Bullet Thesis: The 15 ways Voice AI will transform the real economy

Hi, it’s Michael from firstminute capital. We are a $500m seed fund backed by 130 unicorn founders. In a previous life I was a foreign correspondent for the Financial Times for 10 years, and the founding editor of Sifted. Along with my brilliant colleagues, I invest in European pre-seed and seed software companies like Vocca, Scalera, Granola and ai.work.

—

Last week, I finally answered a call from a debt collection agency—the 27th attempt they had made to reach me. I'd ignored the previous 26 calls, dismissing them as spam or fraud from an unknown number. When I answered, I discovered, to my embarrassment, that I did indeed owe a debt: £16 to a software company from two years ago. As the agent patiently explained the situation, a realization struck me: this company had likely spent £50 worth of human resources chasing my trivial debt. This wasn't just inefficient—it was the perfect illustration of where voice AI could transform a business.

The revolution in voice AI isn't about solving technical problems anymore—that work is largely complete. In 2024-2025, we've witnessed remarkable breakthroughs: speech-to-text and text-to-speech technologies have reached near-human quality, large language models can maintain context through conversations, and AI voices now convey natural emotion and handle interruptions seamlessly. Even more importantly, costs have plummeted, with model prices dropping 60-90% (like OpenAI's GPT-4 real-time API price cut in December 2024).

Now we've reached the year of the application layer. Nearly every vertical SaaS founder I've spoken with in the firstminute portfolio is exploring how to incorporate voice AI into their products—for customer feedback, outbound sales, medical front desk operations, and countless other use cases. We have seen 50+ founders in the past six months pitching industry or function specific software products with voice at their core. This is a big theme in Europe as well as the US, with voice-focused startups comprised 22% of the latest Y Combinator batch, according to Cartesia.

This surge of interest reflects both technical advancements and market opportunity. Despite the ubiquity of texting, email, and social media, phone calls remain the dominant communication mode for most businesses. Industries like healthcare, legal services, home services, insurance, and logistics rely on phone-based communication to convey complex information, provide personalised service, handle high-value transactions, and address urgent needs.

Yet current voice communication is profoundly inefficient. Small businesses miss 62% of calls on average, losing valuable opportunities. Calls go to voicemail after hours, and humans can only handle one conversation at a time. What might business look like if human-level calls with better-than-human context were 1000x cheaper than real people?

Where Voice AI Creates Real Economic Value in 2025

Not all voice-based industries are ripe for AI disruption. The most promising opportunities, we believe, share key characteristics when it comes to phone calls:

• They're business-critical (people are willing to put up with an AI).

• Frequent (so it’s worth scaling).

• Somewhat structured (so the AI can learn and get really good).

• Currently expensive or inefficient when handled by humans (so it’s worth the cost of using AI).

At firstminute, we have come up with a list of 15 vertical use cases cases where we are active looking for voice based startups. If you are building in any of these, please do reach out to michael@firstminute.capital, lorcan@firstminute.capital and sam@firstminute.capital.

1. Debt Collection

Why It Works: Collection agencies make millions of outbound calls to delinquent borrowers (like me, lol ), following strict compliance guidelines. Voice remains crucial because many debtors ignore written notices but will answer phone calls. Real-time conversation enables collectors to negotiate payments and settlements that static letters or emails simply cannot achieve. Voice AI can also replicate highly structured dialogues while ensuring 100% adherence to regulatory scripts.

Market & ROI:
Debt collection is a $30bn+ global industry but traditional agents can only make a limited number of calls per day, and many accounts receive no human follow-up due to staffing costs. Voice AI enables agencies to dial 10× more debtors daily and in theory at least kee compliance with the rules like the Fair Debt Collection Practices Act (FDCPA) in the US and in the EU the Unfair Commercial Practices Directive (UCPD).

Companies building:

• CollectWise (YC F24): AI debt collectors outperforming humans by 2× on late-stage accounts.

• Domu (YC S24): Automates collections calls for banks like BBVA and BNP Paribas; scaled to $870K ARR.

• Kisht: Building voice AI agents for early- and mid-stage collections in the EU.

• GetBill: AI voice debt collection.

