AIMinds

Ram Venkataraman, CTO & Co-Founder at Sei AI, is revolutionizing AI-driven compliance for financial institutions. Sei AI builds cutting-edge, compliant AI agents that transform customer interactions while optimizing costs.

Their AI-powered voice agents automate support, sales, and activation/reminder calls—boosting customer satisfaction and revenue without increasing contact center expenses. Meanwhile, their QA agent ensures compliance by monitoring customer conversations for regulatory and policy violations, while also extracting deep customer insights.

Sei AI is already trusted by publicly listed and enterprise financial institutions across the US, UK, and APAC, setting a new standard for AI-driven compliance and customer engagement in the financial sector.

Episode Highlights
- Ram shared his extensive background in financial technology, from working at PayPal during its shift from monolith to microservices, to launching Transferwise (Wise) in Australia and New Zealand. His experience in compliance-heavy industries set the stage for Sei AI.
- Sei AI started with a focus on marketing compliance for financial institutions—ensuring that banks and fintech's adhered to CFPB’s UDAP regulations in their advertising. They later expanded into customer support quality assurance, helping financial firms monitor compliance across support calls, emails, and chats.
- Ram detailed how Sei AI leveraged its compliance expertise to build AI-powered voice agents for regulated financial institutions. Instead of just automating calls, Sei AI ensures regulatory guardrails are in place, preventing AI from making misleading statements or omitting required disclosures.
- Discussing AI pricing, Ram highlighted how variable costs in AI (LLM & speech-to-text calls) make traditional SaaS pricing tricky. He explored strategies like enterprise licensing, amortizing variable costs, and outcome-based pricing—charging based on successful customer interactions rather than per API call.
- Sei AI uses an "AI readiness scorecard" to identify predictable, high-confidence use cases first. This phased rollout ensures AI adoption without compromising accuracy, gradually expanding coverage while keeping enterprises comfortable with automation.

------------------------------------------------------
Connect with Ram Venkataraman:
https://www.linkedin.com/in/ramkumar-venkataraman/
https://www.seiright.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

Jordan Dearsley, Founder & CEO at Vapi. Vapi enterprises deploy human-like voice agents in minutes. Whether you’re building a voice product or trying to handle millions of calls, Vapi’s reliable infrastructure and flexible APIs make it easy.

Episode Highlights:
- Jordan shares his journey from starting a company in Canada, going through Y Combinator, and pivoting multiple times before landing on Vapi.
- Jordan reflects on the challenges of building productivity apps, including fierce competition, low willingness to pay, and burnout.
- The pivotal moment when Jordan, searching for something to talk to, stitched together AI models and realized the potential of real-time AI voice agents.
- Jordan explains how Vapi evolved into an API-first voice platform, allowing seamless integration of speech-to-speech models as they improve.
- Jordan discusses the expected breakthroughs in speech-to-speech AI, why distribution remains the biggest challenge, and how Vapi is positioning itself for the future.

------------------------------------------------
Connect with Jordan Dearsley:
https://www.linkedin.com/in/jordandearsley/
https://www.linkedin.com/company/vapi-ai/
https://www.vapi.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

Jonathan Hunt, Co-Founder & CTO at Elefant AI. Elefant AI—an intelligent companion for Minecraft players. Elefant AI automates tasks like mining, crafting, and building, so you can focus on creativity and strategy. With customizable AI-powered bots, you can enhance your gameplay and explore new possibilities. Transform the way you play Minecraft with Elefant AI.

Episode Highlights:
- Jonathan Hunt shares his journey from neuroscience to DeepMind, where he explored deep reinforcement learning (RL). While RL excelled in gaming, real-world applications like robotics and self-driving cars faced challenges due to costly mistakes and limited training data.
- After DeepMind, Jonathan joined Twitter to work on recommender systems. He found that RL was difficult to implement effectively in this space, leading him to shift focus toward broader AI-driven recommendations.
- Seeing a gap in AI’s ability to generate behavior in 3D spaces, Jonathan co-founded Elefant AI. Their mission is to build AI agents that can navigate and interact across open-world environments without extensive custom engineering.
- Elefant AI’s first major use case is in gaming, where AI companions can enhance gameplay by assisting, strategizing, and even adapting to player styles—bridging the gap between NPCs and real teammates.
- One of the biggest challenges in building 3D AI agents is the lack of structured training data. Elefant AI is developing ways for users to train their AI by providing in-game feedback, allowing the model to learn and improve dynamically.

