AIMinds

Alex Levin, Co-Founder & CEO at Regal AI. Regal is the AI Agent Platform.

Generative AI Agents are transforming customer expectations and the types of customer experiences business can build. The biggest opportunity is in support, sales and operations calls at consumer businesses.

Episode Highlights:

- Alex challenges the long-standing “hide the phone number” mindset in contact centers. He argues that voice calls consistently outperform emails and texts in customer engagement, with far higher conversion and retention metrics when executed correctly.

- While many companies focus on building AI agents, Alex emphasizes that the real complexity lies in orchestration—pulling in customer data, handling routing, making API calls, and integrating with CRMs. Regal AI’s strength lies in building this full voice engagement stack.

- Regal’s AI agents now resolve ~97% of calls without needing a human handoff. This is a major leap from legacy benchmarks, which typically hover around 20–30%.

- Average handle time becomes less relevant with AI. Instead of optimizing for speed, Regal focuses on outcomes—like resolution, satisfaction, and conversion. Alex notes that AI often reduces call time by 30% anyway, while delivering better consistency and accuracy.

- Voice agents with personality outperform generic ones. Regal experiments with distinct agent styles—like playful, skeptical, or supportive—to mirror brand tone. Techniques like natural pauses, background noise, and tonal variation dramatically increase trust and engagement.

---------------------------------------
Connect with Alex Levin:
https://www.linkedin.com/in/alexlevin1/
https://www.regal.ai/

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Ilan Avner, Director of Product Management at AudioCodes. AudioCodes is a leading provider of advanced communications software, products and productivity solutions for the digital workplace. AudioCodes enables enterprises and service providers to build and operate all-IP voice networks for delivering unified communications, contact centers, and hosted business services, whether in the cloud or on premises.

Ilan Avner is a seasoned product leader with deep expertise in voice, telephony and  voice conversational AI. As Director of Product Management at AudioCodes, he drives the development of AI-powered voice solutions that connect contact centers with conversational platforms, enabling intelligent voice bots, live-agent assist tools, and seamless customer interactions. Ilan is passionate about transforming how businesses use voice to engage, automate, and scale customer communication.

Episode Highlights:
- Ilan emphasizes a “crawl, walk, run” approach for organizations exploring voice AI. Many start with call summarization due to its clear ROI and low friction—reducing agent wrap-up time from minutes to seconds and laying the groundwork for more complex use cases.

- While LLMs and transcription models are increasingly powerful, Ilan points out that telephony integration and voice orchestration remain significant hurdles. AudioCodes steps in to bridge the gap between voice systems, STT/TTS engines, and conversational platforms.

- Ilan highlights the real-world complexity of multilingual environments, such as Spanish mixed with English or code-switching in Hindi. These nuances often trip up transcription models, requiring tuning, provider selection, and sometimes even custom solutions.

- Generating speech isn’t just about good voices—numbers, jargon, and SSML usage remain pain points. Ilan notes that languages like Hebrew lack high-quality TTS, and fine-tuning output to sound natural and intelligible is still a manual, nuanced task.

- AudioCodes’ LiveHub platform enables teams to deploy a voice bot in minutes—connecting phone numbers, conversational AI platforms, STT/TTS providers (like Deepgram), and telephony systems, all from a single UI. It supports full orchestration, escalations, and integration with channels like Teams and WhatsApp.

---------------------------------------
Connect with Ilan Avner:
https://voiceaiconnect.audiocodes.com/
https://www.audiocodes.com/solutions-products/saas/audiocodes-live-hub

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Eliott Hoffenberg, Co-Founder & CEO at Vocca. Vocca helps medical clinics run 24/7 by automating their phone reception with AI.

It handles scheduling, triage, follow-ups, and patient prep through smart inbound and outbound calls. Thousands of providers use Vocca to reduce costs, save time, and deliver instant service—without the wait. 

Episode Highlights:
- Eliott explains how his company builds AI voice agents to automate inbound and outbound calls for clinics—handling scheduling, prep instructions, follow-ups, and more. Voice is still the dominant channel in healthcare, and automation addresses staff shortages and inefficiencies.

