Transforming Code, Care & Classrooms with AI

Assessing the Current State, Future Impact and Investment Potential of AI in Software, Healthcare & Education

Thank you to Vayum Arora for your collaboration on this research.

This article also had input from leading founders in the AI space. Thank you for your time and insights… 

For coding; Amjad Massad (CEO, Replit), Arvind Jain and Matt Kixmoeller (CEO & CMO, Glean), Chris Hatter (CISO, Qwiet.ai), Danny Allan (CTO, Snyk), Jason Warner (CEO, Poolside.ai), Josh Haas (CEO, Bubble), Steve Sewell (CEO, Builder.io)

For care; Abdel Mahmoud, MD (Founder, Anterior), Darius Meadon (CMO, Owkin), Ganesh Padmanaban (CEO, Autonomize.ai), Jeff Chang, MD (Co-founder, Rad AI), Liran Belenzon (CEO, BenchSci), Matt Ko (President, DeepScribe), Nick Talken (CEO, Albert Invent)

For classrooms; Mark Angel (CEO, Amira Learning), Rajen Sheth (CEO, Kyron Learning), Tom Sayer (CEO, Ello)

Introduction

Imagine a world where code writes itself, AI assists doctors in real-time, and every student has a personalized tutor available 24/7. This is our new reality, driven by the latest advancements in artificial intelligence.

This summer, we wanted to develop a perspective on the potential breakthroughs, and investment opportunities in Vertical AI. 

After looking into many different industries, we are excited to see how the integration of artificial intelligence into specific sectors is poised to revolutionize three critical areas: software development, healthcare, and education. We believe these sectors are primed for a leap that will redefine how we code, treat, and learn.

Our research examines the current state, future potential, and investment opportunities in these AI-driven transformations.

Key Findings:

1. Software Development: This is our “Macintosh Moment”. AI tools are democratizing coding, potentially expanding the developer base from 150 million people to billions, while simultaneously raising software quality standards. 

2. Healthcare: AI is alleviating administrative burdens, enhancing patient outcomes, and accelerating drug discovery, with some companies reporting millions in cost savings.

3. Education: AI-powered personalized learning platforms are showing significant improvements in student performance, with up to 40% increases in retention rates.

In our research, we also uncovered a common thread across these sectors: the potential for AI to dramatically enhance human capabilities rather than replace them. From collapsing the software development lifecycle to slashing administrative burdens in healthcare and personalizing education at an unprecedented scale, these Vertical AI sectors are set to kickstart a new era of innovation and efficiency.

In this deep dive, we'll explore how Vertical AI is reshaping these industries, backed by insights from pioneering leaders and real-world case studies. We'll examine the challenges being tackled (like does CapEx spending equate to sustainable, future revenues), spotlight some leading companies, and unpack the long-term implications for professionals, businesses, and society at large.

AI's Transformative Impact on Software Development: The Next Evolution of "Software Eating the World"

The $400 billion global software development industry stands at the precipice of an AI-driven revolution. As artificial intelligence reshapes the landscape, it's not merely posing challenges but creating unprecedented opportunities for engineers and software companies alike.

Democratizing Software Development

The software industry has long grappled with a significant skill shortage, but AI is poised to dramatically alter this dynamic. Jason Warner, former CTO of GitHub, Managing Director at Redpoint Ventures, and now founder of Poolside.ai (an advanced AI-powered coding tool), predicts a seismic shift: "The number of people able to create and code is going to jump from 150 million to billions... magnitudes."

This democratization is already evident. Amjad Massad, CEO of Replit (an AI-powered online coding platform), notes a surprising trend: "Since the start of the ChatGPT revolution, we're seeing way more than just engineers signing up. Non-engineers are extracting even more value from AI coding." We reviewed many Replit case studies and we love this example where someone went from a designer at Brex to a (solo) entrepreneur building multiple successful AI products

Steve Sewell, CEO of Builder.io (an AI-driven platform for visually creating and managing high-performance web pages without extensive coding), compares the impact of AI coding tools to the Macintosh's effect on personal computing: "I think the impact could be as big as the Macintosh had on personal computing. This is our Macintosh Moment of Programming. The GUI unlocks computing for the world. The huge unlock is this is going to make everyone a developer."

