised that “just another cloud solution” is not an approach that will cut it anymore. It was reassuring to read Gil’s take on this and I saw so much of what he wrote in the approach we are taking. I also realised that a lot of investors haven’t yet made the shift in thinking…..away from SaaS and cloud to Systems of Intelligence. Discussions still hinge upon data moats, technological superiority and the AI itself when it comes to our solution. We steer the discussion to domain expertise, networks of intelligence and a hybrid approach to show strong differentiators in our industry. It seems that VC’s may finally be coming around to this reality.
Design can make or break a product. The rise of the iPhone arguably had less to do with its features themselves and more to do with the design. Competitors could boast similar functionality—but the ...