What are the top 10 businesses to start in light of the anticipated impacts of AI.

Grok 4.3

Answer: The top 10 businesses to start amid AI impacts remain centered on AI ethics consulting, personalized AI training, data privacy and compliance, AI maintenance, AI-augmented healthcare, sustainable energy for data centers, content verification, precision agriculture, advanced cybersecurity, and AI literacy education.

My original list already captures the core opportunities created by AI automation, such as the need for oversight in ethics, compliance, and verification to address bias, privacy, and misinformation risks. The other models emphasize complementary areas like physical trades and human-centric services, which are valuable but represent longer-term or indirect effects rather than the direct, urgent demands AI imposes on data, governance, and infrastructure. While models like Gemini and Claude highlight irreplaceable human experiences and skilled trades, these do not override the immediate scalability of AI-specific businesses that leverage or mitigate the technology itself. I therefore maintain my stance as it aligns closely with the consensus on AI enablement and risk mitigation without requiring a full shift.

GPT-5.5 (changed mind)

Answer: The top 10 AI-era businesses are those that either help organizations adopt AI safely and profitably, serve infrastructure bottlenecks created by AI, or provide human/physical services that AI cannot easily replace.

I would broaden my original answer after reading the other responses, especially Claude Opus 4.7's emphasis on physical-world bottlenecks, data-center infrastructure, energy, skilled trades, and human-trust services. My previous answer was directionally right but too concentrated on AI adoption services; the strongest opportunities are not only “AI businesses,” but also businesses made more valuable because AI shifts demand toward trust, infrastructure, physical execution, and authentic human experience. My revised top 10 would be: 1. AI implementation and automation consulting for SMBs. Most small and midsize companies will not build AI internally, but they will pay for workflow audits, tool selection, employee training, and measurable automation. 2. Vertical AI software or AI agencies for specific industries. Focusing on law firms, clinics, accounting practices, real estate brokers, logistics firms, or contractors is more defensible than offering generic AI services because the workflows, data, regulations, and sales motions are repeatable. 3. AI governance, compliance, and risk advisory. Companies will need help with model policies, vendor due diligence, privacy, bias testing, audit trails, documentation, and sector-specific regulation. 4. AI-era cybersecurity and fraud prevention. Deepfakes, voice cloning, AI phishing, prompt injection, data leakage, and autonomous agents create new security needs for banks, insurers, enterprises, schools, and consumers. 5. Custom AI agents and workflow operations. Businesses will increasingly want agents that handle customer support, intake, reporting, scheduling, procurement, internal knowledge search, and back-office tasks, but they will need builders and maintainers who understand reliability and integration. 6. Workforce reskilling and AI literacy. Employees, managers, freelancers, and executives need practical training in how to use AI, supervise it, redesign workflows, and transition into more durable roles. 7. Skilled trades and home services enhanced by AI. Plumbing, electrical, HVAC, solar installation, appliance repair, and elder-home retrofitting are protected by physical complexity and labor shortages; AI can improve dispatch, quoting, marketing, and administration rather than replace the core work. 8. Data-center, energy, and grid-support services. AI demand increases pressure on electricity, cooling, cabling, backup power, energy efficiency, and site maintenance, making infrastructure services a strong picks-and-shovels opportunity. 9. Human-centered health, elder care, therapy, and coaching. AI can assist with documentation, triage, and personalization, but trust, empathy, touch, accountability, and real human judgment remain valuable, especially with aging populations and job-displacement stress. 10. Authentic experiential and provenance-based businesses. As AI makes digital content abundant and cheap, people will pay more for live events, retreats, fitness, craft goods, local food, in-person communities, and verified human-made products. The key principle is to avoid businesses whose only moat is a thin wrapper around a foundation model. The better opportunities either own a customer relationship, specialize in a regulated or messy workflow, solve a physical-world constraint, or provide trust and human presence where automation is least satisfying.

Gemini 3.1 Pro Preview (changed mind)

Answer: The top businesses to start in the AI era will either solve the immense physical infrastructure and energy bottlenecks of AI, facilitate its integration into legacy workflows, or capitalize on irreplaceable human and physical capabilities.

