2024: When AI-SaaS comes of age

2023 was quite an active year in SaaS and enterprise software - after 2022's Eureka moment for generative AI through ChatGPT, we saw a massive surge in startups building 'with' and 'for' AI. While the broader ecosystem continued to see a cooldown, the high pace of activity around AI presented an interesting contrast for us.

In the application layer of SaaS, we see AI as not just an incremental leap in efficiency or productivity, but rather a major platform shift that will fundamentally change the way a lot of SaaS is built and used by businesses. Most of SaaS pre-AI were "products" that would either automate a specific task, or make a certain process you do more efficient through better use of data, abstraction, or design. We already see this changing with SaaS being built post-AI, which are more of "co-pilots" - they are no longer just products you use, rather they actively participate in and move the needle towards achieving the business outcome.

This shift means two things - one, incumbents need to adapt fast and actively incorporate AI assistance into their products or risk being overtaken by emergents. We already see companies like Microsoft, Salesforce, Adobe etc leading the charge on this front. Two, emerging startups in this space need to re-think solutions from the ground up and not just tack on AI to an existing product - the incumbents are already doing that. Their advantage lies in the fact that they can architect a completely novel approach from the ground-up that truly leverages the power of insight and active assistance that AI enables. This possibility is very exciting for us at Blume, and we are quite bullish on multiple industries and functions being disrupted by startups which are building AI-first rather than AI-addon.

We are also actively building conviction towards the possibility that AI will further drive verticalization of SaaS. With custom models trained on industry-specific data rather than horizontal models that can be very noisy and hallucinatory, startups can position very well to provide vertical solutions. The first wave of vertical SaaS (ref. Blume's Vertical SaaS thesis) was driven by industry-specific workflows and use-cases, and we think the second wave of vertical SaaS will be vertical AI-SaaS, driven by industry-specific data, insights and outcomes. These will also be able to leverage API ecosystems, partnerships and network effects much better in their growth journey.

On the development layer front, we are seeing a completely new tech stack being built for an emerging role - AI engineers. Applications that will deploy AI in production will need a toolset that will enable them to effectively evaluate, fine-tune, test, and monitor these models at scale. We are bullish on a new generation of developer tools being built for these use-cases that will complement existing tools rather than replace them. It would be interesting to watch some of these tools take an open-source v/s a closed-source approach and how those decisions impact evangelisation, adoption and monetisation among the developer community.

Lastly, we have seen early conversations around privacy and regulation of AI models and applications in 2023 - we fully expect these to materialise into more concrete expectations in 2024 as enterprises and governments come to terms with understanding how AI leverages data (private or public) and how these applications will be used in the real world. Security, privacy and compliance will be important areas to watch - we expect privacy concerns to drive data security needs, opening up opportunities for startups to build for these challenges in the context of AI.

One important element to underscore across this outlook is that we at Blume love companies that are so much more focused on the problem/solution pair than the underlying tech being used to achieve the outcome. AI is absolutely a much better crutch to get to an outcome, but what matters for a customer in the end is whether the outcome is being achieved. Startups would do well to keep this in mind while building, positioning and selling their new AI-SaaS in the market.