Customers

Customer

PlayHT

Industry

Voice AI

Region

Global

Cloud Platforms

Amazon Web Services

PlayHT Revolutionizes Text-to-Speech with Secure, Real-Time On-Premise Solution Powered by Tensor9

Challenge

PlayHT, a leading provider of AI-powered text-to-speech (TTS) solutions, faced a barrier in customer acquisition. PlayHT's core service revolves around processing customer text data to generate realistic-sounding speech. 

  • Data Security: PlayHT’s potential customers may need to handle sensitive information, such as customer proprietary information or consumer personal information.  Many potential customers were hesitant to entrust their data to PlayHT's cloud-based platform due to security concerns. 

  • Latency: PlayHT’s potential customers needed extreme low latency for the data they needed PlayHT to process. This low latency need is what drives realistic-sounding speech interaction, crucial for conversational applications like AI-driven contact centers, and AI outbound sales reps.

We were experiencing really long and difficult sales cycles with potential clients that needed real-time, private speech generation.
— Mahmoud Felfel, CEO of PlayHT


Solution

PlayHT partnered with Tensor9 to develop a groundbreaking solution: PlayHT On-Premise. Launched in December 2023, this offering leverages Tensor9's Instant Private SaaS technology to securely deploy PlayHT's text-to-speech engine within a customer's Bring Your Own Cloud (BYOC) environment with cloud native technology.

  • Blazing-Fast Performance: By running locally within the customer's environment, PlayHT On-Premise delivers the fastest text-to-speech processing on the market (<150ms), eliminating latency issues and enabling seamless AI-to-human voice interactions.

  • Real-Time Capabilities: The on-premise deployment empowers real-time speech generation, ideal for conversational applications, like AI contact centers, where responsiveness is paramount.

  • Enhanced Data Security: PlayHT’s customer data never leaves the customer’s control, addressing security concerns and opening doors to new markets.

  • Minimal Code Changes: PlayHT's developers benefitted from a smooth transition with minimal codebase modifications required for the on-premise solution.

  • Simplified Onboarding: Customers enjoy a streamlined onboarding experience, typically completed within the same day.

PlayHT partnership with Tensor9 has been a resounding success:

  • Rapid Customer Acquisition: Within a few months, PlayHT signed over a dozen customers, many in real-time conversational AI markets like automated customer service, AI avatars, gaming, employee training, and conversational AI platforms including those operating in healthcare and banking spaces.

  • Market Expansion: Tensor9's technology enabled PlayHT to enter new markets with stricter low latency and data security requirements.

  • Competitive Edge: PlayHT's on-premise solution offers best-in-class latency and data security, giving them a significant edge in the text-to-speech industry.


Tensor9 enabled PlayHT to enter new markets: real-time and private

PlayHT partnered with Tensor9 to create an on-premise offering for their SaaS product called “PlayHT On-Premise”.

Tensor9's innovative technology allows PlayHT's text-to-speech SaaS stack to deploy and run securely within a customer's Bring Your Own Cloud (BYOC) environment enabling a fully private on-premise SaaS. This enabled PlayHT’s customers to keep their data in their own environment and PlayHT to continuously deploy their infrastructure and application code to the customer’s own environment. Furthermore, by physically running closer to the customer's data, Tensor9 eliminated the need to transfer information across vast distances, significantly reducing latency, making PlayHT’s On-Premise solution the fastest text-to-speech provider on the market.

PlayHT On-Premise launched in December 2023. In just a few months, PlayHT signed over a dozen customers, many in the real-time conversational AI space (e.g. automated customer service, AI avatars, gaming, employee training, conversational AI platforms), and some with challenging security requirements (e.g. healthcare and banking). 

Today, PlayHT gets multiple leads a week with the new product offering due to its best-in-class latency and data security benefits. Plus, customers know that the text they process with PlayHT On-Premise is absolutely not trained for anything because PlayHT doesn’t even have access to it.

Tensor9’s was a game-changer for us. It allows us to offer our services to a wider range of customers, especially those in low latency real-time speech driven industries like AI Contact Centers and Scheduling where one hiccup can seriously impact the AI-to-human experience.
— Mahmoud Felfel, CEO of PlayHT

PlayHT launched the new “PlayHT On-Premise” solution with speed, stability and scalability

When PlayHT’s new on-premise offering launched in December 2023, PlayHT had spent about three months getting ready, including the code, sales, marketing, customer support, and updated contracts.

No matter what, a new product feature (deployment model) was launching so we had to make changes; but with Tensor9, the PlayHT lift was much smaller than if PlayHT had built the same capability themselves.

  • Application code barely changed and developers did not have to change their behavior whatsoever. 

  • Infrastructure code improved around auto-scaling algorithms and our health checks because it was manually handled in PlayHT’s traditional SaaS deployment. With on-premise SaaS, PlayHT’s previous operating approach needed to change with BYOC customers so they automated it using Tensor9 as a forcing function to further mature.

Our service is built on high performance and trust, and we needed a way to address these low latency and data security concerns without creating an on-premise solution by fully containerizing our product and losing the flexibility of what cloud native services offer. With Tensor9, our customers not only gained full data security and privacy in the text they needed to convert into speech but also gained the best performing latency on the market because our code is sitting right next to their code with compute owned by and dedicated to that one customer. This incredibly low latency enables our customers that need real-time speech generation.
— Mahmoud Felfel, CEO of PlayHT

PlayHT On-Premise customers enjoy a smooth and fast onboarding experience

PlayHT’s on-premise SaaS customers onboard in the same day, usually with just a few hours to get setup. Their onboarding is the same as the PlayHT traditional SaaS offering but customers interact with their own URL to their own domain. Customer installations starts with a setup IAM that is deleted after provisioning and customers can ingest their own logs for monitor. Auto-scaling happens automatically via our connected controls plane but customers can also choose to override if they know when they have peak demand surges to ensure capacity is available.

Managed at scale by popular SaaS vendors

CareTalkHeath Logo (CareTalkHealth.com)
PlayHT Logo (play.ht)

Call Center Autopilot

Voice AI

Trusted by businesses worldwide

Design Develop Now (designdevelopnow.com) Logo
Vocode.dev Logo

AI Telephony

Enterprise Software

AI Video APIs

AI Support

Speechmatics Logo (Speechmatics.com)

Speech Technologies

Digital AI Workers