- Rather than aiming for overly chatty or expressive bots, the team focuses on clear, predictable behavior. Patients trust the agent not because it feels human, but because it’s reliable, accurate, and always available—key in high-stakes healthcare settings.

- Instead of building just a voice layer on top of other systems, Eliott’s team takes a vertical approach—owning everything from telephony to intent classification to EHR integration. This control lets them move faster, ship safer, and serve complex healthcare needs.

- Many patients don’t realize they’re talking to AI. Eliott stresses that determinism + thoughtful design create reliable, familiar-feeling conversations. Sometimes, less AI is better if it improves trust and clarity.

- Eliott sees healthcare as 50% diagnosis, 50% communication. His vision is to build 24/7 AI receptionists that manage everything outside of clinical care—empowering patients and streamlining operations without needing a doctor for every question.

---------------------------------------------------
Connect with Eliott Hoffenberg:
https://vocca.ai/
https://x.com/eliotthoff
https://www.linkedin.com/in/eliott-hoffenberg/

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Griffin Christenson, Founder at Gryfn.AI. Gryfn.AI offers AI-powered tools to help educators and parents identify signs of dyslexia in children through a fast, research-backed online screening test. The platform provides detailed insights into reading ability and potential learning difficulties, using speech and language processing to deliver personalized reports and recommendations.

As an optometrist and computer science enthusiast, Griffin Christenson, O.D., is merging clinical insight with tech innovation to reshape how we screen for dyslexia. With a background in biology and a doctorate from Pacific University, he’s developed an AI-powered MVP that digitizes and automates dyslexia risk detection—making early intervention more accessible. Now refining the project through global entrepreneurship training, Griffin’s mission is to fuse vision science with cutting-edge tools to improve care. Outside the clinic, he enjoys time with his wife, dog Lambeau, and cat Oliver.

Episode Highlights:
- Griffin shares his journey from becoming an optometrist to pursuing a master’s in computer science, driven by a fascination with AI sparked during board exam prep.

- Inspired by his father and Dr. John Griffin’s work in dyslexia diagnostics, Griffin digitized a paper-based screening test to identify dyslexia risks efficiently.

- The Gryfn AI screener presents words briefly on-screen for users to pronounce, using speech-to-text tech to assess accuracy and risk.

- Gryfn AI aims to open-source its code to expand access to early screening and partner with schools for automated testing.

- Griffin’s wife is launching Clarify Reading for remote dyslexia therapy, and he’s exploring tech-forward ventures in eye care and wearable devices.

---------------------------------------------------
Connect with Griffin Christenson:
https://gryfn.ai/
https://clarifyreading.com/

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Anurag Kejriwal, Founder at Sonnet AI. Sonnet AI, an advanced AI-powered meeting assistant & notetaker designed to keep conversations in context—helping professionals extract meaningful insights and manage relationships effortlessly in their specific domain– with custom text and visual notes. As a Computer Science & Engineering student at Ohio State University, Anurag is building an AI platform that goes beyond traditional note-taking, offering a domain-expert, no-bot-needed assistant that understands context and structure. Unlike other AI meeting tools, Sonnet AI doesn’t just capture words—it provides speaker insights, smart summaries, and an inbuilt relationship tracker, ensuring every conversation remains in context. With features like participant reports and conversation matching, Sonnet helps users keep track of critical discussions across teams, clients, and leads—all in one place.

Episode Highlights:
- Anurag shares his early entrepreneurial experiments—including dropshipping and cold-emailing Instagram theme pages—before eventually diving into the startup world with Sonnet AI while juggling college, a part-time internship, and classes as a CS and math major at Ohio State.

- Originally a sales-focused tool created by other co-founders, Anurag stepped in to reimagine the product with a focus on audio, relationship context, and founder productivity—leading to a new direction targeting startups and legal teams.

- Sonnet AI goes beyond simple meeting summaries by building a relationship tracker—a personalized CRM-like experience that tracks conversations across time and offers context-aware notes, adapting dynamically based on prior interactions.