• Salient (YC S24): AI loan servicing for auto lenders.

• Kastle AI: Voice AI agents for mortgage servicing.

• Skit.ai: Automated and autonomous collections.

2. Healthcare Administrative Calls

Why It Works: Healthcare systems process billions of calls spanning both front-office patient interactions (appointments, refills, intake protocols) and back-office administrative calls (insurance verification, prior authorizations). Voice remains the primary medium for these tasks.

Market & ROI: Healthcare admin costs around $1 trillion a year in the US alone. AI-driven call handling could in theory reduce the admin burden, but also reduce wait times, minimize no-shows and save practices money on staffing. Consider prior authorizations: these calls routinely require 30-60 minutes on hold before reaching a human representative. An AI agent that automatically handles this process could save each nurse or billing specialist several hours daily.

Companies building:

• HelloPatient: AI answering and scheduling for medical clinics.

• Hippocratic AI: Building AI healthcare workers for non-diagnostic roles.

• Abridge: Voice scribe and patient assistant.

• Nabla: Voice scribe helping doctors (firstminute portfolio company!).

• Assort Health: Front desk and triage agents.

• Superdial: AI-powered healthcare answering service.

• Rada (YC W25): voice agents that call insurers on behalf of medical practices.

3. Restaurant & Drive-Thru Ordering

Why It Works: More of a US trend than a European one, but quick-service restaurants (QSRs) and drive-thrus rely on fast, efficient voice interactions to manage high-volume orders. Peak hours can overwhelm human staff, causing errors, long wait times, and missed upselling opportunities. Can you imagine telling your McDonalds order to a voice Ai bot? Yes of course!

Market & ROI: The US drive-thru market alone is more than $140bn annually, with some 43% of US fast-food orders are placed at drive-thrus. AI order-takers can process orders 30–50% faster than humans and operate 24/7.

Companies building:

• Presto Automation: Deployed voice AI order-takers at Checkers, Sonic, and Del Taco; went public in 2023.

• ConverseNow: AI voice ordering platform live in 1,200+ QSR locations across major brands.

• Slang.ai: AI phone agents for restaurants handling both reservations and takeout ordering.

• SoundHound: Providing AI and hybrid agent ordering for enterprise QSR brands.

4. Home Services

Why It Works: Home service businesses—plumbers, electricians, HVAC technicians—depend heavily on inbound phone calls for booking jobs. Customers typically call for urgent, high-value services and expect immediate answers. Voice AI acts as a 24/7 virtual receptionist, answering calls instantly, qualifying service requests, and scheduling appointments directly into service management platforms.

Market & ROI: The US home services market exceeds $600bn annually, with phone leads converting to revenue at 10–15× the rate of web leads. Studies also show that 37% of calls to service businesses are missed, and 41% of weekend calls go completely unanswered.

Companies building:

• Avoca (YC W24): AI phone agents tailored for HVAC, plumbing, and trades, integrated with ServiceTitan.

• Allo: AI answering service for appointment-driven businesses; focused on SMBs.

• Goodcall: Turnkey AI receptionist for small businesses, including contractors and local service providers.

• Sameday AI: AI sales agents designed specifically for the home services sector (plumbers, HVAC, electricians).

• Numa: AI answering platform that integrates with CRMs, used by car dealerships and increasingly by local service franchises.

5. Insurance

Why It Works: Insurance companies depend heavily on phone calls for claims reporting (First Notice of Loss), policy renewals, customer support, and lead qualification. Voice remains essential because customers often prefer speaking to a human—or something that sounds human—especially when dealing with stressful or high-stakes issues.

There are lots of challenges here. There are different licences for advice and non advice sales, and an AI needs to make sure not to overstep the bounds of its licence. Claims communication is also tightly regulated, which comes with heavy fines for mistakes. This could make insurers reluctant to adopt voice Ai for fear of these mistakes. Also at least in the US, insurance licences are state level so can’t have one for the whole US. Still, it’s an interesting area.

Market & ROI: The insurance sector is a $5 trillion global industry, with customer service and claims operations representing major cost centres. Typical insurers handle tens of millions of calls annually—with hold times, agent turnover, and regulatory risk driving up costs.

