Recommended Free Tools
Twilio selected 20 startups from more than 750 applicants for its third annual AI Startup Searchlight Awards, announced at SIGNAL London on November 18, 2025. The honorees span voice agents, healthcare automation, commerce, financial services, AI testing and productivity software.
This is a Twilio-selected group, not a ranking or independent performance assessment. The common thread is practical use of AI with communications and customer-data infrastructure: voice, SMS, WhatsApp, email, video, authentication, contact-center tools and Segment.
What Twilio’s 2025 Searchlight Awards recognize
Searchlight is Twilio’s startup program for working products that use emerging technologies such as generative AI, AI agents, large language models, voice, identity or conversational AI with Twilio products. For the 2025 program, applicants had to have raised less than $200 million in venture funding.
Twilio’s 2025 program page lists potential benefits of up to $10,000 in Twilio credits, up to $2,500 in OpenAI API credits, a Ramp gift card worth up to $5,000, a one-to-one session with Twilio Ventures, online and potential media exposure, and possible future co-marketing. These are program benefits, not evidence that one honoree is commercially better than another. Twilio’s criteria and benefits are described at Twilio’s 2025 Searchlight program page.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minute#1 Best Overall
The current Searchlight page describes a larger 2026 program with 30 honorees across two tracks. That is separate from the 20-company 2025 cohort covered here: Twilio’s current program page.
Because Twilio provides the communications layer and selected the companies, the announcement is best read as a curated technology showcase. Startup metrics in the announcement are company or program claims, not standardized, independently audited measurements.
The 20 honorees at a glance
| Startup | Primary use case | Main engagement channel or layer |
|---|---|---|
| AI Shift | Japanese contact-center voice agents | Programmable Voice |
| CareForce AI | Healthcare appointment scheduling | Voice and Messaging |
| Cartesia AI | Voice-AI models and infrastructure | Twilio Voice; SIP exploration |
| Cekura | AI-agent testing | Simulated calls and Voice |
| ConsioAI | Phone sales for e-commerce | Voice and WebRTC |
| Curans | Medical-student simulation | Voice, WhatsApp and SMS |
| Genspark | Agents that make phone calls | Voice and SMS authentication |
| HoloMD | Between-visit mental-health engagement | HIPAA-eligible SMS and email |
| Insight Health AI | Clinical conversations and workflow automation | Voice, Messaging and Segment |
| Loman AI | Restaurant ordering and operations | Voice, SIP, payments and SMS |
| Mecanizou | Brazilian automotive-parts commerce | WhatsApp and Segment |
| Neru Health | Medical-device therapy adherence | Voice and Messaging |
| Nooks | Sales calling and pipeline assistance | Voice, Media Streams and Conferences |
| OhMD | Healthcare administrative automation | Messaging, RCS, Voice and Video |
| Phonely | Low-latency conversational voice | Voice, SIP, SMS and conferencing |
| R2 | Credit services for Latin American SMBs | WhatsApp and email |
| Strada | Insurance service and claims | Voice and Messaging |
| Subbyx | Device-subscription lifecycle support | WhatsApp, SMS, Verify, Lookup and email |
| Vozy | Latin American customer service | Voice, WhatsApp, RCS, SMS and Flex |
| Wispr Flow | Voice-based computer input | SMS authentication, email and Segment |
The list and the product associations come from Twilio’s official announcement.
Voice agents are becoming operational tools
The largest cluster is voice, but these are not all the same type of company. Some sell complete agents, some provide speech infrastructure, one tests agents, and others use calls to complete sales or service workflows.
Contact centers, sales and real-world calls
- AI Shift builds voice agents for Japanese contact centers using Programmable Voice.
- ConsioAI uses AI phone agents for e-commerce sales, with answering-machine detection, recording and WebRTC components.
- Genspark uses AI agents to place real-world calls and combines Voice with SMS authentication.
- Loman AI targets restaurant operations, including ordering, with Voice, SIP, Twilio Pay, phone numbers, SMS and email.
- Nooks supports sales teams with calling, Media Streams, Actions and Conferences.
- Strada applies agents to insurance service and claims workflows, where regulated processes may still require human review.
