Voices is expanding the way advertising and creative agencies can source voiceover, introducing a new service that places traditional professional voice talent and licensed AI voices within the same production workflow.
Launched in August, Voices for Agencies gives clients two primary options. Agencies can hire professional actors for conventional recorded voiceover or create a Branded AI Voice based on a consenting professional performer. Voices is positioning the latter as an alternative to generic synthetic voice libraries, with compensation, licensing terms and usage rights established around the actor whose voice forms the basis of the AI model.
The launch arrives as agencies increasingly experiment with synthetic speech for advertising, localization and frequently updated content. For Voices, the opportunity appears to be less about choosing between human performers and AI than establishing a business model in which agencies can use both.
Agencies Can Choose Between Recorded and Licensed AI Voices
The traditional side of Voices for Agencies remains familiar. Agencies can cast professional talent for campaigns, with access to union, non-union and international performers. Voices says its VoiceMatch technology can also be used to produce curated talent shortlists within hours.
Branded AI Voice takes a different approach. Rather than selecting an anonymous synthetic voice, an agency begins with professional talent who has agreed to participate in AI voice work. The performer is cast and directed during a recording session, providing the material required to develop a licensed synthetic version of their performance.
According to Voices, the agreement establishes issues including permitted usage, licensing duration and exclusivity before production. The actor is compensated for participating, while the agency receives a voice that can subsequently be deployed according to those agreed terms.
That structure is central to the company’s pitch. As synthetic speech has become more accessible, concerns surrounding unauthorized cloning and unclear training sources have grown alongside it. Voices is attempting to separate its offering from those practices by making performer consent and licensing part of the product itself.
The company describes its approach around three principles: consent, compensation and control. For agencies, that also provides a clearer answer to a question becoming increasingly important when synthetic speech enters commercial work: whose voice is actually being used?
Voices Sees Different Roles for Actors and AI
One of the more interesting parts of the launch is that Voices does not position synthetic speech as the appropriate solution for every project.
The company continues to recommend recorded human performance for work requiring deeper interpretation and longer narrative performances, particularly where nonverbal expression and nuanced delivery are important. That distinction can be especially relevant in corporate narration, where clarity may be the baseline requirement but pacing, warmth, authority, and audience engagement can determine whether listeners actually stay with the material. In other areas, Voices sees potential for human and synthetic performance to operate together.
Advertising provides a straightforward example. An actor might record the principal performance for a campaign, establishing its tone and delivery. A licensed AI version of that performer’s voice could subsequently be used for approved variations, such as changing promotional details or producing additional versions without scheduling another full recording session.
Localization offers another potential application. A campaign created around one performer could potentially be extended into additional languages while maintaining a related vocal identity. Voices also identifies character variations, ambient dialogue and frequently changing live-service content among the areas where licensed synthetic voices could provide additional flexibility.
This creates a different division between human and AI voiceover than simply replacing a recording session with text-to-speech. The actor can establish the original performance while the licensed model handles specific extensions of that work.
Whether agencies adopt that hybrid approach broadly remains to be seen, but Voices is clearly betting that scalable synthetic speech will become part of commercial production and that professional performers can remain involved in that process.
A Deel Campaign Shows the Model in Practice
Voices has already pointed to work involving creative agency SuperBloom and global HR platform Deel as an example of how the model can operate in a real advertising campaign.
SuperBloom initially cast a professional voice actor for Deel through Voices. According to the company, the performer’s licensed AI voice was subsequently used to extend the campaign into multiple languages, with the actor consenting to the AI usage.
The example is important because multilingual campaigns are one of the areas where synthetic speech could have an immediate production advantage. A campaign distributed across several markets can require multiple recording sessions, performers and localized versions. AI voice technology potentially allows certain elements of that process to scale much faster.
At the same time, the Deel example demonstrates the distinction Voices wants agencies to make between licensed synthetic performance and unauthorized cloning. The original performer remains part of the arrangement rather than having their recorded work transformed into an AI asset without an agreed framework.
For professional voice actors, the details of those agreements remain critical. The value of a licensed AI voice depends not only on the initial payment but also on what the model can be used for, how long those rights remain active, whether usage can expand into other campaigns and what restrictions the performer retains.
Voices Reports Rapid Growth in AI Voice Work
Voices says demand for this type of work is already increasing.
According to company-reported figures released with the new service, Branded AI Voice deals have grown 167% year over year. Voices also says nearly 40,000 professional actors participate in AI voice opportunities through its platform, while 80% of its highest-earning talent generate some income from AI-related jobs.
Those figures come directly from Voices and should not be interpreted as independent measures of the voiceover industry’s overall adoption of AI. They do, however, provide insight into how one of the industry’s largest talent marketplaces is seeing synthetic voice develop within its own business.
The numbers also complicate the familiar narrative that AI voiceover exists entirely outside the professional talent market. Voices is attempting to create a category in which actors can earn from both recorded performances and licensed synthetic versions of their voices.
For some performers, that may provide additional revenue from work that would otherwise go to generic text-to-speech systems. It also raises important questions about rates, residual value and how much control talent maintains once a reusable voice model exists.
Those questions will become increasingly important if AI voice licensing grows as quickly as Voices’ figures suggest.
Voices Is Building AI Services Across Multiple Industries
Voices for Agencies is not an isolated experiment. The company has spent 2026 introducing specialized services that combine its traditional talent marketplace with newer voice technologies.
Earlier this year, it introduced Voices for Games, targeting developers that need conventional performances alongside licensed AI voices and voice data. Voices for Customer Experience followed, applying similar capabilities to areas such as conversational interfaces and customer-facing voice applications.
The agency product extends that strategy into advertising and branded content, areas where Voices already has an established relationship with professional talent and clients.
Taken together, the launches suggest a broader evolution of the company. Voices is increasingly positioning itself not simply as a marketplace where clients hire actors, but as a voice production provider covering professional recordings, licensed synthetic voices and the data required for voice technology.
Its approach also reflects the regulatory environment developing around AI performance. Voices says its Branded AI Voice framework was designed with measures including California’s AB 2602 in mind and points to emerging entertainment-industry agreements that establish protections around digital replicas.
For actors, contractual protections will ultimately matter more than whether a synthetic voice is described as ethical or human-first. Consent needs to be meaningful, compensation needs to reflect the value and scale of the usage, and restrictions need to remain enforceable once a digital voice has been created.
A Different Model for AI Voiceover
Voices for Agencies arrives at a point when synthetic speech is rapidly becoming easier for creative teams to access. Adobe recently integrated AI speech generation directly into Firefly, while dedicated voice platforms continue expanding their capabilities and language support.
Voices is taking a somewhat different route by making professional performers part of the synthetic voice product itself.
That does not eliminate the concerns surrounding AI voiceover, nor does it determine whether an actor should agree to license a synthetic version of their voice. Those decisions will depend heavily on the contract, compensation, intended usage and individual performer’s comfort with the technology.
What the new service does demonstrate is one possible direction for the professional voiceover business as AI becomes more common. Instead of treating human voiceover and synthetic speech as entirely separate markets, Voices is building a system designed to move between them.
For agencies, that could mean recording the human performance when creative interpretation matters most and using a licensed AI version when scale, localization or rapid updates become the priority. For voice actors, the opportunity comes with a much more important requirement: remaining a compensated and consenting participant in what happens to their voice after the original recording is complete.