Companies building:

• Liberate AI: Voice agents for 24/7 claims processing, policy changes, and quote explanations.

• Infer (YC W24): AI voice agents qualifying inbound leads and early-stage insurance customer conversations.

• Inaza: AI voice assistants managing underwriting questions and claims intake.

• Hi Marley: AI communication platform for insurers integrating messaging and voice to improve claims interactions.

6. Real Estate & Property Management

Why It Works: Real estate agents and property managers rely heavily on rapid phone communication to qualify leads, answer property inquiries, and schedule viewings. Timing is crucial: leads contacted within minutes are far more likely to convert. Voice AI allows agencies and landlords to instantly engage prospects 24/7, answer FAQs, and schedule showings.

Market & ROI: Real estate generates $2.3 trillion in annual transaction volume in the US alone and most homebuyers work with the first agent who contacts them. Voice AI that responds within seconds could perhaps increase showing rates.

Companies building:

• EliseAI: Voice leasing assistants for property management, raised ~$50M.

• AppFolio (which built Lisa AI): Built-in AI leasing assistant inside one of the largest property management platforms.

• Retell AI: Provides custom voice layers for real estate brokerages and leasing teams.

7. Legal Client Intake

Why It Works: Law firms, especially in consumer-facing practice areas (injury, family law, immigration), depend on quick client intake to win business. Yet many small firms can’t afford 24/7 human reception. Voice AI conducts structured intake interviews, gathering essential case facts, evaluating fit, and escalating high-value cases automatically.

Market & ROI: The US legal services market is valued at $400 billion, but it's fragmented across 450,000+ law firms. Missing a client call can cost a firm $5,000–$50,000+ per case.

Companies building:

• MeetGabbi: Customized AI receptionists and intake agents for law firms.

• Hona: Specialized voice AI for managing legal intake and call triage.

• GoAnswer: Offering legal intake service

8. Logistics & Transportation

Why It Works: Freight brokers, dispatchers, and carriers spend huge amounts of time on the phone coordinating loads, tracking shipments, and confirming deliveries. Voice AI can automate booking confirmations, check calls, appointment scheduling, and payment status updates.

Market & ROI: The US trucking sector alone is worth over $700 billion in revenue per year. Freight brokers spend 30–40% of their working hours on repetitive phone tasks.

Companies building:

• Happy Robot: AI dispatch agents for freight brokers and carriers.

• Fleetworks: Load updates, payment follow-ups, and appointment coordination via voice.

• Retell AI: AI voice overlays for TMS (transportation management system) communication layers.

Source: A16z

9. Market Research & User Studies

Why It Works: Traditional market research relies on costly, slow, human-led interviews and surveys. AI voice agents can in theory simultaneously interview hundreds of participants, follow up with probing questions, and deliver higher-quality qualitative data—at a fraction of the time and price. More open-ended voice data can also yield richer consumer insights for product launches and marketing.

Companies building:

• Listen Labs: Raised $27M to power mass-scale AI-driven voice interviews for major brands (Google, Nestlé).

• Poly.ai: Offering voice survey capabilities alongside customer service agents.

10. Government Services

Why It Works: Government agencies field enormous citizen communication loads—from tax questions to permit requests and service complaints. Traditional human-staffed call centres are overwhelmed, with long hold times and missed calls. Voice AI potentially offers a scalable way to handle frontline interactions and triage inquiries efficiently.

Market & ROI: The US federal government alone receives hundreds of millions of citizen calls annually (e.g., IRS handled 173M in 2022, but only answered ~13%).

Companies building:

• Boost.ai: Public sector conversational AI deployments.

• Pryon: Developing voice and knowledge management layers for public sector use cases.

11. EdTech: Language Learning & Skills Training

Why It Works: Language learning and professional training depend critically on real-time conversational practice—something legacy apps can’t simulate. Voice AI enables dynamic, interactive, on-demand speaking exercises.

Companies building:

• Speak: English-speaking practice app valued at $1B after 2024 raise.

• Praktika.ai: Ai avatars for learning english.

• ELSA Speak: Pronunciation and accent correction using real-time AI feedback.

• Second Nature AI: Sales coaching through AI-driven roleplay and voice simulations.

