Gen AI Unlocking Consumer Innovation

Setting up the scene

We wanted to kick off Consumer AI Decoded by exploring the trends we see and foresee. Trends that we believe will shape the future of consumer business. Along the way, we’ll break down some of the key concepts at the heart of these discussions, ensuring everyone stays on the same page.

Defining Consumer AI

Before diving in, let’s clarify what we mean by Consumer in this context. When we talk about consumer businesses, we’re referring to companies that serve individual end users (like you). These are people who rely on products and services in their daily lives to help them accomplish their goals, including:

✅ Buying groceries
✅ Ordering food
✅ Improving fitness
✅ Asking for information from an AI assistant
✅ Filing a home insurance claim

Our focus, however, is on digital consumer businesses, rather than physical goods. Digital products for consumers are where our core experience and expertise lie. We’re fascinated by how digital products reshape how we live, consume, and interact driving innovation in ways that physical goods cannot.

Consumer AI here refers to AI-powered tools, applications, and experiences built directly for individuals, rather than enterprise solutions operating behind the scenes. Think of popular tools like ChatGPT or Replika.

📊 See the chart below for a clearer breakdown and some examples

Note: Chart was generated using ChatGPT-4, but some logos came out inaccurate, despite multiple prompt refinements

There is a growing overlap between consumer, prosumer, and enterprise propositions. A “Notion” user might begin as a consumer, evolve into a prosumer with API usage, and eventually scale into an enterprise plan at work. It’s something we’ll explore in future posts. For now, it’s important to realize these definitions aren’t always black and white.

The Changing Investment Landscape

From an investment perspective, the landscape for consumer-focused startups has shifted. Over the past decade, growth in consumer investment has slowed. In Europe, despite a 10x growth in VC investment overall, consumer funding has lagged behind. According to Atomico’s European Tech report, the consumer sector's share dropped from ~30% to ~16% in 2024. In the U.S., we see a similar trend: in 2024, the 100 most active VC firms allocated just 6% of funding to consumer startups just half of what it was two years prior. This marks the lowest share since 2018.

This decline is driven by investor preference for safer ROI sectors, post-pandemic corrections, and macroeconomic conditions like high interest rates partially influenced by global conflicts. These factors have made it hard for consumer-focused startups to secure funding.

With reduced funding, innovation in the consumer space will likely take longer to materialise. When it does, AI, specifically Gen AI, will play a critical role.

Why Gen AI?

AI refers to technologies that mimic human behavior, a concept dating back to 1956 when pioneers like John McCarthy and Marvin Minsky convened at the Dartmouth Conference for the first discussion of the topic.

Gen AI, on the other hand, is a much more recent term, emerging with the advent of ChatGPT in 2024 and the underlying transformer model architecture that it was built around. GPT stands for Generated Pre-trained Transformer, a “transformative” topic that would require several posts to explain more fully.

Why Is Gen AI Revolutionary?

Gen AI can now generate:

✅ Text
✅ Images
✅ Audio
✅ Video
✅ Code

It learns from vast datasets at speeds unimaginable just a few years ago. A critical breakthrough behind this was the attention mechanism in transformer models, allowing systems to focus on key elements of a sentence or question. These operate in much the same way we prioritize important cues in a conversation allowing them to figure out what parts of a sentence or question matters most, so they can better understand what a person wants or means. This dramatically improves efficiency and accuracy in tasks like understanding language, generating text, and recognizing patterns.

Companies like NVIDIA played a pivotal role by offering parallelized processing power chips (GPUs) which dramatically accelerated AI training capabilities. This hardware capability when paired with transformer architecture, marked a major turning point. It's no surprise NVIDIA and Microsoft briefly traded places as the most valuable companies in the world in 2024 with over $3 trillion market caps.

A New Era of Gen AI native businesses

20 years ago, innovations such as geolocation, touchscreens, push notification and seamless payments allowed companies like Airbnb, Uber, Instagram, TikTok, DoorDash, Snap, Monzo and others to grow rapidly (in the Mobile/Cloud era).

Today, Gen AI is likely to lower barriers to entry, speeding up development, and reducing operational overhead. This is crucial for consumer startups burdened with rising acquisition and development costs.

📊 See below latest 5 waves of Consumer Technology Innovation

However, while funding in Gen AI is booming, much of the capital is going toward foundational and model layers, leaving consumer applications underdeveloped. There is huge potential to open up a new generation of business that will come at a much faster pace than what we have seen before. We are seeing this today with the example of Lovable which is Europe’s fastest-growing startup reporting $30M ARR in 120 days.

The Gen AI Stack

To understand this investment dynamic, let’s break down the three core layers of the Gen AI stack. Think of building a house:

• Infrastructure Layer: The groundwork where essential components like infrastructure, compute power, and pipelines are built. (e.g., NVIDIA, AWS)

• Model or Foundational Layer: The tools that shape the house and its walls, doors, ceilings. In Gen AI, this is driven by large language models (LLMs), which provide intelligence and make applications function (e.g., GPT-4, Qwen3, Claude)

• Application Layer: The services and experiences enjoyed within the house—like furniture, appliances, and entertainment. This is where consumer-facing applications come to life! (e.g., AI chatbots, creative apps)

Notes: OS (Operating System); ML (Machine Learning); AWS (Amazon Web services = Cloud))

What’s Next for Gen AI Applications?

The Gen AI stack mirrors past tech eras. During the PC era, hardware and OS investment came first (Intel, Microsoft), followed by apps like Photoshop. In the Web era, it was networks (Cisco) and platforms (Google) before Facebook, YouTube, and e-commerce thrived. Investors prioritize these areas because they enable long-term innovation. But once these foundational layers mature, applications emerge rapidly, leveraging the tech to create consumer-facing solutions.

Gen AI is now at a similar inflection point. The consumer explosion is coming.

📊 See below a table breaking down the different layers per Technology wave

Gen AI’s True Value: Cost, Speed, or Growth?

Where does Gen AI offer the biggest impact? The answer is that it depends.

• D2C startups: Value creation is key

• Scaling businesses: Velocity and cost-efficiency dominate

• Incumbents: Often struggle to innovate at scale, constrained by legacy systems.

In the UK, however, despite talk of value creation, many startups are prioritizing AI cost saving use cases over growth. This reveals an innovation gap: startups use Gen AI tactically, not strategically.

The Shift Is Coming

This dynamic won’t last. As Gen AI matures, its value will shift from operational savings to revenue generation and customer transformation reshaping industries and redefining competitive advantages. The next wave of AI-native companies will be those that seamlessly integrate AI into their core operations. They will emerge as industry leaders.

Who will be the Airbnb or Uber of Gen AI? Time will tell. But one thing is certain: Gen AI is transforming consumption. But the real question is: in what ways? And when will we start to truly see its impact?

Stay tuned, as these are the very questions we aim to unpack in future posts..

This revolution is just beginning.

Notes:
Tools used for this post: Chat GPT-4 (to create a Excalidraw look-alike table), Co-pilot & Notebook LM for the Podcast
Sponsors: None - just us, for the greater good..That said, we’re incredibly grateful to everyone who continues to share their views and shape ours!