
Overview
A radio station fills the gaps between songs with spoken links, station identifications and local items. Makers now sell software that writes and voices those links with generative AI and a synthetic voice, and software that takes scheduling requests typed in everyday words. This guide is for programme directors, engineers and producers deciding whether to try one. It sets out what ENCO and RCS state, what their pages leave out, and a trial you can run before anything reaches the air. We did not test any tool on this page. Each statement about a tool comes from the maker's own page as read on 9 October 2026, or from a trade source named in the Sources section.
Kit and Prerequisites
- Your current automation and scheduling system and the version you run
- A week of your own logs, so you know what the links and schedule look like today
- A person who will review generated copy during the trial
- Your station's position on synthetic voices and on telling listeners
Working Through It
Decide which job you are trialling
Voice tracking produces spoken material. Scheduling arranges music. They fail in different ways. A wrong link is heard by listeners at once, while a poor schedule shows up slowly in ratings and complaints. Trial one at a time.
Read what ENCO states about aiTrack
ENCO's own page says aiTrack lets stations mix live talent with AI-generated content and build generative copy insertions with automated voice tracks into playlists well in advance of airtime. It says listener interactions can use pre-set voice types or custom cloned voices. It also says scripts are auto-generated just ahead of playout to keep them relevant. The page says dynamic content insertion for ENCO's DAD automation system was still to come when it was written. A Radio World article from the 2025 NAB Show describes aiTrack as an add-on to DAD for prescheduled song intros, station identifications and localised news and weather. Neither page names a voice engine, supported languages or any review step.
Read what RCS states about RCSAI
RCS's own page returned an error, so this rests on a trade article of 2 September 2026. It says RCS would debut RCSAI in SelectorCloud at IBC2026, that it lets programmers enter scheduling requests in natural language, and that it is the first application on a new engine called RCS Intelligence. The article gives no release date or price, so check whether it has shipped.
Read what we could not read about AI co-hosts
Futuri sells CoHostAI, which search summaries of 2024 trade coverage describe as letting a human host and an AI personality interact on air. The coverage pages returned errors, so we state nothing more about it and none of its figures.
Run the trial off air
Generate a week of links for your own format with names, places and local items you know. Have your reviewer mark every factual error, every awkward phrase and every item that would embarrass the station. Count them. Listen for how the voice sounds after an hour, because repetition shows up in a long listen.
Set the rules before it goes to air
Decide who approves copy, what a synthetic voice may and may not say, who is told, and how a presenter takes over for breaking news. Read our guides on voice consent and on labelling AI-made content before you answer.
Worth knowing:
- Start with the least risky material, such as time checks and station identifications, not news.
- If a tool can clone a voice, settle the presenter's written consent first.
- Keep a person on call who can switch the automation off in a minute.
Knowledge Base
The Longer Answer
The pages describe how generated links reach the air. None describes who checks them.
Why review is the question
ENCO's page describes scripts generated ahead of time and, in one place, just ahead of playout. A script generated minutes before it airs leaves little time for a person to read it. The page does not say whether a review step exists. Ask any maker how a human approves, edits or blocks a link, and what happens when no one has.
Announcements are not shipping products
RCSAI was announced for a September 2026 debut and the article we read gives no release date. ENCO's page speaks of dynamic content insertion for DAD as coming. A feature that is announced should be treated as a promise until you have seen it run on your own system.
Listeners and disclosure
Futuri says its own research found that listeners want to be told when AI is used. We saw that only in search summaries of coverage, and it comes from a company that sells AI hosts. Do not rely on it. Read the rules that apply to your market, and decide your own policy.
Where This Fits
AI voice tracks fill the gaps in a log that Best Podcast Setup for Beginners explains from board operation to on-air delivery.
FAQ
Q: Can AI replace a radio presenter?
A: We did not test any tool and the pages we read do not say. They describe filling links between songs and mixing with live talent. A person decides what airs.
Q: Does ENCO aiTrack work with other automation systems?
A: The pages we read describe it as an add-on to ENCO's own DAD system. They do not mention other systems. Ask ENCO.
Q: Is RCSAI available now?
A: It was announced for an IBC2026 debut. The article we read gives no release date, and RCS's own page was unreachable. Check with RCS.
Vocabulary
Voice tracking Recording the spoken links between songs ahead of time and placing them in the playlist, so the station sounds live. AI voice tracking generates and voices those links with software instead of a presenter.
Synthetic voice engine Software that turns written text into spoken audio in a voice that is not recorded from a person at that moment. ENCO's pages say aiTrack uses engines of this kind and do not name them.
Automation system The software that plays a station's music, links, adverts and jingles on schedule, and logs what aired. ENCO's DAD is one example, and the pages describe aiTrack as an add-on to it.
Natural-language scheduling Typing a request in everyday words, such as asking for a change to a music rotation, and letting software carry it out, rather than working through menus and screens. RCS states that RCSAI offers this.
Sources
One maker's page and two trade articles were read on 9 October 2026. Nothing was tested. Checked 9 October 2026.
- ENCO, aiTrack wins Best of Show. The maker's own page. Describes generative copy insertions with automated voice tracks, pre-set or cloned voices, and DAD support still to come. Names no engine, languages or review step.
- Radio World, Best of Show: ENCO aiTrack (27 May 2025). A trade article. Describes an add-on for DAD that prescheduled song intros, station identifications and localised news and weather. Names no engines, languages or review.
- RedTech, RCSAI in SelectorCloud at IBC2026 (2 September 2026). A trade article. States the IBC2026 debut, natural-language scheduling requests and the RCS Intelligence engine. Gives no release date or price.
RCS's own SelectorCloud page returned HTTP 403, and the Futuri study and CoHostAI coverage pages returned 403 or 429, so nothing in this guide depends on them beyond saying they exist. ENCO's page is undated in the text we read. Radio World gives conflicting accounts of which ENCO product won an award at IBC 2025 in search summaries, so no award claim beyond the NAB 2025 article is made.
Translate this page
- Español
- 简体中文
- हिन्दी
- العربية
- Português
- Français
- Deutsch
- 日本語
- Русский
- Bahasa Indonesia
- 한국어
- Italiano
- Türkçe
- Tiếng Việt
- Polski
- Nederlands
Machine translation provided by Google Translate, on Google’s servers. We do not check these translations and they will get technical terms wrong. The English page is the authoritative one. Following a link sends this page’s address to Google. Your browser may also offer to translate this page itself, which keeps the request on your device.