- Templates, visual notes (like tables, concept maps, timelines), and “no-bot” audio capture are standout features—many of which were shaped by direct user feedback. One standout: users discovering creative ways to use the customizable templates Anurag hadn’t even imagined.

- From audio processing challenges to platform-specific constraints (like no Windows version yet), Anurag dives into technical hurdles and how AI coding assistants have become invaluable for a lean founder-led dev team—despite mixed reactions from academia on their use.

---------------------------------------------------
Connect with Anurag Kejriwal:
https://www.linkedin.com/in/anuragkejriwal1/
https://www.sonnetai.com/
https://www.linkedin.com/company/sonnetai

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Daniel Mishler, Head of Automation at Deepgram. Deepgram is Voice AI platform for enterprise use cases – speech-to-text, text-to-speech, and speech-to-speech APIs for developers.

Episode Highlights:
- Daniel pivoted from naval architecture in the Coast Guard to business through an MBA, then jumped into tech by working on synthetic aperture radar satellites—despite knowing nothing about them at first. His curiosity and growth mindset drove his success.

- Daniel reframes automation as augmentation, emphasizing AI should empower people, not replace them. His goal at Deepgram as Head of Automation is to make the company the most productive and efficient using AI tools, while being realistic about the cultural and technical challenges.

- Daniel credits tools like Cursor and Windsurf with changing his life. Despite no formal coding background and 10 kids at home, he used AI agents to start building. He stresses the importance of starting with a clear PRD and iterating through conversation-first workflows—a process he calls Vibe Coding.

- He built a VS Code extension called Vibe Coder that lets users talk through their ideas, automatically generate a PRD, and then dictate code using Deepgram's transcription tech. He sees this as a glimpse of the future—voice-driven development in AR, smartwatches, and beyond.

- Daniel recommends applying a VC-style portfolio strategy to AI automation: invest in many small, fast experiments, learn from each, and double down on what works. This mindset helps overcome fear, gain buy-in, and keep pace with the breakneck speed of AI innovation.

---------------------------------------------------
Connect with Daniel Mishler:
https://www.linkedin.com/in/daniel-mishler/
https://deepgram.com/

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Brandon Hawkins, Principal Product Manager at Twilio. Twilio is the world’s leading Customer Engagement Platform (CEP), powering direct, personalized, and intelligent customer experiences across every communication channel. Trusted by developers and enterprises in over 180 countries, Twilio helps companies bring AI, data, and security into every touchpoint—from sales and marketing to customer service and growth. With millions of developers building on their platform, Twilio continues to drive innovation in conversational AI and real-time communications, enabling businesses to create magical, scalable interactions with their customers.

Episode Highlights:
– Brandon shares his journey from software engineering to product management, driven by a desire to solve broader, customer-centric problems. He cut his teeth at Story blocks, learning to prioritize under the pressure of rapid growth and product-market fit.

– He reflects on his transition from startups to working at Twilio, highlighting the key difference: in large companies, influence replaces authority. Communication and documentation—like Twilio’s "write it down" culture—are essential for leading cross-functional teams.

– Brandon discusses Twilio's strategic position at the intersection of data, AI, and communications. With products like Conversation Relay and AI Assistance, they offer tools for developers and enterprises to build real-time conversational AI experiences.

– A major technical challenge in conversational AI is handling real-time voice interactions, where latency, interruption management, and voice quality are critical. Brandon notes the emergence of new UX patterns, like “interstitials,” that reflect natural human behaviors.

– He explains Twilio’s two-pillar strategy: building virtual agents to fully automate interactions and human-in-the-loop systems to enhance live agents with AI. Observability and transparency across AI-driven interactions remain a foundational focus for the company.

---------------------------------------------------
Connect with Brandon Hawkins:
https://www.linkedin.com/in/brandonjhawkins/
https://www.twilio.com/en-us

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Casey Traina, CO-Founder & CTO at Overlap AI. Overlap AI is building AI agents that can search, edit, and understand video content for media companies, podcasters , and creators. Their first agent transforms video marketing and can transform long videos into short form content, ready for distribution, and in your specific style. They were part of Y Combinator in Summer 24 and closed their seed round in November of 2024.