While AI lowers barriers, it’s not rendering human developers obsolete. Instead, it’s enhancing productivity and innovation. Warner emphasizes, "We're going to demand higher quality because the best will stand out. It's the nature of the market."

One of my favorite parts of researching this is seeing what kinds of apps people can build on top of these tools. Check out Riley Brown AI (a builder in public, with no coding background, now building complete apps in under 2 hours) and McKay Wrigley

There’s a few other notable coding solutions like Claude Artifact, Cursor.ai, Codeium and Magic. 

Low-Code and No-Code Solutions

The rise of low-code and no-code platforms is further democratizing software creation. Josh Haas, CEO of Bubble (a no-code app development framework), predicts, "I think we'll see a radically different landscape, a real phase 2 of 'software eating the world'."  Josh shared this case study where a Senior Manager of Revenue Operations at a logistics company was able to develop a user interface for managing time-critical organ shipments without any coding knowledge. This allowed the company, Airspace, to not only secure a multi-million-dollar annual contract, but also to facilitate 3,500 organ donations annually.

Matt Kixmoeller, CMO of Glean (an AI-powered enterprise search and knowledge management platform), highlights the potential: "Think of Glean a bit like an employee who has been at your company from Day 1, attended every meeting, read every email, and brings all that information as context to helping you get work done. The real opportunity here is to enable non-technical employees to automate their work."

Revolutionizing the Software Development Lifecycle (SDLC)

AI is set to streamline the traditionally complex and time-consuming Software Development Lifecycle (SDLC), which encompasses planning, designing, developing, testing, deploying, and maintaining software applications. Jason Warner envisions a dramatic shift: "The SDLC will collapse into the models, and as that happens, we'll go from 100 million to 200 million to 500 million to billions of people, and then they'll be able to create much more sophisticated software in that scenario. The SDLC exists largely to do two things: behavior shape and offload cognitive load. Collapsing these into the model pre-guarantees these aspects."

Sewell adds, "AI will make monotonous tasks get done instantly and move 5x faster."

This transformation will have far-reaching economic implications. Haas notes, "The economic impact is that custom software will get cheaper. The lines between building and buying software will move drastically because building will become much cheaper. This will democratize software development, making it accessible even to 20-person companies."

As AI assumes more coding tasks, the nature of software development skills will evolve. Warner predicts a shift towards conceptual understanding: "You could have someone who theoretically writes the next great American novel who has no idea what you mean when someone says first person or third person narrative structure or a three-story arc or the hero's journey. In the same way, I think concepts will become important rather than the nuance of programming. Concepts such as memory allocation, changing variables, and understanding how systems interact will remain crucial."

The learning process itself will change, with Warner noting, "The baseline is going to be using an LLM. Baseline is going to be using an AI system." Massad adds, "but the amount of coding you need to learn to be productive is dropping month-over-month."

Enhancing Software Security

AI is also revolutionizing software security. Companies like Qwiet.ai and Snyk are leveraging AI to detect vulnerabilities, automate threat detection, and respond to potential breaches more effectively than traditional methods.

Danny Allan, CTO of Snyk (a developer-first security platform that leverages AI to identify and remediate vulnerabilities in code), warns of the dual nature of AI in security: "While it helps with productivity, it can also introduce vulnerabilities from open-source code."

Chris Hatter, CISO at Qwiet.ai (a cybersecurity platform that uses AI to detect and prevent vulnerabilities in software), describes their approach to cybersecurity as "autonomous appsec," where AI agents detect, triage, and fix code vulnerabilities. They’re also developing a natural language interface, enabling security teams to ask direct questions like "Where is my biggest risk?" without relying on traditional UI/reporting tools.

