Building an AI Workflow for Your Venue Booking InboxZapier + Claude + Make: auto-classify 100+ pitches a month, draft tier-appropriate replies, and never lose a good act to a slow inbox again

Why your inbox is the bottleneck

If you run a UK grassroots venue that books live music, your booking inbox is almost certainly the single biggest time sink in your operation. Not the sound checks. Not the PRS returns. The inbox.

Here’s the pattern. An artist emails a pitch: band name, genre, a Spotify link, maybe a one-sheet PDF. You need to open it, read it, check the Spotify, skim the socials, decide if it fits your room, and either reply or file it. That’s 8–15 minutes per pitch if you do it properly. At 100 pitches a month, that’s 13–25 hours — a full working week, every month, just on triage.

Most bookers don’t have that time. So pitches pile up, good acts don’t hear back, and the venue ends up booking from a shrinking pool of “acts I already know” instead of the 100 who emailed. The inbox isn’t just slow — it’s actively costing you discovery. (For context, many of those incoming pitches are now AI-automated on the artist’s side too — which is partly why the volume keeps climbing.)

The fix isn’t “hire someone to read emails.” The fix is building a workflow that reads them for you, sorts them by fit, drafts the replies, and puts the shortlist in front of you ready to approve. That’s what this guide builds.

What you need before you start

• A dedicated booking email — e.g. bookings@yourvenue.co.uk. If pitches currently go to a shared info@ inbox, set up a forwarding rule or alias. The workflow triggers on new emails to this address.
• A Zapier account — the free tier handles up to 100 tasks/month. If you get 100+ pitches, the Starter plan (£16/month) covers 750 tasks. Zapier pricing.
• A Claude or ChatGPT account — Claude Pro (£18/month) or ChatGPT Plus (£20/month). Both work. This guide uses Claude because it handles long emails better and hallucinates less on UK-specific context. Free-tier ChatGPT works for low volumes (<30 pitches/month) but hits rate limits fast.
• Your booking criteria written down — genres you book, capacity, typical fee range, nights you programme, acts you already have booked. The AI needs this as a system prompt. 10 minutes to write, reusable forever.

Step 1: Write your venue brief

This is the most important step and takes 10 minutes. Write a plain-English description of what your venue books. The AI uses this as its decision criteria for every pitch it reads. Here’s a template:

Venue: [Name], [City]. Capacity: [number] standing. PA: [yes/no, type].
Genres we book: [list]. Genres we don't book: [list].
Nights: [Fri/Sat/both/weekday]. Typical fee range: £[X]–£[Y].
What makes a good fit: [e.g. "acts with a provable local draw of 30+,
active social presence, have played venues of similar size"].
Red flags: [e.g. "no live gigs in 12+ months, only covers, no UK dates"].
Already booked: [list next 4–6 weeks of confirmed acts so the AI avoids
genre clashes].

Be specific. “We book rock and indie” is too vague. “We book indie-rock, post-punk, and alt-country on Fridays and Saturdays. No metal, no DJ sets, no tribute acts. Typical fee £200–400 for a 3–4 piece with provable draw of 30+” gives the AI enough to make real decisions.

Step 2: Build the Zapier zap

The workflow has 3 steps in Zapier:

The critical design decision: draft, don’t send. The AI classifies and writes; you approve. This isn’t autonomous — it’s a reading assistant that happens to also write first drafts. You can override any classification or edit any reply before it goes out.

Step 3: The classification prompt

This is the prompt that goes into Zapier’s Claude action. Copy it, replace the bracketed sections with your venue brief, and paste it into the “Instructions” field.

You are a booking assistant for [Venue Name], a [capacity]-cap [genre]
venue in [City]. Here is the venue brief: [paste your brief from Step 1].

Read the artist pitch below and return a JSON object with these fields:
- classification (yes-shortlist, maybe, no-thanks)
- confidence (0–100)
- genre_match (true/false)
- red_flags (array of strings, empty if none)
- summary (2 sentences max: who the act is and why they do or don't fit)

Base your decision only on information in the pitch. If information is
missing (no Spotify link, no gig history, no draw estimate), flag that
as a red flag — don't guess. Never invent facts about the artist.

The JSON output makes it machine-readable for the next Zapier step. If you want plain text instead, replace “return a JSON object” with “return a short report” and skip the programmatic routing.

Step 4: The reply templates

Three templates, one per classification tier. The AI drafts these using the classification output, personalised with the artist’s name and genre. You edit before sending.