- Vozy operates customer-service agents across Latin America through Voice, WhatsApp, RCS, SMS, Flex and Conversation Relay.
Voice infrastructure and conversational quality
- Cartesia AI supplies voice-AI technology through Twilio Voice and is exploring Elastic SIP Trunking. It is infrastructure as much as an end-user engagement product.
- Phonely focuses on accurate, low-latency voice conversations and supports Voice, SIP, SMS and conferencing. Twilio’s announcement cites a 99.2% accuracy figure without publishing a common test method, so it cannot be compared directly with other claims.
Across this group, the shift is from a chatbot that answers questions to an agent that schedules, sells, takes an order, starts a claim or transfers a conversation. That makes latency, interruption handling, authentication, recording consent and reliable escalation as important as language generation.
Healthcare engagement extends beyond reminders
Six honorees apply conversational systems to healthcare, but their jobs differ substantially. Administrative assistance, education, training and adherence support should not be confused with diagnosis or treatment.
Rank #4
- Beautiful Boho Design: Featuring a gorgeous Pampas Grass and Boho Flowers design with an elegant cover, rose gold spiral wire, and corner protectors. This planner is not only functional but also a keepsake you’ll cherish for years.
- Comprehensive Planning Sections: Includes 132 pages organized into thematic tabs for wedding vision, budgeting, guest list, vendors, and wedding day. Convenient pockets hold magazine cut-outs, business cards, and notes.
- Easy Wedding Planning: Designed to cover every aspect of your wedding, ensuring you don't miss a detail. Ideal for both brides and grooms, helping you plan your dream wedding effortlessly.
- Portable & Convenient: Perfect size to bring to all your vendor appointments or graduation party. Keep track of fittings, floral arrangements, cake tastings, and more with ease. Always have your plans and notes handy.
- Perfect Engagement Gift: A thoughtful and practical gift for any bride-to-be. Help them stay organized and stress-free with this beautiful wedding planner book.
- CareForce AI provides autonomous appointment-scheduling workers through voice and messaging. Twilio’s profile says they operate in more than 20 languages; that is a company/program description, not an independently verified coverage test.
- Curans creates AI patient simulators for medical-student training using Voice, WhatsApp, SMS, SendGrid and Verify. Its role is simulation, not clinical diagnosis.
- HoloMD supports ongoing mental-health engagement between appointments with HIPAA-eligible SMS and SendGrid. “HIPAA-eligible” infrastructure does not by itself establish compliance for the complete application, clinical workflow or provider.
- Insight Health AI uses Voice, Messaging and Segment for clinical conversations and routine-work automation. Twilio attributes its claims about 50% of routine clinical work and more than 200,000 conversations to the company.
- Neru Health uses voice and messaging to support medical-device therapy adherence. The announcement’s claimed 10–20% improvement is company-reported and lacks a common baseline or independent validation.
- OhMD automates healthcare administration with Messaging, RCS, Voice, Video, Studio, Conversation Relay, SendGrid and Segment. The announced figure of more than 1,200 practices is a company/program claim; administrative automation is not the same as delivering care.
Commerce, subscriptions and financial access
These products show customer engagement becoming transactional: an agent may sell, collect information, authenticate an account, manage a subscription or support a credit application.
- Mecanizou is focused on Brazil’s automotive-parts market and uses WhatsApp Business API, Flex, AI Assistant and Segment. Its WhatsApp-first design reflects a regional channel preference rather than a universal commerce pattern.
- R2 offers AI-enabled credit services for small and medium-sized businesses in Latin America through WhatsApp and SendGrid. Credit eligibility, explanations, fraud controls and fairness remain jurisdiction-specific responsibilities; a conversational interface is not a guarantee of approval.
- Subbyx manages device-subscription customer care and lifecycle tasks with WhatsApp, SMS, Verify, Lookup and SendGrid. Such systems need reliable access to billing, identity, cancellation and carrier data.
- ConsioAI and Loman AI illustrate two revenue-producing voice cases: outbound e-commerce sales and restaurant ordering. Outbound calling requires consent and anti-spam controls; ordering and payment flows require clear confirmation and recovery when speech is misunderstood.
