Case study · Digital Arts Media Services
MCP Server for Tessitura
Your team already uses AI. Now it can use Tessitura.
- Sector
- Performing arts
- Works with
- Claude, Claude Code, ChatGPT, Gemini
- Integrates with
- Tessitura REST API
- Hosting
- Client’s Azure, AWS or Google Cloud — or managed
The challenge
Staff at arts organizations already use AI assistants every day — but not with their most important data. Getting an answer out of Tessitura usually means building a report or exporting a spreadsheet, and the tempting shortcut of pasting patron data into a chatbot is exactly what an organization doesn’t want.
Different teams also need very different access: the box office works with orders, development with donors, leadership with summary numbers, and the public should only ever see what’s on sale.
The solution
Digital Arts Media Services built a custom Model Context Protocol (MCP) server for Tessitura: a secure bridge between Tessitura and the AI tools staff already use. The server exposes a set of well-defined tools — check availability, look up an order, brief me on a donor — and enforces who can use each one.
The AI never calls the Tessitura API directly and never sees credentials. Sensitive fields are redacted on the server, changes wait for approval, and every call is logged — so one governed integration serves every AI tool in the building.
How it works
Staff & tools
- Claude & Claude Code
- ChatGPTCustom connectors
- Gemini-based tools
- Internal apps & agentsIncluding the patron chatbot
Tessitura MCP server
- Sign-inSSO (Entra ID, Google, Okta) or API keys
- PermissionsPer tool, per role, per field
- RedactionSensitive fields removed
- ApprovalsEvery write checked by a person
- Audit & limitsLogs, rate limits, query caps
Tessitura
- Tessitura REST APIDedicated service user
- Test environmentEvery tool proven here first
- Production
- Your Azure
- Your AWS
- Your Google Cloud
- or managed by DAMS
What each team can ask
Each department gets its own tools — and only its own tools.
| Team | Example questions | Access |
|---|---|---|
| Public | “What’s on next month, and are there seats left for Saturday?” | Calendar and availability, read only |
| Box office | “Which orchestra seats are still open for the matinee? Resend this patron’s confirmation.” | Orders, notes, seat holds (with approval) |
| Marketing | “How is sales pacing for the new season, and how many first-time buyers so far?” | Sales data; personal data redacted |
| Development | “Brief me on this donor before tomorrow’s meeting. Who has lapsed since last year?” | Contribution history, contact updates (with approval) |
| Leadership | “How is revenue tracking against budget? How do renewals compare with last season?” | Summary financial and renewal data |
| Developers | “How is this configured? Update this template.” | Configuration, via Claude Code |
Security model
Three ways to sign in
Public tools need no sign-in and only see public information; staff sign in with single sign-on (Microsoft Entra ID, Google Workspace or Okta); systems use API keys.
Permissions per tool, role and field
Each department sees only the tools and fields it needs. Start read-only and add write actions where they earn their place.
Approval for every change
Notes, contact updates and seat holds wait for a person to approve them.
Redaction on the server
Payment details, birthdays and private notes are removed before anything reaches an AI model.
Audit, limits and caps
Every tool call is logged with who asked, what ran and what changed. Rate limits and query caps protect the Tessitura instance.
Your cloud or ours
Hosted in the organization’s own Azure, AWS or Google Cloud account, or managed by Digital Arts Media Services.
Only what a tool returns is sent to the AI provider the organization chooses, under that provider’s terms — and sensitive fields can be redacted on the server first.
Delivery
Discovery
Map the questions and tasks each team wants AI help with.
Tool & permission design
Define each tool, who can use it and what stays private.
Build & test
Develop against the Tessitura test environment first.
Deploy & connect
Go live and connect each team’s AI tools, with setup guides.
Train & evolve
Train staff, monitor usage and add tools as needs grow.
Requirements: access to the Tessitura REST API and a dedicated service user with appropriate permissions.
What it delivers
- Answers in seconds, from real numbers — no reports to build, no exports to manage
- One governed integration for every AI tool and internal system
- Every tool call permission-checked and logged; every change approved
- Patron and donor data that stays out of AI models unless it’s meant to be there
This server can also power the AI Chatbot for Tessitura. For MCP servers on other systems, see Custom MCP Servers.
Tessitura® is a trademark of Tessitura Network, Inc. This solution is independent and not affiliated with or endorsed by Tessitura Network.
More case studies
- Digital Arts Media ServicesAI Chatbot for TessituraA 24/7 box office assistant that works from live Tessitura data — shows, seats, bookings and memberships — with checkout on the venue’s own TNEW site.Read the case study
- Sectorly.aiSectorly.ai PlatformA multi-industry AI platform: governed chatbots, a hotel concierge, appointment scheduling and AI-assisted practice management.Read the case study
Contact us
Want something similar?
Tell us about your organization and the systems involved — we’ll show you what’s possible.
- 202.747.3355
- [email protected]
- 1717 K St., N.W. Suite 900, Washington, DC 20006