Examining willingness to pay for generative AI startup solutions.

Examining willingness to pay for generative AI startup solutions.

Product Management

8

min read

June 27, 2024

Introduction

While 2023 was the year people discovered and conceptualized Generative AI (GAI), overall spend remained low. In 2024, the technology is more mature, businesses & consumers have carefully watched and are beginning to pull out their cheque books. According to Statista, the market size for machine learning grew 27.8% in 2024, while the market for Generative AI grew 76% (Statista, 2024). With this said, wide-reaching implementation of AI systems is slow as businesses stall while they prepare their data for AI(ificiation) and watch how the market & technology plays out.

Many large & data mature businesses are investing in internal AI research, and model generation. These ‘best in class’ organizations are achieving a return on investment of 13% from Generative AI capex compared to the average of 5.9% according to IBM (IBM, 2023). With ROI for SMBs much lower, they remain cautious on spend as they wait for clear use cases, and front-runners in the race to commercially viable GAI.

Today & Beyond

In 2023, AI captured the world’s attention. This year, it will begin to capture real value. Currently, the majority of AI spend, and innovation is focused on model-building otherwise known as ‘infrastructure’. In 2023, Investment in this area was over 3x higher than horizontal & vertical applications combined (CB Insights, 2024).

We expect over the coming months and years, ‘infrastructure’ will become increasingly intelligent & accurate. As general models become better, and advancements occur in longer increments. Price will become a major point of differentiation leading to commoditization of the technology. As margins for ‘infrastructure’ are squeezed, investment & innovation will become increasingly focused on building higher margin applications atop these models.

There’s a necessary lag between investment in model-building & applications. First, these model-builders must invent the infrastructure & underlying technology for the AI boom, then innovators create value by solving problems.

We are only beginning to see the value created by GAI. While there’s more value in this new technology than at any time in the past, the market is likely at just a fraction of its future potential. A strong indication of this is that the largest AI companies in the world OpenAI, Anthropic, and their competitors Google and Meta are continuing to become more commercially minded around the technology.

In 2024, this transition is clear, it seems every few weeks there’s new advancements in the underlying models; with time, new use cases and value propositions will pop up, and the dollars will follow. It’s clear that we are at the beginning of this shift because the biggest organizations behind the technology are still mostly research oriented. The degree to which these companies are commercially minded is a good indicator of where the market is.

How useful is Generative AI in business in 2024?

The market suggests that many businesses are still apprehensive to integrate AI into workflows. Inaccuracy, Privacy, and Cost remain prohibitory factors in the adoption of Generative AI systems. While many workers take advantage of commercially available interfaces like ChatGPT, embedding Gen AI into core business processes is still a slow transition.

While adoption of the technology into core processes is slow, it is gradual. The use of Gen AI within organizations is becoming broader. In 2024, AI is being used in many more business functions compared to 2023. In one McKinsey survey, when compared with 2023, 17% more people use generative AI regularly in their work & personal lives. 55% of respondents claimed to use AI regularly (Quantum Black AI by McKinsey, 2024).

Developing the underlying models is crucial, but educating the general public on how to use such a tool is perhaps more important. Education & teaching enable the innovation and problem-solving needed to develop GAI applications.

So what sectors are paving the way for commercially successful AI? Marketing & Sales has adopted the technology faster than other sectors (Quantum Black AI by McKinsey, 2024). The automation and outsourcing of tasks like copywriting is occurring at an aggressive pace. In both content-based marketing, and direct sales, generative AI is becoming deeply entangled with peoples’ workflows.

The next most common use cases for AI occur within Product development roles. Information Synthesis, Summarization, and document creation are all basic use cases for generative AI within product development.

What are businesses spending on?

Despite the headlines, investment in old fashioned analytical AI which includes machine learning exceeds investment in Generative AI. According to Mckinsey’s report on the state of AI, in most industries, analytical AI spend still exceeds Generative AI spend by more than 20% (Quantum Black AI by McKinsey, 2024).

Broad Consumer & Business willingness to pay for generative AI remains low, while Analytical AI spend remains consistent. However, the rate of acceleration in Generative AI investment means it will soon overtake Analytical AI. With improving underlying models, startups and other ‘application builders’ stand to benefit, but perhaps they’re early to the party.

Are people willing to pay for your startup’s AI product?

