Digital transformation has been underway for more than a decade, enabled by the confluence of three powerful forces: data, mobility, and the cloud. Fed by the information gathered by mobile- and cloud-based systems, data has grown in its influence, and has given rise to increasingly powerful AI systems.
From those systems we now are experiencing the AI-first transformation of business, which supersedes digital transformation and will extend its impact by an untold massive degree.
Consumers’ expectations have extended to every aspect of their lives—and businesses, buoyed by advancements in AI and data, accelerated by the product-led growth motion, and disrupted by decentralization—are creating never-before-seen solutions that enable us to do our best work.
This is more than a trend.
It’s a generational transformation similar to the likes of digital, mobile, and cloud, with the potential—even inevitability—to fundamentally change the way businesses operate and the global economy grows.
The digital transformation of business has been the dominant narrative of the past two+ decades. The AI-first transformation of business, and AI-first technology stack that supports it, will be even bigger.
In turn, those founders are armed to build the AI-first technology stack, empowering customers with transformational products that unlock the potential of individuals and enterprises. Founders building this stack help customers reach goals and uncover opportunities that were previously unattainable, or invisible. This empowers those customers—all of us—to reach ever upward in our desire to do our best work.
Enabling the AI-first transformation of business.
The team at Wing believes deeply in a future that is built on data, powered by AI, and put to work through increasingly autonomous applications.
We also firmly believe humans will not be replaced by AI—but humans working with AI will replace humans working without it.
Data forms the foundation of the AI-first tech stack. The data layer consists of scalable, secure platforms to process and manage data. These platforms store and analyze “ground truth” data.
The next layer is where AI processing occurs, either through a developer’s own AI technology or through the utilization of industry foundation models—and probably both. While industry models are extraordinarily powerful, developers of specific applications can build adaptation layers around them and deploy additional proprietary models above them, improving fit with the target business process.
The AI layer extracts semantic understanding and insights from data, and generates content. It serves that up to a business process layer built with deep understanding of the users, their goals, and the business outcomes they are aiming to achieve.
Atop the stack sits autonomous applications, which go beyond workflow automation to automate all that is automatable, and to arm humans with as much context and insight as possible to address the remainder. These autonomous applications help steer business processes towards optimal outcomes.
Let’s look deeper at each element of this AI-first tech stack, and how it drives the AI-first transformation of business.
Data is the foundation and fuel of our modern enterprise.
Nothing in the AI-tech stack is possible without novel data, and at Wing, we continue to invest—as we have for several decades now—in advanced data platforms geared for crucial data types and workloads.
There is no better expression of data’s importance than Snowflake, a company we first invested in at the seed stage. Snowflake has transformed the way businesses use data—today, its Data Cloud underpins an entire generation of analytics and operational applications literally built on data.
Modern data companies are increasingly data-wealthy, generating and collecting novel data. This data represents a meaningful advantage that other companies can’t overcome with capital or engineering resources. This phenomenon is also an adoption driver for data platforms. Growing product usage, deeper instrumentation, and a desire to fully exploit first party data all require an ever greater capability to store and analyze ground truth data.
In reality, almost all businesses are in the business of facilitating data analysis somehow; by sharing data, analyzing it, operationalizing it, and much more. That opens doors to managing the full data lifecycle, from processing to storage to security and sharing.
It also opens the door to increased data usage in the life sciences, which is why one of our core areas of investment within data lies at the intersection of data and biology, or what we call BioXData. As we seek investment opportunities, we look for companies up to the task of operationalizing their data, and companies that help others in this process.
From data as our foundation (and fuel), we move to the engine of the AI-first tech stack: AI.
The modern enterprise is powered by AI.
The AI layer of the AI-first tech stack is made of systems that help extract semantic understanding and insights from data. These AI systems and platforms not only generate critical insights, but also help to make data and insights accessible within product experiences.