Impact on Jobs

In the near term, AI will boost developer productivity by automating repetitive tasks, enhancing code quality, and accelerating the software development process. This may reduce demand for junior roles centered on routine work. However, long-term prospects are promising as AI shifts focus to more complex, innovative projects, increasing the need for specialized skills in AI development and integration.

As Haas notes, "While AI might displace certain jobs, it won't destroy them. The rapid reduction in software production costs will open doors for new projects, previously too expensive to pursue. This will ultimately drive a higher demand for software creators who can leverage AI to its fullest potential.” The shift from traditional software development roles to AI-assisted and no-code creation represents a relatively smooth transition in the landscape of technological job evolution. This transformation is less disruptive than many historical examples of job displacement due to technological advances.

The skills required in conventional software creation—whether in programming, design, or product management—often translate well to the emerging AI and no-code paradigms. Moreover, the professionals most affected by this change typically possess high incomes and strong technological aptitude. These attributes position them favorably to adapt and retrain compared to workers in many other fields facing technological disruption. As a result, while the transition is significant, it offers a more gradual and manageable path for those willing to evolve their skillsets.

Investment Potential

The intersection of AI and software development offers a rich landscape for investment. Warner highlights the immense potential: "We are just at the beginning. This phase will be the most transformative in the history of software development." As AI becomes more integrated, it democratizes software creation while simultaneously elevating quality standards. Sewell captures the opportunity: "When programming becomes as intuitive as using GUIs, we’ll unlock entirely new possibilities." The convergence of reduced development costs, enhanced quality, and broader accessibility sets the stage for unprecedented growth and innovation, making this sector a prime target for investment.

Danny from Snyk: “I expect the long-term impact of AI will more than justify its investment.  They key will be understanding what problem you are facing and then measuring and ensuring that the AI is tuned to specifically address this issue.”

AI's Revolutionary Impact on Healthcare, Drug Discovery & Material Sciences: The Dawn of a New Era

The $10 trillion global healthcare industry stands on the brink of an AI-driven revolution. Within this vast sector, R&D plays a crucial role, with annual spending exceeding $182 billion. As AI technologies mature, they promise to reshape healthcare delivery, research, and administration, offering solutions to longstanding challenges while creating new opportunities for innovation and investment.

Alleviating the Administrative Burden

One of the most pressing challenges in healthcare today  is the overwhelming administrative burden on healthcare professionals, which not only contributes to burnout but also detracts from patient care quality.  AI presents a transformative solution by automating routine tasks and streamlining workflows, allowing clinicians to focus on what matters most—patient care.

Abdel Mahmoud, MD, founder of Anterior (developer of 'clinical-decision support as a service' or ClinDaas), explains: "Imagine if healthcare admin was as effortless as swiping a credit card. Our AI Clinical Co-pilot, Florence, takes over the monotonous admin work, reducing hours of paperwork to mere minutes, allowing clinicians to simply review and confirm.”

Ganesh Padmanaban, CEO of Autonomize.ai (creator of AI Copilots for healthcare professionals), emphasizes the supportive role of AI in healthcare: "Autonomize's AI Copilots are designed to augment the capabilities of healthcare professionals, enhancing their efficiency and effectiveness without replacing them. We are giving healthcare knowledge workers a superhuman sidekick!"

Burnout has become a critical issue in healthcare, with one in five clinicians planning to leave the profession within the next two years due to overwhelming stress. Matt Ko, President of DeepScribe (an AI-powered medical transcription service that automates the process of documenting patient encounters), emphasizes the importance of addressing this crisis: "Burnout is one of the biggest issues in medicine today. DeepScribe is life changing to physicians. DeepScribe is a game-changer, cutting documentation time by 75% and improving job satisfaction and patient connection."

The impact of such AI-driven solutions can be significant. Padmanaban shares a compelling example: "When a leading health plan integrated Autonomize's AI copilots into their prior authorization process, the automation resulted in a 40% reduction in processing time. This enhancement in workflow efficiency not only yielded millions in cost savings but improved patient access to necessary treatments."