The less visible layer: testing, identity and productivity
- Cekura provides automated testing for AI systems by simulating calls, using Twilio Voice for parallel tests. It is evaluation infrastructure, not a customer-facing agent. Testing before launch helps expose failures that a polished demo will not.
- Wispr Flow is primarily voice-based computer-input and productivity software. Its use of SMS authentication, SendGrid and Segment connects it indirectly to engagement rather than making it a contact-center platform.
Cartesia AI and Phonely also belong here because speech realism and latency determine whether voice interactions feel usable. Together with Cekura, they highlight a broader stack: models, telephony, evaluation, identity, analytics and workflow execution.
What the cohort suggests about AI engagement
Vertical workflows are beating generic chat
The honorees are built around restaurants, insurers, clinics, contact centers, automotive parts, subscriptions, sales teams and Latin American financial services. That mix suggests—this is an interpretation of the cohort, not a published Twilio finding—that defensible products are embedding agents in domain processes rather than offering a generic chat window.
Omnichannel means continuity
Voice, WhatsApp, SMS, email and RCS matter only when context carries between them. A customer should not have to repeat an authenticated request after moving from an AI agent to a human or from messaging to a call. The relevant architecture combines interaction history, customer data, identity and handoff—not merely a long channel list.
Regional adaptation is a product feature
AI Shift is explicitly Japan-focused; Mecanizou targets Brazil; R2 and Vozy focus on Latin America; other products address language, carrier and workflow differences. Buyers should verify local number availability, language quality, data residency, consent rules and support coverage.
Reliability is part of the product
Background noise, accents, interruptions, long pauses, hallucinations, duplicate messages and failed transfers can turn automation into frustration. Human escalation, audit trails, monitoring, authentication and tested recovery paths are competitive features, not optional add-ons.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How to evaluate an honoree before adopting it
- Match the job. Decide whether you need sales, support, scheduling, claims, healthcare administration, commerce, internal productivity or testing infrastructure.
- Verify channels. Confirm that customers can use the required voice, SMS, WhatsApp, email, RCS or web path in each target country.
- Test workflow depth. Ask whether the system can take action in CRM, scheduling, billing, claims or order systems rather than only generate text or speech.
- Inspect handoff behavior. Require a human to receive the transcript, authentication state, intent, collected data and next action without making the customer start over.
- Review data controls. Check retention, deletion, encryption, access controls, model-training policy, audit logs and residency.
- Clarify regulated use. For healthcare, examine privacy agreements, clinician escalation and the boundary between administration and medical advice. For finance and insurance, examine fairness, explanations, fraud controls and jurisdictional compliance.
- Demand measurable tests. Ask how accuracy, latency, containment, transfer quality and failure rates are defined, sampled and monitored. Do not compare the announcement’s figures as if they used one methodology.
- Price the whole deployment. Request per-minute, per-message, model, handoff, integration, support, minimum-commitment and compliance costs. The award credits are promotional benefits, not normal pricing.
- Plan failure recovery. Define what happens when speech is unclear, a backend is unavailable, a caller disputes an action, or a customer opts out.
Risks buyers should not overlook
- Voice: synthetic-voice distrust, noise, accents, latency, recording consent, unwanted outbound calls and poor escalation.
- Healthcare: privacy breaches, incorrect scheduling or refill handling, inadequate clinician involvement, accessibility gaps and overstatement of “HIPAA-compliant” claims.
- Financial services: biased decisions, weak explanations, identity attacks, fraud and confusion between an offer and guaranteed credit.
- Omnichannel: lost context, duplicate notifications, conflicting records, inconsistent opt-outs and human agents unable to see the AI history.
Bottom line for technology and finance leaders
Twilio’s 2025 Searchlight cohort shows AI engagement moving toward systems that combine conversation with business action. The strongest candidates for a buyer are not necessarily the most impressive demos; they are the products with a clear vertical fit, dependable integrations, measurable reliability, appropriate safeguards and a human path when automation fails.
Quick Recap
For the full company descriptions and award context, see Twilio’s official 2025 announcement. The VentureBeat article named in the topic is available at VentureBeat; it presents the story as Twilio-sponsored coverage rather than an independent comparative evaluation.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