Wing company Pinecone is a great example. We led their seed, and since, they’ve quickly become a necessary AI-powered layer to manage and search vector embeddings. Their platform enables businesses to build sophisticated and scalable AI-centric applications.
Some companies within this space offer analytical AI, used to classify, organize, or reason about data. Others now offer generative AI, which creates data in the form of content or content enhancement. These systems are designed not just to support ad hoc analysis, but to be productized in support of business processes.
Next, AI processing occurs, either through a developer’s own AI technology or through the utilization of industry foundation models—and probably both. While industry models are extraordinarily powerful, developers of specific applications can build adaptation layers around them and deploy additional proprietary models above them, improving fit with the target business process.
Over the last decade, we’ve unlocked the utility of new data types, from images and audio to unstructured text and even within biology. Paired together, data and AI enable true personalization. Today, foundation models are being rapidly adopted, often combined with internal models to improve fit for particular business problems.
As these data and models become must-haves, AI is quickly becoming an embedded element of every next-generation business application. This leads us to the rise of autonomous applications.
Autonomous applications steer business to optimal outcomes.
Business applications exist to accelerate and improve business processes, and the next generation of these applications will be powered by AI.
Legacy applications were workflow-driven and sought to automate a defined business process. AI-powered applications go beyond workflow to automate the automatable, and to arm humans with as much context and insight as possible to address the remainder.
Some of these apps will become truly autonomous with no humans in the loop (“level 5”), and the business process runs on its own. Performance is measured simply in outcome delivery.
More typically, the application will be semi-autonomous, helping steer the business process and enabling human operators to more consistently and efficiently achieve superior outcomes.
Autonomous applications can be used differently, and allow modern enterprises to compete in creative ways. Exciting autonomous application companies can already be seen in horizontal enterprise functions like sales (Gong, SetSail), marketing (Copy.ai), engineering (Jellyfish), recruiting (Findem), employee service (Espressive) and even content moderation (Spectrum Labs). Other enterprise functions such as customer success, finance, HR and product management will be fertile ground for new autonomous applications as well.
Vertically-focused, AI-powered applications are also important as the data tornado touches down in specific industries. Life science promises as profound a transformation as any. Advances in data collection, such as low-cost sequencing, have radically scaled data availability. Mathematical and computational approaches open the door for new research tools, medical equipment, diagnostics, and even therapeutics enabled by data and AI. Pharma and biotech are in dire need of computer and data science chops, and companies like Seer (proteomics data platform) and Asimov (AI-powered synthetic biology) aim to help.
The impact of AI on user experience is surprisingly under-appreciated. Great consumer-grade design is part of the equation, yielding a measure of delight that drives adoption and usage—a familiar story.
What’s new is how this usage generates proprietary data through the operation of the application itself. Smart business applications will combine “synthetic data” with domain-specific machine intelligence to raise the bar. Streamlined, context-driven experiences are now possible, free of the drag and distraction of workflow pathways irrelevant to the user’s goals.
Autonomous applications can offer a whole new level of dynamic, optimized, and yes, delightful experiences geared to specific needs of customers, employees, and business partners. Obsession with experience will be pervasive in the world of AI-first businesses, and will help transform the nature of work itself.
The adoption of autonomous apps is accelerated by the product-led growth motion, which brings us to the upward driver of the AI-first tech stack: PLG.
Adoption and outcomes are accelerated by product-led growth.
The product-led growth motion enables these layers of data, AI, and autonomous applications to be adopted faster than ever before.
End user experience has become the driving factor behind product adoption, which has created an acceleration wherein adoption is limited only by product and awareness. PLG will continue to accelerate the pace of adoption of innovations, broadening their reach and expanding their impact—in turn accelerating the use of data and AI in enterprise software across the board.
Enterprise software customers are demanding more open, accessible ways to adopt and consume software. The PLG business model is enabling a new wave of companies to be built, by decentralizing decision making and boiling it down to its most atomic level: the end user. And while PLG does not have to be AI-centric, it is a powerful driver of the AI-first transformation of business.
