How Has AI Changed Our Acquisition Strategy?

If you've been following my journey with Waterglass, you know we focus on acquiring and scaling niche software businesses with the help of AI.

One of our guiding principles, however, was that AI not only helps to run businesses more efficiently, but that we will be able to fulfil our mission of running them fully autonomously in the future.

It's a simple playbook: find undervalued SaaS products, improve their distribution and operational efficiency, and grow them systematically.

Today, I want to share how we've – over time – rethought our acquisition strategy in the age of Generative AI, and why the old rules no longer apply for SaaS.

Let's dive in 👇

The pre-AI playbook

At Waterglass, agentic workflows have been part of our operational toolkit for some time now, especially in the context of

• customer support
• content generation (blog posts, support articles, ad copies)
• market research
• coding assistance

Our traditional evaluation process for SaaS businesses focused on a set of questions, including:

• Has the business achieved product-market fit?
• What distribution channels are unique to the business? And are they defensible?
• How have MRR, CAC, and churn developed over the past year(s)?
• Where is the lowest hanging fruit for immediate growth?

But we also looked for businesses where we could apply AI to create efficiency gains post-acquisition. The underlying belief is simple: solid distribution channels combined with operational efficiency enables us to build strong, lasting businesses.

This approach served us well. We acquired Pxl, applied AI enhancements where possible, scaled it up, and generated solid returns (4x'd MRR since acquisition).

The AI overhang

The democratization of AI capabilities through foundation models has fundamentally shifted how products and services are built.

So much so that AI wrapper startups started popping up left and right.

For context, AI wrapper startups are often little more than UX layers over existing LLMs, making them easy to launch but hard to defend or scale in the long run.

The video-editing SaaS that got disrupted

Three weeks ago, we were evaluating a promising video-editing SaaS targeting content creators.

The product was simple yet elegant: you upload a video, and it automatically cuts, transcribes, and adds subtitles, optimizing them for platforms like Instagram and TikTok.

The metrics looked great:

• Growing user base with >$10k MRR
• Strong engagement and retention
• Clear market need (every content creator struggles with this)
• Experienced founders with relevant background (who even pitched their business on Dragons' Den!)

Under our old evaluation framework, this would have been a slam dunk. Good product, clear value proposition, scalable distribution opportunities.

But we passed.

Why? The AI risk was too high. The core functionality – understanding video content and reformatting it – was exactly the type of task that AI models are already today pretty good at.

Turns out, validation came faster than expected.

Just one week after we decided not to move forward with the acquisition, both Google and OpenAI released updates to their video generation models that are now almost able to perform the same task – with better quality and faster processing times.

AI as both threat and opportunity

We assess every acquisition target through this opportunity-first lens:

How can AI 10x this product's value proposition? What new capabilities could AI unlock for existing users? Could AI help us reach markets this product couldn't access before?

The framework is simple: we look for products where AI enhancement creates exponentially more value than AI disruption destroys.

Yes, AI can disrupt. But it can also supercharge.

Today, AI helps us operate more efficiently. Tomorrow, it might handle entire business operations end-to-end.

When we evaluate acquisitions now, we ask: Could this business run itself with minimal human intervention in 3-5 years? If the answer is yes, that's not a threat to our model – it's the ultimate scale opportunity.

Evolving our sourcing strategy

Finding truly defensible businesses has become significantly harder, but not impossible.

We've refined our sourcing approach to focus on:

• Niche markets with fewer competitors and slower-moving dynamics
• Industry-specific tools that require domain expertise
• Workflow-heavy products where AI augmentation makes sense
• Data-rich applications where proprietary datasets provide data moats

The key is looking for products where AI becomes a feature, not a replacement.

What's next

Our mindset has evolved. AI hasn't killed the acquisition game; it's forced us to be sharper and more strategic about what we pursue.

We're actively looking for products that foundation models can't easily replace—the overlooked niche gems that solve specific problems for specific people in ways that generic AI tools simply can't match.

The AI revolution isn't slowing down, and neither are we.

If done well, spotting the diamonds can make (or break) your acquisition strategy – but that’s always been true, regardless of AI.