Saltar al contenido principal
Artículo

AI goes “public”: Navigating the new era of access and awareness in 2024

By Tony Ko and Tosia Morris
person giving lecture

Imagine: it’s Q4 2024 and, as the accountable executive for AI, you’re preparing to present to the board on the company’s AI initiatives.

But there’s an issue: while you have robust, forward-looking ROI projections, nothing has been monetized yet. As you prepare for the meeting, you feel the pressure to deliver a compelling narrative that inspires the board’s confidence, laying the groundwork for continued funding for meaningful AI advancements in the coming years.

This year has the potential to be a landmark in the field of artificial intelligence. Unlike 2023, where AI investments were largely experimental, 2024 sees deliberate and substantial investments. Much like a groundbreaking initial public offering (IPO), the democratization and widespread access to AI — catalyzed by OpenAI’s release of ChatGPT in late 2022 — has thrust AI into the public domain, engaging a broader demographic with newfound stakes for AI’s success. With AI predicted to become a $200 billion global industry by 2025, we see levels of investment emulating the serious financial commitments post-IPO in the business world.

Drawing on Gartner’s Hype Cycle, this new faction of AI stakeholders has skyrocketed to the “peak of inflated expectations,” eager to unlock the most promising potential of this breakthrough technology. However, as we’ve seen from AI’s past, the greater the gap between popular expectations and understanding of this technology, the deeper the letdown as early experiments and implementations fail to live up to the grand visions. This disillusionment can lead to widespread disengagement, jeopardizing technological advancement.

With AI at such an exciting precipice, the imperative for 2024 is to close the gap between the bold aspirations for this technology and popular understanding of AI’s current capabilities, sustaining engagement through cycles of experimentation and improvement necessary to propel AI to true mainstream adoption.


AI’s metaphorical IPO — a shift in awareness and access

Prior to 2023, AI was the purview of a small percentage of the population who endured the AI hype cycles, persisting through the multi-decade “AI winters” that bookended the 1980s (more on this later). Today, much like a wildly successful IPO, AI’s recent emergence into the popular sphere has created a paradigm shift where nearly everyone is now an AI stakeholder.

Slalom has seen firsthand the uptick in organizations carving out significant portions of their 2024 operating budgets for AI investments. From $30 million at a well-known charitable trust to $40 million at a midwestern insurance company, these substantial investments stand in sharp contrast to 2023, where organizations were primarily using discretionary funds to experiment with AI. With 70% of executive leaders increasing AI investments this year, the tangible interest and investment underscores a pivotal moment in AI’s trajectory, reflecting market demand for its transformative potential. However, sustaining this momentum will be key to propelling AI innovations toward their full realization and broader societal impact.

Just as going public opens a company up to heightened scrutiny and shareholder expectations, the plethora of new and substantial investments has increased the pressure on accountable executives to deliver equally significant returns. In the open market, investors who are dissatisfied with the performance of their stock may choose to sell their shares, which, at worst, can result in IPO failure. Similarly, AI’s stakeholders need to see returns on their investments or they will disengage. This gap between the expectations of new stakeholders for transformational outcomes and their understanding of the technology poses a serious threat to sustained engagement, a threat that has the potential to bring on another AI winter and delay the inspiring and investible advancements in our world.


Echoes of the past — lessons from AI’s history

While generative AI (GenAI) has recently taken the world by storm, the broader field of AI has had a much longer and more varied trajectory.


placeholder

Popular hype surrounding AI has crested and fallen repeatedly since its inception in the 1950s. During phases of heightened AI activity, or AI summers, there’s a notable surge in innovation and investment as early successes drum up popular hype around the promises of AI. Conversely, during the lows, interest and funding take a downturn as early systems struggle to achieve mainstream adoption. These low periods, or AI winters, significantly set back technological progress as investors pull back on funding necessary to ensure AI’s success.

Following the early golden years of the 1950s and 1960s, AI experienced its first winter when, in the 1970s, interest and investments substantially waned as the pace of technological progress slowed and AI failed to deliver on its early promises. In the 1980s, research projects focused on autonomous vehicle technology (like Carnegie Mellon’s Navlab and Mercedes-Benz’s Eureka PROMETHEUS Project) thrust AI back into the spotlight, leading to renewed interest. However, these early-stage developments didn’t reach the scale or public use that was projected, leading to the notion of a widespread “failure.” Once again, technology hype outpaced reality, and AI experienced its second winter. While interest in AI started to pick up again in some circles starting in the early 1990s and has continued steadily throughout the 2000s, the recent and unprecedented spike has put AI in the spotlight like never before.

Tracing back to the AI winters surrounding the 1980s, we see that hype without substantial progress can lead to disillusionment, which has material impacts on the progress and adoption of AI. This historical perspective underscores the importance of managing expectations in this new era of AI’s “public offering.”


Parallel narratives — IPO successes and AI’s potential

Whether it’s a company offering its shares to the public for the first time or a cutting-edge technology dominating the public’s attention, there are certain characteristics that make a debut venture successful.

Clear vision and value proposition:

A clear and well-understood vision and value proposition not only provide direction and differentiation but also are critical for stakeholder confidence and buy-in. We saw this at Google in 2004, where, despite a rocky start, the company’s ultimate public success was largely due to its clear business model and the undeniable utility of its search engine. To succeed in this new era of accessibility, AI must demonstrate clear value and practical applications, and AI initiatives must be grounded in realistic expectations. WeWork’s failed IPO, driven largely by unrealistic valuations and an unsustainable business model, serves as a cautionary tale of how misaligned expectations can lead to disengagement.