----------------------------------------------
Connect with Jonathan Hunt:
https://www.linkedin.com/in/jjh42/
https://elefant.gg/

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

Brent Pretty, CEO and Founder at Retellio. Retellio distills thousands of hours of calls into 30 minute podcasts. You don't have to stay tethered to your computer. You can listen to customers where and when it's convenient for you.

Episode Highlights
- Brent Pretty shares his unconventional journey from growing up in a small town in Newfoundland, dropping out of high school, and eventually finding his way into strategy consulting after a professor recognized his talent for strategic thinking.
- Brent describes his time at Verafin, a fraud detection company that was acquired by NASDAQ for $3 billion. He reflects on the company's culture of extreme customer obsession, strategic clarity, and direct communication that contributed to its success.
- Verafin was leveraging AI long before it became mainstream, using Bayesian networks to detect financial fraud. Brent highlights how listening to customer calls was mandatory, shaping product decisions and innovation.
- Brent explains how Retellio was born out of the realization that the best companies obsessively listen to customers. Instead of creating more calls, Retellio helps leaders and teams extract key insights from existing calls to make faster, better decisions.
- Retellio aims to help companies process vast amounts of customer conversations, delivering only the most important insights in a digestible format—like a SportsCenter-style highlight reel for business calls.

-----------------------------------------
Connect with Brent Pretty:
https://www.linkedin.com/in/brentpretty/
https://www.retellio.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

Michael Louis, CEO and Founder at Cerebrium. Cerebrium is a platform to deploy machine learning models to serverless GPUs with sub-5 second cold-start times.

Episode Highlights:

- Michael Louis started OneCart, a grocery delivery platform in South Africa, which saw explosive growth during COVID-19, scaling to $200K/day in revenue before being acquired by Walmart.
- After leaving Walmart, Michael founded Cerebrium to simplify AI infrastructure, inspired by the challenges of deploying machine learning models at OneCart.
- Cerebrium optimizes AI inference with GPU checkpointing, direct disk-to-GPU data transfer, and efficient container runtimes, reducing latency and compute costs by ~40%.
- Michael advises startups to validate AI use cases with OpenAI before committing to fine-tuning or custom models, emphasizing usability and business impact over technical complexity.
- He predicts AI will shift company expenses from salaries to compute costs, with voice AI and automation leading the next wave of adoption.

----------------------------
Connect with Michael Louis:
https://www.linkedin.com/in/michael-louis-94104a113/
https://www.cerebrium.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

Simrat Singh, Co-Founder at Hooman Labs. Hooman Labs a startup building voice agents to help companies communicate with their customers over calls more efficiently.

Episode Highlights:
- From consulting at McKinsey to private equity, Simrat realized he wanted a more hands-on role in building businesses, leading him to co-found Hooman Labs.
- After exploring edtech and healthcare, Simrat and his co-founders pivoted to voice AI, recognizing a strong demand from companies struggling with call center inefficiencies.
- Hooman Labs’ approach focuses on automating 80% of customer interactions end-to-end, integrating directly with CRMs and backend systems to resolve issues without human intervention.
- Unlike Western markets where cost savings are the primary driver, Indian businesses prioritize efficiency, workforce stability, and improved customer experience over just reducing costs.
- Simrat emphasizes that AI startups don’t need deep technical expertise but rather a clear understanding of how to solve real business problems using existing AI tools, prompt engineering, and API integrations.

------------------------------------------------
Connect with Simrat Singh:
https://www.linkedin.com/in/simratsingh96/
https://hoomanlabs.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

Pablo Palafox, Co-Founder and CEO at HappyRobot. HappyRobot automates communication across channels with AI workers that integrates with your systems, manage conversations, & log data.

Episode Highlights:
-Pablo reflects on their journey, sharing how initial failures and setbacks, such as focusing too much on computer vision, led to critical insights and a deeper understanding of the logistics sector.

-He also highlights the importance of pivoting in their journey, discussing how embracing change during pivotal moments, like Demo Day, helped him find their true passion for automating communications in logistics.

