As AI labs such Anthropic unleash application-specific models, whispers in the AI community suggest a grave existential threat to purveyors of traditional SaaS software.
These worries have grown as Anthropic has introduced plugins for various industries, including legal, design, and finance, and a cybersecurity model, Mythos.
However, not all software vendors see the progression of generative AI as a cause for panic. For Mike Ettling, chair and interim CEO of SaaS insurance software vendor Sapiens, there is no need to fear that this is the end of SaaS; instead, it is the vendors without the right moat that are most imperiled.
In this Q&A, Ettling discusses what it means to have the right moat for SaaS vendors and what they should focus on now.
As someone watching the tech market and as a leader of a SaaS vendor, how much are you concerned about the encroachment of generative AI?
Mike Ettling: I am not that paranoid about the SaaS apocalypse. We are living in a mixed-up world now, where public investors are running an investment strategy that is basically: what is the next Anthropic press release?
Everything is going to come down to who has the best moats. Who can move the fastest to protect their moat? And with a moat, you have years and decades of ontology about the customers you serve, and you can use that ontology to build AI tools that really transform a business.
So, in the insurance industry, many big insurance tech providers are doing a lot for companies. However, these companies still have people doing underwriting, handling claims and doing manual work around our systems. Yes, there are a lot of startups coming in and saying, ‘Oh, we can do that.’ But then you look at the projects, they do not own the ontology. So, they cannot really change how those buildings of people operate and are needed, because they do not really own the underlying moats.
My view of this apocalypse is that, as we start to see it a bit in certain deals, once a company truly establishes its moat and can demonstrate that it can lead to AI revenue, valuations are coming back. On the other hand, there will be an element of SaaS apocalypse. There would be a lot of catastrophes and failures as this all shakes out.
But my view is, we have been through all this before, the internet, the steam train, electricity, and we will figure it out.
You are saying that unless you have data, you are in trouble as a company. What about those software vendors that are just starting out? Can they survive?
Ettling: It is about owning the context as the software vendor, which is the moat. For example, a software company that owns core HR data, the core record for people data, owns the context [for that data]. The software company that is just doing all design, all representation, goals and success management, they own peripheral data. They do not own the call data. They are going to struggle. So, they lack much context. They therefore do not have a big moat, and they are going to struggle. to survive.
What if a software vendor partners with one of the big AI vendors? Could that help them survive? Especially because Anthropic has made deals with vendors across different industries, such as legal.
Ettling: Yes, there may be software vendors that have been doing things around case management at a legal company or in data research. LLMs are going to obliterate those companies. Those software vendors are not needed.
LLMs can do this through contextual analysis of data, and creating insights is what they are good at.
What LLMs will disrupt is the whole legal pyramid recruiting model, where you recruit a hundred clerks, and eventually ten become partners.
You do not need a hundred clerks anymore. You might need twelve and Anthropic. It is going to disrupt that whole legal, professional pyramid model that many of these professions run on.
But providers are offering ERP, core accounting, time management and billing for legal firms. Those are not going to be disrupted because LLMs are not going to take over doing that.
What about the argument that many make, which is that while LLMs and AI agents do not take that over right now, they will eventually?
Ettling: If anything, those ERP vendors should now be aggressively building agents to do what those competitors in the case management peripheral side of the legal space did. They should be building those agents and offering them to their customers. You cannot just use Claude. You cannot just use ChatGPT. You must build the context, you must build the rules, which have got to come from somewhere. And that comes from the people with the core moat and the core data, processing the core transactional data.
If they move fast enough, they will cement their moats, build more AI on top of them, and we will be safe, whereas the ones who are in have been on the periphery and will never be able to get them out. Their modes will be eroded, and they are, the companies will take that over.
What is the main challenge that software vendors will have to face in this phase of AI?
Ettling: The biggest thing vendors are going to have to focus on is speed of execution. This world is now moving at a pace I have just never seen before. We do not know what will be in the October press release from Anthropic. Their pace will be the differentiator between people who succeed and those who do not. That is an investment, because many organizations do not have the ability. They have lost their pace.
Can highly regulated industries like insurance afford to keep up?
Ettling: Contrary to some of the arguments you see out there, AI helps highly regulated industries. The problem in regulated industries is ensuring that all compliance is in place and that the system does the right thing. Sometimes, regulated industries have additional testing requirements.
Payroll systems are a good example. The reason there are not many global payroll platforms in the market is the cost of maintaining them and of keeping up with all legal changes and requirements, country by country, to keep the system relevant. That was always the barrier to entry. With AI, that is now totally changed. We are going to see a lot of payroll startups in global payroll because of the advent of AI.
AI will help regulated industries move faster.
What would your advice be to organizations and software vendors that are panicking because of the pace at which AI is moving?
Ettling: What is exciting for me is the fact that there is not really a playbook. So, what I am trying to do in all my companies is get the best brains in agentic AI, to brainstorm the topics, because we are making out the playbook as we go.
My advice to companies is, firstly, to stay calm. Secondly, get the best talent. Do not be constrained by your organization’s hierarchy.
Any software company must build with agentic AI now. You cannot build software the old way anymore.
Editor’s note: This interview has been edited for clarity and conciseness
