The admin is eating the arts, and venues are pushing back
Ask anyone who runs a small venue, a festival or an independent label in this city what takes up their week, and almost nobody says music. They say email. They say the spreadsheet. They say the forty-seven messages asking whether the show is all ages, whether there is parking, and whether the support act is still on at eight.
None of that is why any of these people got into the industry. It is simply what has accumulated around the work, and it has been getting worse for a decade as every ticketing platform, every socials channel and every submission form added another inbox to check.
The maths of a small venue
It helps to be specific about where the hours actually go, because the answer is rarely the thing people complain about loudest.
A room with a two hundred capacity running four shows a week is fielding a few hundred enquiries in that same week across email, Instagram, Facebook and whatever the ticketing provider forwards. The overwhelming majority are variations on six questions. Door times. Age restrictions. Accessibility. Parking and transport. Whether tickets are still available. Whether the bar takes cash.
The cost is not the answering. Each one takes ninety seconds. The cost is that they arrive at eleven at night, on a Sunday, and during load-in, which is precisely when nobody is at a desk. So they queue up, and by the time somebody works through them the person asking about accessibility has already decided not to risk it.
Then there is the submission pile. Any promoter with a reputation gets more demos and pitches than they can listen to properly, which produces the outcome nobody wants: good acts ignored because the pile was too big to work through, not because anyone judged them and passed.
What the software is genuinely good at now
The tools that have arrived over the past two years are unimpressive in exactly the way that makes them useful. They are not making the art. They are handling the queue.
The pattern that works is narrow. Take the six questions that make up most of the inbox, connect them to the information already sitting on the venue’s own website, and let the thing answer at two in the morning in whatever channel the question arrived in. Anything it cannot answer with confidence goes to a human in the morning, flagged, with the conversation attached.
The same approach applies to the submission pile, though it needs more care. Sorting incoming pitches by genre, by whether the act is local or touring, by whether they have actually read the brief, does not require taste. It requires reading. A machine can do the reading and hand a human an ordered pile, which is a very different proposition from letting it decide who gets booked.
Venues that have done this well report the same thing: they did not save money on staff, they stopped losing the enquiries that arrived outside office hours. Those two outcomes look similar on a spreadsheet and are completely different in practice.
Where it falls over, and it does fall over
This is the part that usually gets left out of the sales pitch, so it is worth stating plainly.
These systems answer with total confidence whether they are right or wrong. The tone does not change. That means anything involving money, a refund, an accessibility commitment or a promise to an artist needs a person signing off before it goes out, because a confidently wrong answer about wheelchair access is far worse than a slow one.
They are also hopeless at the thing the industry actually runs on. They cannot tell that a band is about to break, cannot read a room, cannot recognise that the act who played to thirty people last year now has a following and should be moved to a Saturday. Curation is judgement built on relationships and years of turning up, and no amount of processing replaces it.
And there is a version of this that actively damages a venue: using the technology to sound like a person when it is not one. Audiences work it out quickly, and in a scene built on being genuine that is an expensive thing to get caught doing. Say plainly that the first response is automated and that a human reads everything. Nobody minds a fast answer. People mind being fooled.
The bit nobody enjoys reading
Two practical points before anyone signs up to anything.
If a system holds customer conversations, somebody needs to know where those conversations are stored and who else can see them. For a venue holding accessibility requirements and contact details, that is not a trivial question.
And there is now a disclosure expectation attached to this technology in a growing number of markets, which is heading in one direction only. Building the habit of labelling automated responses now costs nothing and saves a retrofit later.
Starting small, which is the only way it works
The failed version of this project is always the same: somebody buys a platform, tries to automate everything at once, it produces one embarrassing answer in public, and it gets switched off within a fortnight.
The version that survives is narrower and duller. Pick the single highest volume question, which for most venues is door times and age restrictions. Automate only that. Watch it for a month. Check what it got wrong, and only then add the next question.
Notice that none of those steps is a question about software. Which job goes first, which one stays human permanently, and what happens on the day the system answers something wrong in public are decisions about how the organisation runs, and they are worth settling on paper before any money changes hands.
That is the unglamorous middle of what AI consulting is meant to deliver, and it doubles as a useful filter. Any provider who wants to talk about platforms and pricing tiers before they have asked how your week actually works is selling the wrong thing, and there are a lot of them about at the moment.
What this is actually for
There is a version of this conversation that treats automation as a threat to a scene built on people, and it deserves a straight answer.
The threat is real if the technology is pointed at the wrong thing. A venue that automates its curation, or a label that lets software decide who gets a listen, will produce exactly the flattened, algorithmic outcome that people fear, and will deserve what happens next.
But that is not what is happening in the rooms where this is being used well. What is happening is that somebody stopped spending nine hours a week answering the same six questions, and used those nine hours to go and see three bands they had not seen before.
That is not a small thing. The scene has always run on people who had enough time and energy left over to pay attention. Anything that gives some of that time back is worth a look, provided it stays firmly in the inbox and well away from the stage.