• NanoBits
  • Posts
  • Why every Industry we called "AI-Resistant" is now all-in (+ What's Coming Next)

Why every Industry we called "AI-Resistant" is now all-in (+ What's Coming Next)

The State of AI 2025 Report by Bessemer Venture Partners - Part 2

EDITOR’S NOTE

Hello Nanobits Readers,

My dentist called me last week for a routine checkup. Instead of the usual "hold please while I check our books," the receptionist said, "Let me ask our AI assistant when you are due for a cleaning."

My dentist. Using AI. In 2025. 😲

This caught me off guard because dentistry has always been the poster child for "traditional" industries that resist technology. If AI can transform how dental practices operate, what does that mean for every other industry we once considered "technophobic"?

Last week, we explored how AI is rewriting the rules of startup success and disrupting enterprise giants. This week, we will be diving into the vertical AI revolution, how AI is penetrating industries that were supposedly resistant to technology, transforming consumer behavior, and what Bessemer's report and data reveals about what's coming next.

Plus, we will reveal Bessemer’s five bold predictions for 2025-2026. Spoiler: One of them involves your browser becoming something entirely new (well, no surprises, people are loving Comet!) 🌐

THE VERTICAL AI PLAYBOOK

For years, we have dismissed certain industries as "technophobic", resistant to adopting new tools and stuck in old ways. Healthcare, legal, real estate, home services, these were the sectors where technology adoption moved at a glacial pace.

Plot twist: These industries weren't technophobic at all. Traditional SaaS just failed to solve their high-value, language-heavy, and multimodal tasks. AI is finally meeting these users where they are, with products that feel less like software and more like real leverage.

Bessemer's data shows that vertical AI adoption is accelerating fastest in industries previously considered resistant to technology. The pattern is clear: AI is solving problems that traditional software couldn't touch.

Every successful vertical AI company follows a similar pattern. Here's the winning formula:

Compelling Wedge: Start by solving a core pain point that's language-heavy or multimodal. The best wedges are intuitive and embed into existing workflows.

Context is Key: Build defensibility through domain expertise, deep integrations, and multimodal interfaces built for vertical-specific needs.

Built for Value: ROI is clear from day one. These tools unlock 10x productivity, not incremental improvements.

The Audio Pattern: Voice and audio processing appear repeatedly as the "miraculous wedge" across verticals. When AI can understand and act on spoken language, it transforms workflows that were impossible to automate before.

Industry Transformation in Action

Let's look at how AI is actually transforming these supposedly "resistant" industries:

Healthcare

The Problem: Massive documentation burden causing provider burnout

AI Solution:

  • Abridge automates clinical note-taking with generative AI

  • SmarterDx helps hospitals recover missed revenue through automated coding

  • OpenEvidence delivers instant medical literature answers at point of care

The Impact: Providers spend more time with patients, less time on paperwork

Legal

The Problem: Manual document review and research taking days or weeks

AI Solution:

  • EvenUp generates legal demand packages in minutes instead of days

  • Ivo automates contract review and natural language search

  • Legora accelerates legal research and collaborative drafting

The Impact: Trial attorneys can scale caseloads dramatically

Education

The Problem: Teachers drowning in administrative tasks

AI Solution:

The Impact: Teachers focus on teaching, not paperwork

Real Estate

The Problem: Labor-intensive property management workflows

AI Solution:

  • EliseAI automates prospect communications and lease audits

The Impact: Property managers handle more units with less effort

Home Services

The Problem: Inconsistent customer service and sales coaching

AI Solution:

  • Hatch acts as AI-powered customer service teams

  • Rilla analyzes in-person sales conversations and coaches reps at scale

The Impact: Better customer experience and higher conversion rates

The Common Thread: These aren't just "AI features" bolted onto existing software. They are AI-native solutions that reimagine entire workflows around what's now possible with language understanding and automation.

Three fundamental questions will determine vertical AI's future:

Today's vertical AI companies integrate with legacy systems, but as they mature, will they start competing directly with them instead?