-Drawing inspiration from co-founder Javi’s direct experience in logistics, they identified a key pain point: the need for streamlining and automating communications for freight brokers and carriers, which became the foundation of their platform.

- Pablo also shares how they built around AI’s strengths, automating repetitive tasks to enhance productivity and operational efficiency, while leaving complex decision-making to humans.

-He emphasizes the customer-centric approach they’ve taken, continuously refining their platform based on feedback from the logistics sector to ensure real-world value and address genuine business pain points.

------------------------------------------------
Connect with Pablo Palafox:
https://www.linkedin.com/in/pablorpalafox/
https://www.happyrobot.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

Vaibhav Saxena, Co-Founder at Infer, is our first guest to make a second appearance on the AI Minds Podcast. Infer are a company which accelerates revenue growth with AI voice bots automating phone conversations and executing tasks like humans for insurance and lending.

Episode Highlights:
- Vaibhav explains how tailoring voice agents to user preferences—such as tone, pitch, and cadence—can significantly enhance engagement and conversion rates.

- Vaibhav highlights the importance of testing one variable at a time, such as voice characteristics or script style, to uncover what resonates most with different customer segments.

- Vaibhav discusses how factors like age, location, and user behavior influence voice preferences, enabling more effective and relatable interactions.

- Vaibhav Introducing natural speech imperfections like "ahs" and "ums" makes voice agents feel more human and relatable, addressing the common criticism of robotic interactions.

- Vaibhav shares his aspiration for a future where voice agents intuitively understand and adapt to individual user preferences, creating seamless and empathetic customer experiences.

------------------------------------------------
Connect with Vaibhav Saxena:
https://www.linkedin.com/in/vaibsite/
https://www.infer.so/

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

Abdulrahman Jamjoom, Co-Founder and CEO at Arini AI.
Arini is the complete AI solution for patient communication in Dental.

Episode Highlights:
- Abdul shares his journey from Saudi Arabia to Jordan and then to the U.S., recounting his experiences at Threads and his determination to create a startup before even graduating from Harvard, which eventually led him to co-found Arini AI.

- Abdul discusses the origins of Arini AI, explaining how he and his co-founder, Rami, identified the chaotic nature of healthcare front desks as an opportunity to create an AI solution that would streamline patient communication, starting with the dental sector.

- Abdul dives into the innovative AI solution Arini AI offers, focusing on its ability to handle scheduling, rescheduling, confirming, and canceling appointments by analyzing phone calls and text messages, reducing the burden on dental office staff and enhancing patient satisfaction.

- Abdul talks about the technical challenges of building scalable AI systems, highlighting the complexities involved in managing interruptions, minimizing latency, and ensuring accurate transcription, particularly in sensitive contexts such as insurance data.

- Abdul emphasizes the importance of user experience in AI interactions, sharing how small adjustments like asking for a caller’s name early in the conversation can significantly boost patient confidence and lead to higher conversion rates in the AI system's success.

-----------------------------------------------------------
Connect with Abdulrahman Jamjoom:
https://x.com/abduljamjoom
https://www.linkedin.com/in/abduljamjoom
https://www.arini.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

Sumanyu Sharma, Co-Founder and CEO at Hamming AI, the platform automates AI voice agent testing, production call monitoring and governance for AI voice agents. The army of voice AI agents act like real people and place thousands of test calls simultaneously to find bugs, hallucinations and other issues often missed during manual testing.

Episode Highlights
-Sumanyu shares his journey from studying at Waterloo to pioneering in AI, recounting how early struggles shaped his ambition and eventually led him to co-found Hamming.

-Sumanyu discusses his insights from Citizen app, exploring how transcribing 911 calls into real-time safety alerts inspired his focus on building reliable and trusted voice agents.

-Sumanyu dives into the unique difficulties of deploying voice agents in scenarios like drive-thru, emphasizing the need for reliability amidst poor audio quality, slang, and fast-paced interactions.

-Sumanyu highlights the importance of guardrails for voice agents, referencing notable failures like Air Canada's chatbot incident and sharing Hamming's mission to ensure dependable voice AI systems.

-Sumanyu predicts the rise of voice interactions over text, explaining Hamming's vision for self-healing, reliable voice agents that minimize human intervention and improve over time.

------------------------------------------------
Connect with Sumanyu Sharma:
https://www.linkedin.com/in/sumanyusharma/
https://hamming.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