Market readiness and stability:

Product confidence is key to securing investment in any new venture, and investors are more likely to sustain engagement if they believe the market will remain relatively steady over time. In 2004, Facebook initially struggled with issues around its offering price and market stability. Similarly, AI is currently facing an uphill battle for stakeholder understanding. While users generally expect precise and predictable results from technology, the reality is that most products and services, including GenAI, have a standard degree of error tolerance. To fully realize the value of GenAI solutions, we must demonstrate and educate to ensure market readiness. Seeing AI solutions in action provides stakeholders with a concrete understanding of what the technology can accomplish, clarifying abstract concepts and showcasing the potential impact of the solutions.

Global reach and scalability:

The ability to appeal to a global audience and potential for expansion are essential for any venture vying for market dominance. Alibaba’s record-breaking IPO highlighted its dominance in the global e-commerce market due, in part, to its scalable business model, which enabled it to serve customers across different regions and countries. Drawing a parallel, AI must be similarly scalable and universally applicable to tap into its full potential. And we’re seeing this trend play out in real time: as AI technology continues to improve, new techniques including hyperparameter optimization have significantly reduced the cost of fine-tuning and training models as well as key components of retrieval-augmented generation (RAG) architectures. In addition, Pinecone’s vector database boasts cost improvements of up to 50x, while Groq’s language processing units (LPUs) have shown notable progress in reducing latency for enterprise usability. These technological advancements reduce the barrier to custom models that can be leveraged for beyond-intuition-based use cases (e.g., How do I attain my goals?) to achieve accuracy-based use cases (e.g., What should my target goals be?). While these advancements are promising, AI must continue to make gains to sustain current levels of funding as well as increase experimentation and adoption among smaller players.

Innovation and market demand:

The ability to create new and unique products or services that meet the changing needs and preferences of consumers is what truly differentiates the leaders from the pack. Snowflake’s immediate market success, driven by its innovative cloud-based data warehousing, underscores the importance of meeting specific market demands. While many of AI’s new stakeholders are currently at the peak of inflated expectations, to maintain sustained engagement, AI must align to real-world needs by delivering innovative solutions to today’s challenges. Simultaneously, one of AI’s key challenges in 2024 will be keeping up with market demand as use cases continue to explode. Once the viability and feasibility of a use case—such as customer service copilots, email/document assistants, and claims processing (e.g. healthcare/insurance)—are proved, dozens of use cases with varying degrees of complexity can be sourced to keep AI teams busy for six to 12 months. Without a mechanism to regulate the rate at which work is processed, stakeholders may become disengaged or dissatisfied due to delays in delivering AI solutions. Therefore, ensuring efficient throughput is essential to keep stakeholders engaged and satisfied with the AI development process.

Pace of technological advancement:

Throughout AI’s history it’s clear how closely the pace of technological advancement for innovation is tied to levels of interest and engagement. Recently, we’ve seen how quickly OpenAI captured the public’s attention and mindshare, reaching 100 million users and achieving $2 billion in revenue just 14 months after its initial release. Even as the company achieves this milestone they are continuing to innovate, expanding modalities beyond text to include image and video. As technologies like OpenAI rapidly advance and capture public interest, it becomes increasingly important to make these technologies more accessible and user-friendly. The complexity of technology can sometimes act as a barrier to broader adoption and engagement. By simplifying technology, it becomes more inclusive and easier for individuals to understand and use, ultimately fostering greater innovation and participation in technological advancements.

Velocity of incumbents:

With every new product release or new start-up company we will see hundreds of companies become irrelevant and, at the same time, thousands of new companies spawn. DeepMind cofounder and author Mustafa Suleyman predicts “the coming wave of technologies threatens to fail faster and on a wider scale than anything witnessed before.” This proliferation of AI advancements will increase the risk for executive leaders of falling into analysis paralysis. To adapt and succeed in this new era of rapid AI innovation, it is imperative for organizations to focus on three key outcomes: (1) redefining relationships with their customers, (2) redefining their relationships with their employees, and (3) redefining their industry.

For those following the market, it can often feel as though there are more IPO disappointments than successes as initial launches fall short of expectations, given the hype. In examining high-profile IPOs over the past two decades we see the importance of clear value, market readiness, scalability, realistic expectations, pace of innovation, and velocity of incumbents in determining the success of any new venture. With AI currently making its metaphorical market debut, we can apply these insights to take advantage of AI’s current trajectory and make plans to mitigate any impact on progress, successfully avoiding another AI winter.


Bridging the gap — the mandate for 2024

The key to AI’s success in this new era is to align the excitement of “going public” with the practicalities of AI’s current capabilities. This involves educating AI’s new stakeholders, setting realistic expectations, and focusing on responsible AI growth. The substantial AI investments in 2024 mean executives will be accountable for showing a return on these investments. To help our customers navigate these challenges, Slalom has built an AI business value calculator to help organizations focus on high-value AI investment cases while leveraging intelligence to compare business case ideas across industry and function. Returning to the opening predicament of our AI executive eager to show tangible results and secure continued funding and engagement, the goal is to help executives allocate their budgets to realize the maximum potential of AI investments. Now, as you walk into the boardroom, equipped with the right tools and secure in the knowledge that your stakeholders are educated on the realities of AI’s current capabilities, you confidently make your pitch and secure funding to sustain the innovation required to redefine your organization’s relationship with your employees, your customers, and your industry.

As AI enters 2024, the balance between technology innovation and realistic expectations is more crucial than ever. This year has the potential to mark a significant leap in AI’s journey, provided we heed the lessons of AI’s history.




Let’s solve together.