Ko said, “On average we cut down documentation time by roughly 75%. Surprisingly human scribes are much more inaccurate than you would expect. Against our benchmarks, expert human scribes on average perform at an accuracy rate of 63%. Using the same benchmark, DeepScribe performs at an accuracy rate of 95%.

Several other innovative companies are making strides in this area. Abridge (AI-powered clinical note generator) uses proprietary AI to convert patient-clinician conversations into structured clinical notes in real-time, saving clinicians an average of two hours daily. Ambience (developer of an AI operating system for healthcare) has developed an AI operating system that enables a reported 78% reduction in documentation time. Nabla (provider of AI-assisted clinical documentation) currently helps 15,000 physicians, significantly reducing stress and improving patient care by streamlining clinical note generation.

Enhancing Patient Outcomes

Beyond administrative efficiencies, AI is making significant inroads in improving patient care and outcomes. Hippocratic AI (developer of safety-focused Large Language Models for healthcare) has developed the healthcare industry's first safety-focused Large Language Model (LLM), focusing on patient-facing, non-diagnostic applications. Their AI agent, "Diana," can perform tasks such as patient follow-ups, care coordination, and post-discharge management.

We connected with Jeff Chang, MD, co-founder of Rad AI (developer of AI solutions for radiology). Jeff started med school at 16 and is the youngest radiologist and the second youngest doctor on record in the US: "Rad AI's solutions have been adopted by > 40% of all US health systems and 9 of the 10 largest US radiology practices. On average, we save providers one hour per 9-hour shift, while reducing provider fatigue/burnout and improving the quality of patient care."

Other notable companies in this space include Imagen AI (provider of AI-powered medical imaging analysis), enhancing medical imaging analysis, and Viz.ai (developer of AI-powered stroke detection software) which analyzes CT scans to identify potential strokes and alert specialists in real-time.

Revolutionizing Drug Discovery and R&D

AI's impact on healthcare is extending far beyond patient care, reshaping the landscape of scientific research and drug discovery. By leveraging machine learning, natural language processing, and predictive analytics, AI accelerates the traditionally time-consuming and costly drug discovery process.

Companies like Owkin (an end-to-end AI-biotech that uses cutting-edge causal AI to unlock precision drug discovery, development and diagnostics) are at the forefront of this revolution. Darius Meadon, CMO of Owkin, explains: "By leveraging our network of leading academic partners and medical experts, we access world-class multimodal patient data through privacy-enhancing technologies. This strong data foundation allows us to use AI to connect the dots across scales – from molecules to cells to the whole body – to capture the causal links of complex biology."

BenchSci (a disease biology Generative AI platform) is making significant contributions to this field. Liran Belenzon, CEO & Co-Founder of BenchSci, highlights their vision: "We don't think about AI as a way to replace what scientists do. Instead, our approach is centered around how we can take what AI excels at...and build upon this to empower scientists."

Other companies like Insitro (machine learning-powered drug discovery platform) are using machine learning to analyze liver biopsy samples and genetic data, leading to more accurate disease state predictions and potential new drug targets. Huma AI (provider of AI-powered post-market surveillance solutions) has enhanced post-market surveillance capabilities for a leading IVD manufacturer, allowing for competitive insights analysis to be completed in minutes.

Beyond healthcare, we also learned about R&D advancements in material science. Nick Talken, CEO and founder of Albert Invent (AI-powered invention platform) explains: "Traditionally, it might take 200 iterations to reach a breakthrough. Albert collapses that down to 50 or even 20 cycles. We help scientists design the most intelligent experiments that provide the most relevant data, so they get to their breakthrough faster."

The Future of Healthcare & Scientific Research Jobs

As AI continues to integrate into healthcare, it will likely augment rather than replace human roles. In the short term, AI will enhance the capabilities of doctors, hospitals, researchers, and administrators by streamlining workflows and improving diagnostic precision.