Every reply sounds like it came from a human who read the pitch — because an AI did read the pitch. The difference between “we’ll keep you on file” from a template and “we’ll keep you on file” from an AI that actually parsed the genre and gig history is that the second one includes a real reason. Artists notice.

Step 5: Escalation rules

Not every pitch fits neatly into three buckets. Build these escalation rules into your Zapier workflow:

• Confidence below 60% — route to your inbox with a “needs human review” flag instead of auto-drafting a reply. The AI isn’t sure; don’t let it guess.
• Known agent or promoter — if the sender domain matches a known agency (e.g. Paradigm, CAA, Primary Talent), bypass classification and flag as priority. Agent pitches are pre-vetted and time-sensitive.
• Multiple red flags — if the pitch triggers 3+ red flags (no links, no gig history, generic template, wrong genre), auto-draft the no-thanks reply but also log the pitch to a spreadsheet for your monthly pattern review.
• High-profile act — if the Spotify monthly listeners exceed [your threshold, e.g. 50k], flag as priority regardless of genre match. A 50k-listener act emailing a 150-cap room deserves a fast human reply even if they’re slightly outside your usual genre.

The escalation layer is what separates a useful workflow from a dumb autoresponder. It handles the 80% that’s routine and routes the 20% that needs judgement straight to you.

Make.com alternative

If you prefer Make.com over Zapier, the workflow is almost identical. Make’s free tier gives you 1,000 operations/month (vs Zapier’s 100 tasks), so it’s better value for high-volume inboxes. The setup:

• Trigger: Watch Gmail/Outlook module
• Action 1: HTTP module → POST to Claude API (or use Make’s built-in OpenAI module with ChatGPT)
• Action 2: Router module → branch by classification
• Action 3: Gmail/Outlook “Create draft” module per branch

Make’s visual builder makes the routing logic easier to see and debug than Zapier’s linear step chain. For venues processing 100+ pitches, Make is the better tool. For under 50, Zapier’s simplicity wins.

What this costs

• Under 30 pitches/month: £0. Free Zapier + free Claude. Hit rate limits occasionally but manageable.
• 30–100 pitches/month: ~£16–20. Zapier Starter (£16) or Make free tier + Claude Pro (£18). The sweet spot for most UK grassroots venues.
• 100+ pitches/month: ~£35–50. Zapier Professional or Make Core + Claude Pro. Multi-room venues, festival offices, venues with national profiles.

Compare that to the cost of not replying: a good act that doesn’t hear back books somewhere else. A bad act that slips through costs £200–800 in lost door, replacement fees, and staff time. The workflow pays for itself if it catches one bad booking or saves one good one per quarter.

What this doesn’t do

This workflow reads emails and drafts replies. It does not:

• Send anything without your approval — every reply sits in drafts until you hit send
• Book acts — it classifies and recommends; you decide and confirm
• Replace your judgement on culture fit — the AI can match genres and check red flags, but “would this act work on our stage on a wet Thursday in November” is still your call
• Handle contracts, fees, or deposits — those happen after classification, on the platform or via your normal process

The boundary is clear: the AI reads, you decide. If you want the vetting layer that goes deeper into each shortlisted act (social audit, draw verification, EPK red flags), read AI for vetting acts — it picks up exactly where this workflow’s “yes-shortlist” bucket leaves off.

Where this fits in the cluster

This is post 7 of 9 in the AI for UK Music Venues cluster. If you’re starting from scratch, begin with the foundations:

• Foundations: 12 ChatGPT prompts for venue bookers (copy-paste inbox replies), best AI tools for venues (what to pay for), AI to fill your venue (marketing layer).
• Intermediate: AI for venue programming, AI for venue compliance, AI for vetting acts.
• Advanced (you are here): AI booking inbox workflow. Next: AI for venue analytics (live), AI for festivals & multi-stage venues (Sep 2026).
• The full venue cluster: see AI for UK Music Venues (2026) for all 9 posts in the series.
• The other side: Want to understand how artists are building and automating their pitches? The AI for Musicians cluster covers the artist’s workflow from the other end of the inbox.

Annual refresh commitment

This guide was published on 15 May 2026 and is refreshed every May. We re-verify every reference, recommendation, and data point once a year. Next scheduled refresh: May 2027. If any claim is outdated before then, email hello@gigxchange.app and we will update it within 24 hours.