Schedule agent
Builds the next week from the template and the running orders: next episodes, repeats and rotations, then shows you what changed.
9eye is broadcast scheduling and traffic software with AI agents built in. They plan the week, keep every block on time, place ads, promos and fillers, and handle live changes — inside your rules, with your approval.
From ad placement and promo scheduling to playlist timing and as-run reconciliation: each agent knows one part of a TV channel’s day. Give it your skills once, then let it work when an operator asks, on a schedule, or by itself when something happens.
Builds the next week from the template and the running orders: next episodes, repeats and rotations, then shows you what changed.
Automatic playlist timing. Balances programme segments, break lengths and fillers so fixed starts — the news at 19:00 — land on the exact frame.
AI commercial scheduling: places the day’s spots in the right breaks, keeps clash codes apart, respects break lengths and minutes per hour, and explains each spot.
Promo scheduling that works: promos run ahead of the programme they promote, never inside it, with the passes each campaign needs.
Fills what is left between programmes and in each break with fillers, bumpers and idents from your rotation lists, to the frame.
When sport overruns or news breaks in, it re-flows the rest of the day and moves the spots that no longer fit.
Before playout export, checks every item — timing, gaps, overlaps, missing media, expired rights — hands each problem to the agent that can fix it, and checks again.
As-run reconciliation: matches the as-run log to the plan, explains each difference and lists the spots that need a make-good.
Reads new titles from spreadsheets, e-mails or your MAM into the catalog, and writes EPG texts in several languages.
Plain-language answers about anything in the system, with links to the airings, orders and items behind them.
Build agents for your own workflows, or connect one you already use — Claude, ChatGPT, Grok or your own — through MCP.
How it connectsThe agents work in a closed loop. The playlist check agent finds a problem and hands it to the agent that can fix it. Every fix is checked again, round after round, until every check passes. If a problem can’t be fixed within your rules, the operator is notified — so no known mistake goes to air.
Every check you set for the channel: timing, gaps, overlaps, media, rights, clash codes and your own rules.
A late start to the timing agent, an empty gap to the filler agent, a clash to the ad placement agent.
Inside its skills and your rules, logged with the reason.
A fix must not cause a new problem. If anything is left, the loop runs again.
The day is ready for your approval, then playout.
What is wrong, why, and what the agents tried — long before air.
Build the checking system each channel needs. Add validators from the built-in set or write your own rules in plain words, decide which ones block air, and connect each validator to the agent that fixes it. That is what drives the loop: validators find problems, their agents solve them, and the validators check again.
Start from built-in checks, or write a rule in plain words. Each channel has its own set.
Some checks must pass before playout export; others only warn the team.
Pick the agent that fixes a failure. Leave it to the operator, and the operator gets it.
Write a skill — your channel’s scheduling rules and way of working, in plain words — and give it to an agent. From then on it works that way every time, on every channel you choose. Change the skill, and every agent that uses it follows.
Fill every break between 18:00 and 23:00 with the day’s ads, promos and bumpers, so each break ends on the frame.
Moving a programme, or dropping a booked spot.
An operator asks in plain words, from any screen. The agent does the job with its skills and shows the result.
On requestEvery evening for tomorrow, every Thursday for next week. Set the time once and the work is ready before your team arrives.
Every day · 18:45A new order, an overrun reported by playout, a missing media file — the agent starts on its own, the moment it happens.
On an eventHowever it starts, the agent follows its skills, works within the roles you grant, and waits for approval before anything goes to air.
9eye speaks MCP, the Model Context Protocol — the open standard for connecting AI to software. Connect Claude, ChatGPT, Grok or an agent you built yourself, and give each one a job on your channels.
Use the AI provider your company already trusts, or several side by side.
One per channel, per daypart or per market — each with its own job and skills.
Connected agents get only the roles you grant them, and nothing they change goes on air without approval.
A wrong spot costs money and a wrong programme costs viewers. The agents are designed around that.
Nothing reaches playout until a person approves the change.
An agent acts with the roles of the person who runs it, on the channels they may touch.
Each step is logged with its reason, and every change can be undone.
Schedule data stays in your company’s space and is not used to train AI models.
The agents work on a full traffic and scheduling system that your team uses directly, every step connected from catalog to as-run.
Series, films, ads, promos, idents and fillers, with segments, air windows and clash codes.
A channel’s week of blocks and linked ad breaks, drawn once and repeated in a few clicks.
Series and programmes placed across weeks: first airings, repeats and rotations.
The broadcast day to the frame, programmes and breaks side by side, checked before air.
Exported to your automation in BXF and common playout formats, with the EPG alongside.
What actually aired, reconciled against the plan, with every difference explained.
Early access starts with a demo on a channel like yours, then a trial on your own catalog and templates.
It plans what a TV channel airs: the catalog of programmes and commercials, the weekly templates, the orders that place series and spots, the frame-accurate playlist sent to playout, and the as-run log that confirms what went on air.
They take on the repeated work of a traffic and scheduling team: planning the next week, placing ads and promos, keeping fixed starts on time, filling gaps, re-flowing the day after a live overrun, checking playlists before export and reconciling the as-run log.
The ad and promo placement agents read your booked spots and campaigns, then place them in breaks by your rules — clash codes, break lengths, minutes per hour, time windows — and give a reason for every placement. Spots that do not fit are proposed for other breaks.
No. Agents can start on request, on a schedule or on an event, but every change waits for a person to approve it before it reaches playout. Each step is logged and can be undone.
They work in a closed loop. The playlist check agent checks every item and hands each problem to the agent that can fix it; every fix is checked again until all checks pass. A problem that cannot be fixed within your rules goes to an operator with the reason, and nothing reaches playout without approval.
Yes. Each channel has its own set of validators: built-in checks such as fixed starts, gaps, clash codes and ad minutes per hour, and custom rules you write in plain words. You choose which ones block air and which agent fixes each one when it fails.
9eye speaks MCP, the open standard for connecting AI to software, so any MCP-capable agent can work with your channels, and you can build as many agents as you need. Connected agents get only the roles you grant.
Yes. Linear broadcast, FAST and streaming channels share the same core, with programme guide export for streaming platforms. What differs between channels is settings and rules.
No. Your schedule data stays in your company’s space and is not used to train AI models.