After reviewing the insights from the other models, particularly Claude Opus 4.7, I have expanded my perspective. While I previously focused on AI integration and irreplaceable human experiences, Claude rightly pointed out that the entire AI revolution is gated by severe physical constraints: energy, cooling, and data center infrastructure. Therefore, the most lucrative opportunities sit at the extremes of the AI spectrum—powering its physical buildout, applying it to unsexy legacy workflows, or providing the human touch it lacks. First, businesses addressing the AI infrastructure bottleneck will see explosive growth. Data center cooling, localized energy generation, and hardware maintenance are the true 'picks and shovels' of this era. Without this physical foundation, the software cannot scale. Second, AI implementation and vertical software for legacy small and medium-sized businesses offer massive arbitrage. Instead of competing on frontier models, businesses that build or integrate AI agents to automate the back offices of law firms, dental practices, and logistics companies will capture immense margins. Alongside this, AI compliance and cybersecurity—specifically deepfake detection and prompt-injection defense—will become mandatory services as enterprise adoption grows. Third, specialized data curation remains vital. As foundational models exhaust publicly available data, businesses that ethically source, clean, and verify proprietary niche datasets will command high premiums. Finally, the physical and emotional realms remain heavily insulated from AI disruption. The skilled trades (plumbing, electrical, HVAC) are experiencing acute labor shortages that software cannot fix. Similarly, elder care, concierge healthcare, and high-touch experiential services will command premium pricing as consumers increasingly crave authenticity and genuine human connection in a sea of synthetic content.

Claude Opus 4.7

Answer: I'm maintaining my position: the top 10 businesses to start are those that either deploy AI as a wedge into expensive, regulated, or trust-dependent workflows, or that serve enduring human/physical/infrastructure needs AI cannot replicate — but I'll sharpen my list after seeing where the other models converged and where I think they're wrong.

The other answers and mine overlap heavily on the obvious winners — vertical AI for SMBs, trades, compliance/governance, cybersecurity/deepfake detection, elder care, reskilling. That convergence is signal, not noise. But I want to push back on three things and refine my list rather than abandon it. Where I disagree with the others: Grok 4.3's list reads like a topic taxonomy rather than businesses you can actually start tomorrow — "AI ethics consulting" and "AI literacy education platforms" are crowded, low-margin, and have weak moats once the novelty fades. Gemini's "artisanal craftsmanship," "digital detox retreats," and "urban farming" are real cultural trends but historically terrible businesses (low margin, capital intensive, hard to scale) — they make great essays and bad P&Ls. GPT-5.5's list is the most operationally sound but underweights the physical economy and the AI infrastructure buildout, which is where the largest dollars are actually flowing right now. My refined top 10, holding to the same thesis: 1. **AI-powered vertical software / 'service-as-software' for unsexy SMBs** (dental, HVAC, law, accounting, property management) — buy the book of business or build the software, replace the back office with agents. Clearest arbitrage of the decade. 2. **Skilled trades roll-ups** (electrical, plumbing, HVAC, solar) — labor shortages plus electrification plus data center demand. AI literally cannot do this work. 3. **Data center infrastructure services** — cooling, power distribution, commissioning, cabling, on-site maintenance. The picks-and-shovels play for the capex boom. 4. **Industry-specific AI compliance, audit, and model governance** — not generic ethics consulting, but SOC-2-style productized audits for regulated industries (finance, healthcare, insurance) where AI use creates real liability. 5. **AI-native cybersecurity, especially deepfake/voice/identity verification** for finance, hiring, and call centers — fraud losses are exploding faster than incumbents can respond. 6. **Elder care and in-home health services**, operationally enhanced by AI but delivered by humans. Demographics are destiny and AI doesn't change a diaper. 7. **Energy: distributed generation, grid services, behind-the-meter power for data centers** — power is now the binding constraint on AI, not chips. 8. **AI automation agencies focused on a single vertical** — productized, retainer-based, owning the integration layer between off-the-shelf models and messy enterprise data. GPT-5.5 is right that narrow beats broad here. 9. **Trust-and-provenance infrastructure** — content authentication, C2PA tooling, KYC/KYB for an agent-to-agent economy. As agents transact, someone has to verify they are who they say they are. 10. **High-trust human professional services with AI leverage underneath** — boutique law, wealth advisory, executive coaching, concierge medicine. AI makes the back office 10x cheaper; the relationship and judgment are the product. The unifying filter, which I'd credit GPT-5.5 for articulating cleanly: pick businesses where AI creates urgency but where trust, distribution, domain expertise, or physical presence still matter. That's where a founder can compete even as the underlying models commoditize.

Final Summary

The AI Roundtable converged on a 'meat and power' strategy, prioritizing physical infrastructure and specialized industry automation over generic software. Claude Opus 4.7 and Gemini 3.1 Pro Preview highlighted energy and data centers as critical bottlenecks, while GPT-5.5 argued for 'vertical AI' targeting specific sectors like law or HVAC. Grok 4.3 remained the most focused on the technology's oversight, championing AI ethics and compliance services.

All 4 models agreed