Hi, I’m Clement. The Big Picture is a newsletter for founders and investors who want to make sense of what is happening in the SaaS industry.
The Big Picture
Data points and market trends
Revenue and Product Engagement Figures of AI-Powered Agents
I recently came across several GenAI startups that shared some very interesting figures:
11X, which offers an AI-powered Sales Development Representative (SDR), is “approaching $10M in ARR” just two years after its inception—an impressive growth trajectory.
Dust, which creates specialized AI assistants for businesses, surpassed $1M ARR a little over a year after launching.
Additionally, Dust reported a 70% WAU/MAU ratio, indicating strong user engagement.
Harvey, a verticalized AI agent tailored for the legal industry, shared several graphs highlighting high engagement and retention rates for their AI assistant. For instance, they have a 70% retention rate for users who started using Harvey a year ago.
My takeaways from these numbers:
Benchmarking these revenue and engagement numbers shows that these companies are on a fast or very fast growth path. In a currently slow-growing SaaS environment, it shows that you do have fast-growing AI-first SaaS startups out there.
What I also find interesting is that this first wave of AI agents focuses on automating repetitive tasks in white-collar jobs. It feels like they are a direct evolution of Robotic Process Automation (RPA) products like UIPath. Traditionally, RPA software has been more suited for enterprise customers due to its complexity and setup requirements. The new wave of AI agents appears to democratize the market with more accessible products for a broader audience.
The engagement rates are particularly interesting, as a big concern for many AI startups is whether they provide real value to users. In these cases, it seems that the answer is yes.
The very short time span between the founding of these companies and their current revenue levels also shows how GenAI tech has become accessible. It's impressive to see that they could build and sell AI agents in such a short amount of time.
Article of the week
Research and thoughts
Will we soon live in the B2B matrix?
Looking at many AI powered SaaS that emerged the past year and a half, it feels like we may soon live in a B2B 'Matrix' (I’m speaking about the movie with Keanu Reeve), where many aspects of our daily work will be simulated or managed by AI.
I’m particularly thinking about two emerging trends in B2B software:
Digital twins
Simulated environments
1. Digital twins SaaS refer to virtual copies of existing physical systems (like production lines, fleet of vehicles, buildings…), processes, or even human roles (virtual employees), that allow businesses to simulate, monitor, and optimize operations in real-time.
A good example of this trend is Samp. Samp creates digital twins of industrial sites, enabling real-time monitoring to detect and prevent potential issues. By comparing the current state of an industrial site with its original design, the platform helps companies quickly spot problems, improve safety, and reduce unexpected maintenance.
I also include in this trend 'digital twin workers,' such as AI-powered receptionists, which are designed to handle the repetitive tasks that employees would typically perform. They are basically like virtual colleagues (cannot wait to see them doing jokes in slack channels).
2. Simulated environments SaaS refers to software that uses AI to create realistic simulations of work environments or physical processes. This allows professionals to test, train or experiment on these simulated environments/processes in ways they couldn’t with the real life entities.
A good example of this trend is Soma Lab. Soma Lab provides medical students with simulated patient cases to practice diagnosing and treating real-world scenarios. By using AI-powered simulations based on real patient data, the platform allows students to train in a risk-free environment.
Solutions in search of problems?
While I find the ideas behind these concepts intellectually appealing, I can’t help but wonder if they sometimes feel like “solutions in search of problems”. As with many emerging technologies, we often try to apply them everywhere, even when there’s no real need. It was typically the case with blockchain technology, for instance, where the technology is intellectually appealing, but we probably tried too hard to apply it to every possible use case, even when it wasn’t needed.
If you look at GenAI, it excels at analyzing large datasets, detecting patterns, and making predictions based on historical data. This makes digital twin and simulated environment models a perfect match for GenAI's strengths and explain why we see an increasing number of startups applying using them. But are we making the same mistake with GenAI that we did with Blockchain tech?
The way I tend to think about it at the moment is through:
Cost reduction: Does a simulated environment or digital twin cut costs significantly? In some cases, yes—like CAD engineers simulating fluid dynamics or stress analysis in software rather than in labs. Simulated environments are changing the way we design and build physical products.
Labor shortage: Can it help address labor shortages? This is a major issue in many industries, and we may increasingly rely on digital twins and simulations to fill these gaps.
Productivity boost: Does it enhance productivity? I recently read about call centers in the Philippines that have reduced human training time from three months to one by using simulated customer calls. So GenAI can have a real impact on getting workers up to speed.
Obviously not everything needs to be simulated and virtualized, but it seems that we’re in the case where some of these solutions will find real problems to solve.
Business Breakdown of the Week
Interesting startups and products
BuildForMe: Wordpress AI assistant
What is BuildForMe?
BuildForMe is an AI powered assistant for Wordpress. Their AI-powered tool helps users create, maintain, and enhance WordPress websites by automating tasks such as SEO optimization, widget customization and content creation.
What I find interesting:
I know a bit about the WordPress ecosystem because my wife uses WordPress for her business.
As a non-tech person, she paid an agency to set up her WordPress website and hired several independent consultants to maintain and optimize it.
And, wow, did I witness some bad behavior and incompetence in this field (like the time a consultant told her to delete all the images in her site's media library to make her website faster…).
The global market for WordPress services—including design, development, and support—is estimated to be worth billions of dollars. Some estimates suggest the entire WordPress economy (including hosting, themes, plugins, and services) could be valued at over $10 billion annually.
So, it’s a massive market ripe for disruption by AI-powered agents, as GenAI could replace many of the tasks currently handled by humans.
I don’t know if Buildforme is particularly good, but I chose this startup because I believe it’s a field with interesting opportunities to replace human-powered services with AI-driven solutions.