Long-term, AI will empower healthcare professionals with advanced tools for precision medicine, predictive analytics, and personalized care, fostering a collaborative environment between humans and AI.

"AI in Albert isn't about replacing scientists; it's about empowering them," says Nick Talken of Albert Invent. "We help scientists get the same product to market twice as fast, or make it twice as good. That's a 4x improvement, which is a tremendous value unlocked."

The CMO of Owkin adds, "The great thing about AI is that it has the potential to uncover patterns that humans could never find. For example, Owkin has developed an algorithm that can look at a scrap of tumor tissue and accurately predict which genes are being expressed and where. This will enable richer downstream analyses for researchers, and it allows researchers to discover new biology."

Investment Potential

Matt Ko, President of DeepScribe, emphasizes, "At the end of the day, the key word in healthcare is care, and machines cannot replace the human touch."AI offers substantial investment opportunities, particularly in high-cost, high-volume processes. For example, AI platforms addressing prior authorization—where traditional methods can cost up to $100 per request—yield significant returns by saving time and reducing costs. As Ko notes, "DeepScribe has increased the number of billed diagnoses by 36%, leading to massive financial outcomes for provider organizations while improving patient care."

Looking ahead, Ko predicts, "In the next 3-5 years, we believe that 90% of patient conversations will be recorded, presenting a unique opportunity for ambient companies to leverage that data to deliver point-of-care insights that enhance patient outcomes."

Companies like BenchSci also demonstrate significant ROI potential. Their ASCEND platform has shown a 40% time reduction in R&D portfolio performance for top customers, with Novartis saving over $14 million between 2018 and 2021 by deploying BenchSci's technology.

As AI evolves and integrates more deeply into healthcare, it promises to drive long-term value by addressing inefficiencies, enhancing patient care, and accelerating medical breakthroughs. For investors, healthcare AI represents a frontier where technology meets one of humanity's most critical needs, offering significant returns and meaningful impacts on global health outcomes.

The convergence of AI and healthcare is not just transforming an industry; it's reshaping the future of human health and well-being. As we stand on the brink of this new era, the opportunities for investors to benefit from this revolutionary change are both exciting and profound.

AI in Education: Revolutionizing a $6 Trillion Industry

The global education industry, valued at over $6 trillion, stands on the brink of an AI-driven transformation. Despite its critical role in societal progress, education faces numerous challenges impeding its efficiency and effectiveness. Artificial intelligence offers promising solutions to these longstanding issues, opening new avenues for personalized learning, administrative efficiency, and improved accessibility.

Personalized Learning: Tailoring Education to Individual Needs

Traditional education systems often follow a one-size-fits-all approach, which fails to to address the diverse needs and learning paces of students.  AI-driven personalized learning platforms are breaking this mold by customizing educational content to match individual learning styles and requirements.

Tom Sayer, CEO of Ello (an AI read-along companion), emphasizes this shift: "AI has the potential to offer an experience perfectly tailored to each child. Rather than teaching to the average, each student can have their own personal teacher who can provide support for exactly where they are on their learning journey."

Mark Angel, CEO of Amira Learning (an AI-driven reading instruction platform), explains: "Amira 'morphs' into a 1:1 personal tutor by constantly learning how a student learns and then adjusting the approach to instruction accordingly. Amira's uses Reinforcement Learning to evaluate the audio stream, the word being read, the history of previous mistakes, and the student's reading 'profile' to make an in-the-moment judgment about how to aid that student best."

Other startups are making significant strides in personalized learning. Studdy (an AI-powered study assistant) leverages generative AI to provide personalized study plans and instant feedback, resulting in a 25% increase in assignment completion rates at one university. Memrise (an AI-enhanced language learning platform) employs generative AI for language learning, leading to a 40% increase in vocabulary retention for Spanish learners after three months of daily practice.

Angel further adds, "As doctors use the MRI to see inside the body as their 'assistant' in diagnosis, Amira is an MRI for the reading brain, enabling teachers to be enriched as other knowledge workers have been by AI."