The rise of the individual in the B2B customer universe is inextricably linked to the rise of PLG as the preferred go-to-market mechanism. The expansion of customer types and buying centers has increased the overall opportunity space available to new entrants, and gives them advantages compared to incumbents who have built their businesses around an older, narrower customer set.
Before it established itself as a reliable accelerant to business, PLG first acted as a disruptive force that shook up the traditional ways technology was adopted. The next major disruptive force on the adoption of new technologies and products is decentralization.
The new network effects of decentralization.
Decentralization has proven to be a massive disruptive force, enabling new services previously impossible in classical systems—both at the infrastructure layer with the decentralized architecture of blockchain, and at the app layer, now comprised of decentralized applications with no single point of control or failure. The incorporation of AI into decentralized systems is only adding to the power and applicability of this new architecture.
Decentralization is reshaping how products are adopted in terms of the hyper focus placed on the end user (versus IT departments or enterprise heads). It’s also empowering these and other individuals relative to corporations, through things like DAOs.
In this way, blockchain will facilitate novel designs in terms of where ownership and power lies between the individual (the consumer, creator, or employee) and the collective (a DAO perhaps, or a traditional enterprise). The implications of this will create major shifts in the fabric of our society and economy.
We’re in the early days of blockchain and web3—startups providing web3 infrastructure, developer tools, and security are just in their infancy. But the signs of growth and significance are promising, especially if you follow, as we do, developer interests. We invest where the best and brightest technologists spend their time—and they are increasingly spending their time on web3.
The forces of decentralization and AI come together in the life sciences, where a major disruption is taking place.
BioXData: A transformational time in healthcare and medicine.
Healthcare and medicine are at an inflection point.
Drug development is rapidly evolving to a new world where drug creation is decentralized and data-enabled. Like how information technology was disaggregated over the last two decades with contract manufacturing, applications, and internet distribution, pharma’s drug development stack is already being decentralized: today, a whopping 80 percent of discovery for new drugs is by biotechs—not big pharma—with pieces of development and manufacturing fast to follow.
We are living in a perfect storm of enablers: machinery to generate novel data in the -omics, dramatically faster throughput, and the golden age of AI to parse the data and generate valuable biological insight.
Today’s companies can be purpose built to marry biology and computation with the goal of selling to pharma leveraging new business models.
At Wing, our BioXData thesis is built on a strong belief that while AI is reshaping almost all industries, few will be as greatly impacted as healthcare and medicine. The data-driven, AI-first transformation in medicine and healthcare is early but already underway.
Healthcare was previously verticalized in brick and mortar locations with patients seeing providers at predefined points in time. Unfortunately, disease progression simply doesn’t work that way. Today, with care via the internet available for the first time, tech is addressing healthcare’s top problem: access.
A provider in Salinas can see a patient in San Diego via cellphone or laptop, enabling more continuous diagnosis and intervention. And, the decision maker is increasingly the provider, who is empowered to do their best work in payor-reimbursed arrangements.
This is a distinct business model evolution that is paving the way for personalized care and management of costly afflictions starting with diabetes, mental health, and musculoskeletal health (to name some early few).
At Wing, we’re thrilled to partner with intrepid founders at ground zero who also believe in data-powered, tech-enabled solutions in healthcare and medicine.
Do our best work—together.
Much like the advent of mobile and cloud technologies drove innovation and excitement, we believe the impact of the AI-first transformation of business will expand what’s possible and ultimately elevate our human potential to do our best work.
The traditional constraints that limited individuals and enterprises are, in many ways, gone. That transforms the rules of engagement—and the rules of competition.
We must remember, though, that the while we’ve discussed the ways in which AI is the core of this transformation, it is humans who are the heart and soul of what makes the transformation possible, and even worthwhile. To that end, we are excited and humbled to work side by side with the exceptional founders pioneering this new space, and to do our best work—together.