Episode Hightlights:
– Casey shares how he left Stanford after being inspired by a talk from Sam Altman, choosing to jump into the AI startup world rather than wait four years to graduate.

– Initially, Overlap AI aimed to create a "TikTok for podcasts," but after podcasters kept asking for clips, they realized the real demand was for automated short-form video creation.

– Overlap AI tracks the performance of posts, learns from their success, and continuously improves future clips to maximize engagement and virality.

– Casey explains why video is still an unsolved challenge in AI, noting that even leading models struggle to grasp context and meaning in video content.

– Overlap AI's roadmap includes enhancing production quality to match senior-level video editors, incorporating adaptive subtitles, color grading, and more advanced visual effects.

-------------------------------------------
Connect with Casey Traina:
https://www.linkedin.com/in/caseytraina/
https://x.com/caseytraina
https://overlap.ai
https://www.linkedin.com/company/joinoverlap/
https://x.com/Overlap_AI

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Mac Klinkachorn, Co-Founder & President at Trellis. Trellis automates manual PDF tasks at scale with AI agents. Trellis integrates end-to-end with your systems of record and allows users to define infinitely complex workflows to capture their business processes.

Episode Highlights:
– Mac Klinkachorn shares how he founded his first company at 15, using physics to detect water leaks, and how that experience shaped his problem-solving mindset.

– Mac discusses his transition from pursuing physics at Stanford to exploring AI and machine learning, realizing the potential of language models in real-world applications.

– Mac explains how he and his co-founder identified a massive bottleneck in handling unstructured data, leading to the creation of Trellis, an AI-powered solution for automating PDF workflows at scale.

– The conversation dives into the complexities of PDFs in industries like healthcare and finance, where non-standardized formats make automation difficult, and how Trellis overcomes these issues with AI.

– Mac reflects on whether improving AI models will eventually make tools like Trellis obsolete and explains why workflow automation and validation are key challenges beyond just data extraction.

-------------------------------------------------
Connect with Mac Klinkachorn:
https://www.linkedin.com/in/macklink/
https://runtrellis.com/

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4

Shraey Bhatia, Co-Founder & Chief Scientist at Leaping AI, a pioneering company revolutionizing enterprise businesses, call centers, and sales with self-improving Voice AI. Founded in Germany in August 2023, Leaping AI is part of Y Combinator's prestigious W25 batch, driving the future of AI-powered communication.

Episode Highlights:
- Shraey Bhatia transitioned from academia to startups after completing his PhD. His research initially focused on fact-checking and hallucinations in AI. Seeking new challenges, he co-founded Leaping AI, a voice AI startup, after moving from Australia to Berlin.

- The startup started in August 2023 with a vision for voice AI in sales, but after feedback, they pivoted to customer support with upselling features. Their breakthrough came when a German enterprise took a chance on them, helping them prove the tech could work.

- Initially focused on German enterprises, Leaping AI faced long sales cycles. In contrast, the U.S. market had faster decision-making and a stronger appetite for AI innovation, making it a better fit for scaling the company.

- Leaping AI is not just a call automation tool—it continuously evaluates and refines its own performance. It improves itself through AI-driven feedback loops, optimizing customer interactions over time without human intervention.

- After realizing Germany’s AI regulations and funding environment were limiting, the team spent time in San Francisco, applied to Y Combinator, and got accepted into YC Winter 2025. Now, they are scaling Leaping AI globally while maintaining a presence in Germany.

--------------------------------------------
Connect with Shraey Bhatia:
https://www.linkedin.com/in/shraeybhatia/
https://www.leapingai.com/

Connect with Demetrios:
https://www.linkedin.com/in/dpbrinkm/

Connect with Deepgram:
https://deepgram.com/
https://www.linkedin.com/company/deepgram
https://twitter.com/deepgramai
https://www.facebook.com/deepgram/

Join the Deepgram Discord Server!
https://discord.com/invite/xWRaCDBtW4