  1. Legacy Systems: Will vertical AI extend existing systems or replace them with AI-native platforms?

  2. Incumbent Competition: Will established players' scale and distribution beat startup innovation?

  3. Data Moats: Can vertical AI companies maintain advantages when industry data is fragmented and privacy-sensitive?

These questions will define whether we see gradual evolution or complete disruption.

CONSUMER AI

While most of 2024 focused on AI as a productivity tool, 2025 is seeing a fundamental shift in how consumers interact with AI.

Consumers across age groups are increasingly turning to general-purpose LLMs, particularly ChatGPT and Gemini, for daily or weekly assistance (with an estimated 600M and 400M weekly active users as of March 2025, respectively.)

Beyond Productivity: The Emotional AI Shift

What began as "ChatGPT for writing emails" has evolved into something much deeper:

Mental Health & Wellness: AI journals like Rosebud and gamified companions like Finch help with self-reflection and habit building

Creation Democratization: Tools like Create.xyz, Suno, Runway, and Moonvalley are turning every consumer into a creator

Voice as the New Interface: Unlike Alexa/Siri, LLM-powered voice AI handles open-ended conversations and emotional support

Perplexity: The Search Revolution

Perplexity has emerged as the breakout consumer AI success story, fundamentally changing how we interact with information online. With its model-agnostic orchestration and blazing-fast UX, it's becoming the go-to for AI-native search.

The Big Shift: With the launch of Comet (Perplexity's agentic browser), we are seeing the evolution toward ambient, proactive agents that don't just answer questions, they anticipate needs.

Clear Unsolved Consumer Pain Points

The most obvious consumer AI opportunities remain untapped, not from lack of demand, but because current technology still requires too much manual user action.

Use Cases Waiting for Infrastructure:

  • Travel: Fragmented booking processes cry out for personalized, end-to-end concierge services

  • Shopping: E-commerce could be fundamentally reshaped when agents handle browsing, comparison, and checkout autonomously

BESSEMER’S 5 BOLD PREDICTIONS FOR 2025-2026

1. The Browser Becomes the Dominant Agentic Interface

Your browser won't just navigate the web, it'll become a programmable environment where AI agents observe and execute tasks across the entire digital world. Think of it as the new operating system for AI.

2. 2026: The Year of Generative Video

Just as 2024 was the breakout year for AI images and 2025 for voice AI, 2026 will be when generative video crosses the mainstream chasm. We are nearing the tipping point in controllability, accessibility, and realism.

Key Players to Watch:

  • Google's Veo 3: Setting benchmarks for video realism

  • OpenAI's Sora: Leading in creative applications

  • Moonvalley's Marey: Focusing on cinematic quality

  • Open-source models: Will they catch up like they did with images?

The next 12 months will clarify generative video's market structure across three key battles:

Lab Dominance vs. Open Innovation: Google's Veo 3 sets realism benchmarks while companies like Higgsfield prove you can build powerful products on existing models. Unlike images, video's computational intensity has limited open-source progress but Bessemer predicts strong open models will emerge in 2026.

Speed as a Moat: Teams like Lemonslice are experimenting with real-time video generation, where latency becomes a competitive advantage for livestreaming, virtual influencers, and gaming.

Key Use Cases to Watch:

  • Cinematic tools for creators and studios

  • Real-time generation for interactive applications

  • Photorealistic storytelling and virtual production

  • Personalized content and social identity

The IP Challenge: Growing copyright complexity means startups must proactively address licensing, training data sourcing, and creator royalties, not just for legal protection, but for long-term trust and differentiation.

3. Evals and Data Lineage: The Missing Infrastructure Layer

The biggest bottleneck in enterprise AI isn't the models, it's evaluation. Public benchmarks like MMLU don't reflect real world workflows, compliance needs, or business contexts. Most companies can't reliably assess if their AI actually works for their specific use cases.

The 2025-2026 Shift: AI evaluation will go private, grounded, and trusted—unlocking 10x enterprise deployment. Instead of chasing leaderboard scores, companies are building internal eval suites measuring real business metrics: accuracy, latency, hallucination rates, and customer satisfaction.