Administrative Efficiency: Reducing Educator Burnout

Administrative inefficiency is a major contributor to teacher burnout, with over 44% of U.S. K-12 teachers reporting frequent burnout. AI can alleviate this burden by automating tasks such as grading, scheduling, and resource management.

Platforms like MagicSchool (an AI-powered teacher assistant used by over 2.5 million educators) and SchoolAI (an AI integration platform for schools) are leading this charge. MagicSchool's AI assistant, used by over 2.5 million educators, directly addresses burnout: "Save time, fight burnout, & promote sustainability," while SchoolAI has helped over 2,000 schools integrate AI tools that simplify classroom management.

Accessibility and Inclusivity: Bridging Educational Gaps

AI is also helping to bridge the accessibility gap in education, providing remote learning solutions and adaptive technologies to reach underserved regions and students with diverse needs.

Tom Sayer of Ello shared a touching example: "In literacy, one extra benefit of Ello for students with learning challenges is the safe and non-judgmental space it provides to practice reading. One heartwarming story that shows this involves a child with selective mutism who - unbeknownst to their parents - picked up Ello, which had been purchased for a sibling, and began reading out loud."

Other companies like Riid (an AI-powered test preparation platform) are using generative AI to create customized test preparation materials, leading to significant improvements in standardized test scores.

Impact on Jobs

We spoke with Rajen Sheth, CEO of Kyron Learning (an AI video teaching platform). Before founding Kyron Learning, Rajen was VP, Google Cloud AI & Industry Solutions: “I see teachers becoming the orchestrator of student learning and being able to use a variety of tools to help them do their job.  This frees the teachers to have more time to spend individually with students and drive their learning.  I see students being able to learn the material that they need at the pace that they want in the style that is best for them.  AI is one significant part of this picture, but it relies on working closely with teachers to use AI to solve real problems.”

The integration of AI in education will likely lead to job transformation rather than wholesale replacement. In the short term, AI will augment the capabilities of educators and administrators, potentially displacing some roles focused on routine tasks. However, over the long term, AI is expected to empower educators with advanced tools for curriculum development, predictive analytics, and personalized professional development.

This shift will create new roles such as AI-assisted instructional designers, educational data scientists, and AI ethics specialists in education. These professionals will work at the intersection of pedagogy and technology, ensuring that AI tools are effectively and ethically integrated into educational settings.

Investment Potential

The education technology sector presents compelling investment opportunities, as evidenced by the success of companies like Dreambox Learning (an adaptive learning platform) and Osmo (a hands-on learning system with AI integration). These platforms have demonstrated significant improvements in student performance across various subjects.

Addressing concerns about the cost of AI implementation, Tom Sayer of Ello offers an optimistic perspective: "Right now, the way Ello runs, the cost of supporting a learner using the AI we use is easily outweighed by a relatively affordable monthly subscription payment. Investing in the people and development to build the systems is expensive, but running the Ello app is quite affordable in its current state."

Mark Angel of Amira Learning adds, "As AI tutors approach and surpass the goodness of human tutors there is an order of magnitude potential for us to increase unit costs while still being an order of magnitude more affordable than human capital."

The market potential is substantial, with nearly 70% of 4th graders in the U.S. reading below grade level. This presents a significant opportunity for AI-driven solutions to improve educational outcomes at scale.

Rajen shared, “My view on this is that the cost of AI is going to keep going down as models become more efficient and infrastructure gets better.  Already, GPT4 is a fraction of the price that it was just a year ago. The value is extremely high because it augments the staff that organizations have.  In so many cases already, using AI is 1/10th of the cost of doing a task manually, and that will get better and better over time.”

As AI continues to evolve and integrate more deeply into educational systems, it promises to drive long-term value by addressing critical inefficiencies, enhancing personalized learning experiences, and improving accessibility. For investors, the EdTech AI sector represents a frontier where technology meets one of society's most fundamental needs, offering the potential for both significant returns and meaningful impact on global educational outcomes.