The Concierge: AI Worship Assistant

The Concierge is a chat-based assistant inside DailyWorship that helps with worship planning. Open it from the side nav. Ask it anything related to your service, your songs, or the liturgical moment you're preparing for.

What the Concierge is good for#

Service planning. "We're doing a Maundy Thursday service. Suggest a song flow that moves from celebration into the upper room and ends in stillness."
Theme-finding. "I want a song about God's faithfulness through suffering. Help me think through the angle."
Prompt drafting. "Help me write a generation prompt for a song based on Psalm 51 for an Ash Wednesday service."
Tradition-aware suggestions. "We're Anglican, second Sunday of Advent. What would a song selection for our church look like?"
Lectionary follow-up. "What scriptures are coming up next Sunday in the Revised Common Lectionary?"

The Concierge has access to your library, your organization's shared library, and the public discovery feed. It can recommend tracks you already have, tracks your team has shared, or new generations.

How to open it#

Sign in.
Click Concierge in the side navigation.
A new conversation starts.

Past conversations are kept in the side panel. Open any to continue where you left off.

What context the Concierge knows#

The Concierge sees:

Your account profile (name, denomination if you've set one)
Your active organization workspace (if you're in one)
Your library titles, themes, and styles (for recommending tracks you already have)
The current liturgical week (for traditions that follow a calendar)

It does not see:

Your billing or payment information
Other users' private libraries
Anything you haven't said in the current conversation

A note on privacy#

Conversations are stored against your account so you can resume them. You can delete a conversation at any time from the conversation list — when you delete it, it's removed from our servers within seven days.

We do not train models on your conversations. The Concierge uses commercial AI services (currently Google's Gemini 2.5 Flash) and your conversations are sent to those services for inference, governed by their respective privacy terms. We do not use your conversations to fine-tune or train any model.

What the Concierge is not#

A replacement for pastoral discernment. The Concierge suggests; the worship leader decides.
A search engine for the open web. It works within DailyWorship's content and its training knowledge of Christian theology and worship practice. It doesn't browse the live web.
A theology referee. When you ask it a doctrinal question, it will give you a historically-orthodox Christian answer where one exists. It will tell you when a question is contested within Christian tradition rather than pretending there's a single answer.
A way to bypass moderation. Generation prompts you ask the Concierge to help write still go through the same moderation as if you wrote them directly.

Tips for getting better suggestions#

Be specific about the moment. "A song for the offering moment" is better than "a song." "A song for the offering on Christmas Eve" is better still.
Name what you've already used recently. The Concierge can avoid recommending songs you used last week if you mention it.
Tell it about your team's strengths. "We have a strong vocalist but no full band right now" shapes recommendations toward acoustic-friendly arrangements.
Give it scripture references when you have them. Scripture-grounded prompts produce more focused suggestions.

Common questions#

Does the Concierge know what songs my church sings?

Only if they're in your DailyWorship library or your org's shared library. It doesn't have access to your CCLI history or external setlist databases.

Can it generate songs directly?

The Concierge drafts generation prompts you can take to the Studio. It doesn't run the generation itself — that's a one-click handoff so you stay in control of the spend.

What if it gives a wrong scripture reference?

Tell it. The Concierge will check and correct itself. If you spot a pattern of errors, report it via support — we use these to refine the prompts that drive the assistant.

Related Articles: