Voice AI Weekly #03 - 5th Week of July 2026
We have compiled five noteworthy news items from the global voice AI industry this week. A variety of developments took place within a single week, ranging from voice AI at the telecommunication provider's network level to on-device STT and agent platforms for contact centers.
1. DevRev Expands Organizational Memory with Voice AI for Customer Centers
On July 23rd, DevRev added voice AI functionality to its customer support platform, 'Computer'. It is designed to apply the organizational memory previously used in chat and email to real-time calls as well, allowing agents to reference customer records, tickets, order history, documents, and even codebases during consultations. They emphasized that the conversation context and authorization system remain seamless when transitioning to human agents.
DevRev stated that the focus of the voice AI competition has now shifted from 'natural conversation' to 'problem solving'. In a survey released by Gartner last March, only 20 percent of companies reduced their agent workforce through AI adoption. Ultimately, the ability to maintain context during calls can be seen as the deciding factor in the success of automation.
2. OpenAI Unveils 'Presence', an Enterprise Voice and Chat Agent Platform
On July 22nd, OpenAI introduced 'Presence', an enterprise voice and chat agent deployment platform. It supports GPT-Live-based real-time voice conversations and can be applied to customer support, sales, IT services, and HR tasks. When applied to their own English customer service line, they revealed that 75 percent of incoming calls were handled without human intervention. BBVA Mexico, SoftBank, and IAG's Retail Insurance Australia were named as early adopters.
It is also characterized by being released through a limited early access program rather than a self-service API. Given that they disclosed actual operational metrics for contact center automation rates, it is likely that 'what percentage was automatically processed' will become the standard comparison metric in the AICC market in the future.
3. Relatient Adds Clinical Request Automation to Healthcare Call Center Voice AI
On July 22nd, Relatient, a healthcare patient engagement platform, announced the addition of a feature to automatically handle clinical requests to its Dash Voice AI. It recognizes clinical inquiries, prescription refill requests, and billing questions left by patients over the phone using only voice and directly registers them as actionable tasks in existing workflow tools such as EHR (Electronic Health Record). The company explained that clinical requests make up 35 percent of calls received by hospitals, and organizations that process more than 1,000 voicemails per week can save over 300 hours of work time with this feature.
The approach of recognizing and classifying clinical inquiries to connect them to tasks, rather than just taking appointments, aligns with MAGO's approach, which does not stop at speech recognition but extends to intent analysis and report generation. The value of domain-specific STT is expected to grow, especially in healthcare call centers where specialized terminology is abundant and the risk of misclassification is high.
Source Relatient - Relatient Expands Dash Voice AI with Clinical Request Automation
4. Stenograph-Speechmatics Integrates On-Device STT into Court Reporting Software
On July 22nd, court reporting software provider Stenograph announced that it has directly integrated Speechmatics' on-device speech recognition engine into its program, CATalyst VP. This eliminates the need to run a separate speech recognition program and allows for the same performance in courtrooms or offline deposition sites without an internet connection. They also highlighted that sensitive testimony does not leave the reporter's computer.
The trend of demanding on-device STT in highly regulated and terminology-dependent domains like courtrooms and healthcare aligns with the domestic financial, healthcare, and public sectors. This can be seen as a signal that customers requiring both domain specialization and air-gapped deployment will increase.
5. SoundHound Enjoys Double Celebration with 52% Revenue Growth and Acquisition of LivePerson
SoundHound AI's revenue for the first quarter of 2026 reached $44.2 million, up 52 percent year-over-year. The previously announced acquisition of LivePerson is expected to close in the second half of 2026, and SoundHound expects this acquisition to add at least $100 million in revenue by 2027. The company's annual revenue outlook for 2026 is around $232 million.
The trend of conversational AI companies scaling up through contact center platform acquisitions is likely to be mirrored domestically as well. A strategy that sells not just standalone technology but channels and customer bases together is expected to determine the success of the competition.
If you are considering adopting Voice AI, please contact us at contact@holamago.com.
See you again next week in Voice AI Weekly.