Generative AI buyers can be broadly categorized in three ways; those looking for pre-built solutions (like any product available on the GPT store), those looking to customize their tools with their own data and use cases (like RAG based LLMs), and those looking to develop their own models from scratch (like the models built by Deloitte, and other consulting firms).

Building models from scratch is a laborious, expensive ordeal usually reserved for only the most well-capitalized businesses, think the big consultancies, and big banks. So, to examine the market opportunity for small-scale Generative AI startups, we will consider only the first two categories, those looking for off-the shelf, and customizable solutions.

A McKinsey survey found that half of respondents are using Generative AI through publicly available models with little or no customization (Quantum Black AI by McKinsey, 2024). So, if you’re an AI startup building a solution on top of a third-party model, roughly half of Generative AI adopters are not willing to pay for your solution.

There are just a handful of model building companies competing for 50% of the pie, while there are many more organizations competing for the other 50%. This explains why the Generative AI startup market feels so saturated right now. There’s many entrepreneurs and organizations chasing the relatively few dollars being spent on these types of solutions right now. I firmly believe that these businesses are ‘getting in early’, if you can survive the transition and adoption phase, the market opportunity could be staggering.

Of the respondents seeking an industry-specific, off-the-shelf, or turnkey solution, most are in the Professional Services, Advanced Industries, Healthcare, and Financial Services.

Let’s take a look at some successful Generative AI startups in these industries.

Professional Services: Blue J is a Canadian based startup, their Generative AI Tax & Legal research tool is changing the way Accountants & Lawyers perform research. As a rapidly scaling startup with significant revenue growth YOY, Blue J is a leader in Generative AI solutions in the Professional Services Sector (BlueJ, 2024).

Advanced Industries: GridRaster builds spatial computing platforms for the Aerospace industry, their AI powered Extended Reality (XR) design tool is being used by impressive customers like the US Air Force & Space Force in the design of new aircraft, and Nvidia in the design of new Graphics Technology (GridRaster, 2024).

Healthcare: Atomwise is a YCombinator backed startup using Generative AI in the discovery of new pharmaceuticals. With an April 2024 study showing significant improvements in identifying complementary compounds when compared to traditional HTS discovery, Atomwise is a leader in demonstrating Generative AI value creation in the healthcare industry (AtomWise, 2024).

Financial Services: Collect AI is using Generative AI for receivables and debt collection. With NLP & Generative AI at the heart of their prediction tool, Collect AI is changing the way businesses calculate risk for consumers and borrowers, and how they collect the cash owed to them (Collect AI, 2024).

Conclusion

To conclude, Businesses & Consumer willingness to pay for AI solutions lags behind investment in the development of these new technologies, but with new technologies, it always does. In 2024, investment in AI is growing rapidly, but many buyers remain in an ‘experimentation’ phase. Despite the hype, only a small fraction of AI companies are succeeding in the current market. Those building Generative AI infrastructure are benefitting the most right now. However, the commodization of AI infrastructure will lead to lower costs for everyone involved. Combined with an increasingly powerful technology, applications of Generative AI that solve tough problems will continue to pay dividends to the businesses that survive long enough to identify strong use cases.

So, if you’re reading this as a small-scale GAI application startup, have patience. Buyers are:

• cleaning their data

• training their employees

• surveying their options

• waiting for more advancements in the underlying models

• and letting the early adopters make mistakes.

TLDR; Spend on Generative AI (GAI) is largely focused on R&D within the underlying models. These models represent the ‘infrastructure’ behind GAI, necessarily there’s a lag between developing the infrastructure and capturing value. Spend on GAI applications remains low, however, as the cost for better models goes down, and the general public become more skilled in using this technology, applications will be able to capture value that infrastructure couldn’t. So far, Industries that have strongly validated their GAI use cases include Professional Services, Advanced Industries, Healthcare, and Financial Services. Examples of companies succeeding in these industries include Blue J, GridRaster, Atomwise, and Collect AI respectively.

© Lancey Software Inc. All rights reserved.

Legal

SOC 2 Type II Compliant

Legal

SOC 2 Type II Compliant

© 2024 Lancey Software Inc. All rights reserved.

© Lancey Software Inc. All rights reserved.

Legal

SOC 2 Type II Compliant

Legal

SOC 2 Type II Compliant

© 2024 Lancey Software Inc. All rights reserved.