• Duolingo: Premium subscription with AI voice conversation features.

• Khan Academy (Khanmigo): Voice-based tutoring features for learning support.

• Sonia AI: Voice AI coach for professional skills development and career training.

12. Hospitality

Why It Works: Hotels, resorts, and vacation rental operators handle thousands of phone-based guest inquiries—everything from booking confirmations and concierge requests to room service and late check-ins. Voice AI could automate many of these high-frequency, low-complexity conversations. Unlike webforms or email, guests expect and prefer real-time service via voice.

Companies building:

• HostAI: Omnichannel assistant for hotels and short-term rentals, handling booking, check-in, and service calls.

• Nowadays: AI event planning and hotel scheduling assistant.

• EliseAI: Initially focused on real estate, now expanded to hospitality CRM and voice support.

• HiJiffy: Specialized AI voice and chat concierge for hotels in Europe and LATAM.

• Regal.io: Voice-based upsell and rebooking systems used by hospitality brands.

—

As well as these vertical use cases, here are three horizontal cross-function ones as well:

1. Outbound Sales & Lead Qualification

Why It Works: Sales development representatives (SDRs) spend enormous time cold calling, chasing leads, and following up. Voice matters: real-time conversation dramatically increases connection and conversion rates compared to email or text. AI removes the tedious parts while preserving authentic engagement.

Market & ROI: Outbound calling represents an estimated $85B market in the U.S. alone. AI voice agents can manage 3–5× more daily call volume than humans, with consistent follow-up cadences.

Companies building:

• telli (YC F24): AI phone agents that convert

• 11x (YC W24): Fully autonomous AI SDRs.

• Nooks: AI voice outreach + dialing platform.

• Artisan AI: Voice-first AI sellers.

• Hyperbound: AI role-play training to sharpen sales skills.

• Sameday AI: AI sales agents tailored for SMB outbound.

• VoiceGenie: Generative AI-powered voice bot that reaches out, understands, & responds to your buyers empathetically.

2. High Volume Recruiting

Why It Works: Recruiters conduct tens of thousands of phone screens monthly. Voice remains essential because it allows real-time assessment of communication skills and enthusiasm—things resumes can’t capture. AI handles repetitive initial interviews at scale, freeing up recruiters for higher-value tasks. Also because many applicants really want jobs, they seem willing to put up with an AI recruiter.

Market & ROI: The staffing and recruitment industry generates over $650B annually. High-volume hiring (e.g., warehouses, customer service) is particularly suited: replacing human screeners with AI can halve time-to-hire and reduce early-stage costs by up to 80%.

Companies building:

• Mercor (YC S22): AI voice interviews tailored for engineering candidates.

• Micro1 (YC W23): AI hiring for remote teams.

• Screenloop: Combines AI-driven interviews with predictive analytics.

• Talkpush: Specialized in volume hiring, especially in emerging markets.

3. Customer Support

Why It Works: Despite advances in chat and email, 70% of customer service interactions still involve voice, particularly for urgent or high-emotion cases. Voice AI now has the emotional range, interruption handling, and conversational memory to manage complex queries and route to humans when needed. Crucially, it also brings consistency—no more agent variation or dropped calls.

Market & ROI: Customer service is a $460B global market, with large enterprises spending $100M+ annually on call centre operations.

Companies building:

• Sierra (ex-OpenAI team): Full-stack customer experience automation platform with advanced voice capabilities.

• Decagon: AI agents tailored for customer support across verticals.

• Forethought: Voice-augmented AI support that handles tiered escalation and knowledge base queries.

• Nebesta: Ai native customer support BPO (from ex Ankostore founder)

• Parloa: German voice AI startup focused on multilingual customer support automation.

• Poly.ai: Leading voice AI player focused on Fortune 500 customer service deployments.

• Cartesia: Offers voice-native LLM agents for complex support workflows.

To sum up…

The telephone has been around for 147 years, but only now are we reinventing what it can do. Voice AI represents a massive opportunity to transform countless workflows that still rely on human-to-human calls.

—

If you are building in any of these spaces, above we’d love to hear from you. Do reach out at michael@firstminute.capital, sam@firstminute.capital or lorcan@firstminute.capital.