The New Evaluation Stack:

  • Private, use-case specific evals on proprietary data

  • Business grounded metrics integrated into production

  • Continuous eval pipelines with real-time feedback

  • Lineage and interpretability for regulated industries

Rising Infrastructure: Companies like Braintrust, LangChain, Bigspin.ai, and DataHub are building the evaluation and lineage tools that will become a strategic layer of the AI stack, not just development features, but procurement requirements.

As model performance converges, the differentiator won't be raw accuracy, it'll be knowing exactly how, when, and why your AI works in your specific environment.

4. A New AI-Native Social Media Giant Could Emerge

Major technology shifts historically create new social giants: PHP enabled Facebook, mobile cameras birthed Instagram, and mobile video propelled TikTok. Generative AI capabilities will likely spawn the next breakout social platform.

Potential Forms:

  • Networks where AI agents help us never miss birthdays, friend updates, or local events, making us our best selves online and offline

  • Platforms populated by emotionally intelligent AI influencers and AI clones (like Character.AI and Replika hint at)

  • Social spaces where AI, not humans, are the main characters

5. The Incumbents Strike Back: M&A Surge Ahead

After two years of AI-native startup disruption, enterprise giants are striking back by acquiring their way into the AI era. Expect a 2025-2026 M&A surge as incumbents buy capabilities they can't build fast enough.

The Battle Lines: Vertical software faces the starkest choice, evolve or become obsolete. As AI-native startups automate industry workflows (insurance claims, legal briefings, revenue management), traditional SaaS players must acquire to survive. High-service, regulated industries will see the most consolidation.

Beyond Features: This isn't about bolting on AI, it's structural reinvention. AI tools are becoming so embedded they resemble intelligent service providers, forcing incumbents to rethink their entire value proposition.

Infrastructure Play: Strategic acquisitions will also target model orchestration, evaluation, and memory systems as enterprises buy AI-native stack building blocks.

Founder Takeaways: Be ready for strategic interest, build strong technical moats, and understand where incumbents are falling behind. The giants are confirming through acquisitions that AI-native is the new standard

THE FOUNDER’S EDGE IN THE AI COSMOS

We are deep in AI's unfolding galaxies. Success isn't just about velocity, it's about vector, moving fast in the right direction. The most iconic companies won't just ride the AI wave; they'll shape it by aligning exponential capability with real-world clarity.

AI is now operational, generating revenue, building relationships, and rewriting industry rules. Yet much remains unresolved: memory, context, governance, agency. The map is fuzzy, but the frontier is real.

Top Takeaways for AI Builders & Professionals:

Two Winning Archetypes: Supernovas hit ~$100M ARR in 1.5 years (but with fragile retention), while Shooting Stars grow like stellar SaaS over 4 years with strong product-market fit.

Key Strategic Advantages:

  • Memory = Moat: Persistent memory and personalization create unbreakable lock-in

  • Systems of Action: Don't bolt AI onto legacy; reimagine entire workflows

  • Start with AI Wedges: Solve narrow, high-friction problems, deliver 10x value, then expand

  • Browser-First: Build for the browser, it's becoming the new operating layer for agentic AI

  • Private Evaluation: Build continuous eval infrastructure from day one

  • Speed Wins: Onboarding in hours, not months, collapses competitive moats

Market Realities: Vertical AI is the new SaaS, incumbents are acquisitive, and in a world of automation, human taste and judgment become the ultimate differentiators.

END NOTE

The companies winning today aren't just building faster software. They are designing systems that see, listen, reason, and act, embedding intelligence into the fabric of work and life.

The next 18 months will be crucial. Generative video will go mainstream. The browser will become an AI playground. New social platforms will emerge. And somewhere in a garage or dorm room, the next AI breakthrough is probably taking shape.

The question isn't whether AI will transform your industry, it's whether you'll be leading that transformation or watching from the sidelines.

Image Credits: CartoonStock

Share the love ❤️ Tell your friends!

If you liked our newsletter, share this link with your friends and request them to subscribe too.

Check out our website to get the latest updates in AI

Reply

or to participate.