The pitch deck for any tour management software in 2026 mentions AI somewhere. Most of it is marketing dressing on the same scheduling and routing tools that existed five years ago. Some of it is real. The question for the production manager or tour manager who has to actually run a touring operation is which is which.

I’ve spent the last few months watching what tour managers across mid-tier Australian touring crews are actually using. Some of it is genuinely useful. Some of it is the same software with a chatbot bolted on. Here’s what’s separating the signal from the noise.

The categories that are working

Itinerary generation has moved past the demo phase. The current generation of tools takes a list of cities and dates, the artist’s tech rider, the crew rider, the venue list, the support act details, and produces a draft day sheet that’s roughly 80% of what the tour manager would have produced manually. The 20% that isn’t right is mostly the local intelligence — the venue manager who’s notoriously slow to confirm, the hotel that pretends to have parking but doesn’t, the support act’s unusual load-in requirement.

That 80% is real value. A tour with twelve dates that used to take a tour manager three solid days to draft can be drafted in six hours with the AI tool, leaving more time for the local-intelligence work that actually matters.

Receipts and expense reconciliation is another category that’s working. The tools that ingest the photographed pile of receipts at the end of each day, sort them by category, match them to the budget lines, and flag the outliers are saving real time on tours where the bookkeeping was a nightmare. The accuracy is not 100% but it’s good enough to make the manual review work an order of magnitude faster.

Crew comms summaries is the sleeper category. The tour manager who comes off a long bus drive to seventy unread messages across three group chats and one production WhatsApp can now get a useful summary of what they’ve missed. It doesn’t replace reading the actual messages — but it surfaces the urgent ones and lets the rest wait.

What’s still mostly hype

Predictive booking and routing optimisation is the area where the marketing is loudest and the production reality is thinnest. The pitch is that the AI optimises the routing across a tour for cost, drive times, and venue availability. The reality is that the constraints in actual tour booking — promoter relationships, agent preferences, hold dates from venues, support availability — don’t fit into the optimisation formulations the tools use. The tour managers I’ve talked to who’ve tested these tools mostly use them as a sanity check, not as a primary booking tool.

Crowd flow and security AI for venue use is similar. The vendor demos are impressive. The deployments in actual venues are running with significant human supervision and the value over the existing security workflow is contested. This may change. It hasn’t yet.

Audio mix assistance is a category that exists but is more useful as a starting point than a final mix. The FOH engineers I know who’ve tried it use it for sound check rough-in on smaller shows. They don’t use it for headline shows.

The integration problem

The tour managers getting the most out of their AI tools are the ones who’ve thought about how the tools fit into the existing workflow. The tour budget lives in one tool. The day sheet lives in another. The hotel and travel bookings live in a third. The crew comms live in a fourth.

Tools that integrate across this stack — that read the budget when generating the day sheet, that update the hotel bookings when the routing changes — produce real productivity gains. Tools that operate in isolation produce productivity gains that get eaten by the work of moving information between tools.

This is the practical sticking point. Most touring crews don’t have the bandwidth or budget to do serious systems integration work. The tools that ship with sensible default integrations across the common stack get adopted. The tools that require integration engineering tend to sit on the shelf.

For larger touring operations or production companies running multiple simultaneous tours, the integration work becomes worth doing properly. A few have engaged outside specialists to build the integration layer. One AI consultancy we’ve talked to has done this kind of integration work in adjacent industries, and the patterns translate. Tour management is not a hugely different problem from other field operations once you abstract it.

What tour managers are still doing manually

The work that hasn’t moved is the work that depends on relationships and judgment.

Promoter and agent relationships. Knowing which promoters pay on time, which need chasing, which add on extras after the fact. No AI tool I’ve seen handles this well.

Crew management. Knowing which sound engineer is having a hard time, which roadie is reliable on long drives, which video tech is two months from quitting. These are human-management questions and the tools don’t help.

Crisis management. The night the bass amp dies thirty minutes before doors. The morning the bus breaks down outside Adelaide. The promoter who pulls the show with eight hours notice. These situations get handled by experienced people on phones, not by AI tools.

This is fine. The point of AI tools is not to replace these decisions. The point is to free up the cognitive bandwidth to make them better, by automating the calendar admin and budget reconciliation that was eating the tour manager’s day.

What I’d tell a tour manager getting started

The advice I’d give a tour manager who’s curious about AI tools is straightforward. Pick one workflow stage where you’re losing time. Adopt one tool that addresses it. Live with that for a tour cycle before adding another. Don’t replace the tools that already work. Don’t assume the AI will replace your judgement.

The tour managers who’ve gotten the most out of these tools are the ones who treat them as productivity tools, not as substitutes for the experience and relationships that make a tour work.

The tour managers who’ve gotten the least out of them are the ones who tried to overhaul their whole stack at once, didn’t finish the migration, and ended up running two parallel systems badly.

The middle path — measured adoption, thoughtful integration, realistic expectations — is the path that’s working in May 2026.