One of the most common questions Tyler Anson Newberry hears from young worship leaders is “How do I know if a song is good?”
For Newberry, a Houston-based worship leader who has been in ministry for over 30 years, evaluating a new song takes only a few minutes.
“I have an internal song theology I use when I read lyrics,” said Newberry. “I have this algorithm in my mind built on personal experience. I can look at the first verse, chorus, bridge, and make the judgment.”
Over decades of leadership in Southern Baptist and nondenominational churches, Newberry has honed his sense of what makes a worship song useful for a congregation—he can tell if it’s singable or if the lyrical content aligns with the church’s theology of worship. When younger leaders struggle with song selection, Newberry thinks he can offer some tools to help.
“You lose touch with how much you know. I forget that I have years of experience informing my decisions,” said Newberry. “A lot of younger worship leaders don’t have a mental theology grid for the worship songs they encounter.”
He’s also noticed an absence in song-evaluation resources from denominations. Without denominational guides or standardized theological training, Newberry says many worship leaders end up selecting songs based on musical preferences rather than theological content.
“A lot of young worship leaders don’t know that they need Bible training,” said Newberry. “They have been sold a faulty bill of goods from Instagram and YouTube that tells them that all they have to be is talented.”
Just over a year ago, Newberry and his wife, Leslianne Newberry, left full-time church ministry to start Asaph’s Table, a nonprofit that offers support and training to worship leaders in the Houston area.
Newberry’s conversations with young leaders about song selection prompted him to design SongTheology, a lyric-evaluation tool powered by artificial intelligence
Newberry developed a “worship theology grid,” a set of criteria to assess and score any searched song. He wrote the language for the grid, and the tool relies on the AI tool Claude to perform the analysis of lyrics (pulled from Google) and provide theological feedback.
SongTheology allows a user to search for a song by title and see its score out of 35 possible points, given according to the strength of the lyrics in each of Newberry’s seven categories: “scriptural grounding,” “doctrinal clarity,” “God-centeredness,” “gospel presence,” “clarity vs. ambiguity,” “congregational usefulness,” and “pastoral durability.”
The tool also generates an explanation that sums up the overall score and lists “flags to note.” For example, the popular song “Gratitude” by Brandon Lake scores 20 out of 35, with two criticisms (“flags”) for its lack of “gospel clarity” and “emotion without theological grounding.”
The evaluation says the song “speaks of gratitude and worship of a king but never names Christ, mentions the cross, grace, forgiveness, or any redemptive work that would ground Christian worship.”
Newberry says that when he looks at the search data, he can see users are searching for contemporary songs and older hymns. “Come, Thou Fount of Every Blessing” is the most-searched song. Newberry suspects this may be because people often don’t know the meaning of the word Ebenezer in the second verse. He gets a laugh out of seeing that folks are searching songs like “Zaccheus Was a Wee Little Man” to see what SongTheology will turn up.
Another AI-powered resource, Asaph—an “AI worship planning assistant”—offers to help worship leaders create “setlists built on Scripture” through “instant analysis of underused themes, gaps in theology or overused emotions.” Rather than individually evaluating songs, Asaph is designed to track and evaluate the emotional and theological balance of a church’s worship sets over time.
Asaph’s founder, Alex Moyse, is a worship leader in Sydney with deep roots in the contemporary worship music world. His father was a professional musician who played guitar for the band Air Supply in the 1980s, went on to be an early member of Hillsong Church, and produced a few of the church’s first worship albums.
Moyse grew up seeing the growth of contemporary worship music and the ways production technology added new, exciting elements to a worship service—as well as new stressors for worship leaders.
When Moyse first launched Asaph, he provided worship-leading training content that he thought would be a draw for potential users.
When the idea failed to gain traction, Moyse asked a friend and fellow worship leader why, and the answer was “It’s not urgent.”
“He told me, ‘I don’t have time to onboard my team and do the whole training thing,’” Moyse said. “I realized that we have to solve the time-crunch problem.”
For Moyse, the goal of giving AI-powered tools to worship leaders is to free up some of their time. The organization and selection of music is a demand on time that he thinks could be offloaded onto a tool like Asaph that can store all the music a ministry is using and synchronize with other platforms like Planning Center (which many churches use to communicate with and schedule volunteers).
Mark Porter, a British researcher who studies Christian congregational music and teaches at the University of Erfurt, experimented with Asaph to see how it worked and what kinds of setlists it would generate.
“It can put together a coherent musical set,” said Porter. “I think anything that pushes people to be more reflective about the selection of music can be a good thing. But there’s the question ‘Does this tool do that?’”
Porter pointed out that the AI models behind Asaph matter. The people using AI to build these tools also have a point of view that informs the instructions they are giving the technology. Like SongTheology, Asaph relies on Claude.
Moyse says that he doesn’t want to help worship leaders outsource their discernment to Claude and that his team is trying to use the large language model in a limited way.
“AI is not trained to think about this stuff,” Moyse said. “We are helping the AI in a nuanced and specific way. It’s doing a lot of the grunt work, and there is still a human in the loop.”
Both Newberry and Moyse say they have created their own evaluation criteria and are using Claude to apply them. Neither tool is endorsed by a particular denomination or tradition.
Denominations like the United Methodist Church and the Christian Reformed Church have created tools to guide the adoption of contemporary worship music by local churches. These tools are optional guides, but they do equip church musicians with language to determine whether a song aligns with their beliefs or to identify the elements of a song that reflect a conflicting theology of worship.
Lifeway Worship and SBC Worship are unofficial filters for Southern Baptist churches, but neither platform provides explicit theological evaluation of music. Worship leaders serving in nondenominational or loosely networked churches have to find their own song-evaluation resources.
Newberry hopes SongTheology can be a tool for “the regular people sitting in the pews who wonder, Am I basically listening to junk food?”
Many of the mentorship and training opportunities that shaped Newberry as a worship leader have disappeared or become harder to find, he said. Even in Texas, where it seems as if there’s a church on every corner, he said too few leaders have the time and willingness to mentor budding worship leaders.
Porter suggests that worship leaders who want to use a tool like SongTheology or Asaph should consider what kind of work they’re trying to offload before bringing it into the process of planning congregational worship.
“You can’t use it to shortcut your own reflection and engagement,” said Porter. “Ask, ‘Does this tool help me to think more deeply? Or does it just reduce this all to a mechanical task?’”
Because these tools analyze songs based only on lyrical content, their ability to assess and describe the music is limited. For Porter, this is a significant shortcoming.
“Some elements of music are easily captured by a computer model. But songs have histories of being interpreted in different ways in different historical contexts. Over time, they acquire meanings and uses that aren’t captured in the text.”
Although Asaph can’t analyze the local history, harmonic content, or dynamic range of a particular song, it’s designed so that churches can track their own musical practices over time. Eventually, Moyse would like to allow churches to see the songs other congregations in their areas or denominations are using.
“I am hoping that this tool can help invert the system that brings worship music from the industry down to the local church,” said Moyse. “A local church could write a song, input it into Asaph, and a nearby church could discover it.”
Newberry sees SongTheology as one of many resources he’s trying to build for worship leaders, and he doesn’t expect everyone to be comfortable with it.
“To those who don’t want to bring AI into this process, I get it,” said Newberry. “This is just my attempt to help guide people into a new way of thinking. It’s a tool to help you begin to not need this tool.”
