I posed the same three real-life Bible study questions to the top three AI Bible study tools, and I’m going to compare their answers. I’m excited about this one; AI is opening up some awesome tools for different kinds of Bible students.
The video linked above is sponsored by Logos Bible Software, where I worked for nine years and still have many respected friends. I love Logos, I was just in their beautiful new offices the other day to see some of those friends, and I still feel honored to promote their excellent software. If a new subscriber purchases my Ward on Words Study bundle, you’ll get six months of Logos for free. Go to logos.com/wardonwords to see it. That means six months of AI usage within Logos, which is what I’m here to talk about today: Which AI tool is best for Bible study?
I’m going to land on Logos, but not because it’s the sponsor of this piece, but because it is the tool I myself am turning to most often, by far. I’m going to give the other two tools a fair shake, I promise. Because what I care about here is actually helping people do better Bible study, no matter the tool. Some people are good with paper, and that’s fine. Somehow the great theologians of the church managed without computers, let alone AI. And some of you are going to watch this video and decide that Logos isn’t the tool for you. I think that’s okay. Many of us, however, have enjoyed using Logos for a good while, and the question is how or whether the advent of the AI era ought to shake up our toolset—even within Logos. I’m going to argue that AI makes Logos even more valuable to people who teach the Bible regularly and care about being responsible, but that some users will have reason to be satisfied with the other tools I mention.
I actually started the script for this video almost a year ago, back in January, and I found that things were changing so fast that I hardly wanted to put out a video. But that time gap will prove to be interesting; it will show some of the changes that have occurred in the tools I’m going to mention. I am not going to pretend that I have looked at every single AI Bible study tool, but I’ve had people sending me links now to new tools for many months, and I’m showing you the three that have risen to the top for me. If I’ve missed out on a new tool that you think I should know about, please let me know. AI is causing an explosion. Some good things are going to shake out into permanence, I hope.
The top three tools
With those prelims out of the way, let me turn to the three tools we’re going to examine.
1. The Logos Study Assistant
The first is the Logos Study Assistant, including the brand-new and amazing research mode. AI in Logos basically uses your Logos library as your large language model (LLM); it draws (or at least it can draw) from only your books—or from only the books in the Logos library as a whole. I love this, because it means I’ve pre-vetted the sources that AI draws from. The Logos Study Assistant also gives me clear footnotes with links so I can follow up. It even highlights the portion of each resource from which it drew a given answer. That is so, so cool. I’m really enjoying this. For me, this is the most game-changing Logos addition in many years. I worked there. I do not say this lightly. This is not hype. AI is the reason you should sign up and pay for a Logos subscription. It really is that important if you teach the Bible regularly.
2. Lumenology.app
The second AI Bible study tool is called Lumenology (affiliate link—as are others in this piece). It’s a new tool that looks like it has real potential. Its founder has been emailing me a bit this year to tell me what’s new, and Lumenology is growing. When I first picked up Lumenology, it used only public-domain resources, and not necessarily high-quality ones, but literally the day I sat down to finish this script, I got a letter from the founder announcing that he’s considering taking Lumenology toward the model used by Logos. And I see it in the software. I see even more potential here than I did when I first picked up this tool. But my review will, of course, focus on what Lumenology is today, not what it might be tomorrow.
3. Claude / ChatGPT / Gemini / Etc.
The third tool is Claude slash ChatGPT slash Gemini slash I hardly care which model you use, and they’re all changing so fast that I’m just going to stick with Claude because I stumbled into their ecosystem early and haven’t seen a reason to leave. In other words, can regular AI do well enough that you don’t need to put money into Logos or Lumenology?
I have been using AI for Bible study as I prepare sermons and Sunday school lessons, as I write books and articles and YouTube scripts and blog posts. John Piper warned his brother elders years ago that we are not professionals, and he was right. I am a servant of Christ, first and foremost: a slave. But insofar as I earn my living from a combination of Bible teaching and church web design, I don’t know what else to call my use of Bible software but professional. Bible teaching is my calling, and it’s my profession. I get frustrated quickly with tools that aren’t pro tools.
My Three Questions
So I have looked at my own search history from real-life work, doing the kinds of things that Bible teachers do, and I’ve selected three questions that I think are representative of the kinds of things that AI can help you do. I’m going to pose these questions to each major AI Bible study tool, and I’m going to compare their answers.
1. Present instances of the Messianic secret in the book of Mark.

The first question I brought to my three AI Bible study tools was fairly easy and straightforward, but not the kind of thing you could search for via keywords: “Present instances of the Messianic secret in the book of Mark.”
I was teaching a grad New Testament class at Carey Theological College, and I wanted to talk about the Messianic secret without diving in too deep; I wanted to give it maybe five minutes in my lecture. I wondered if the AI tools would give me a summary of passages that I could use to do further study. I knew I could search for “Messianic secret” in a Bible dictionary or two in Logos, but this was faster in a too-busy life, and what I needed was pretty straightforward: a list of Bible passages, and maybe a pointer or two or three to further resources.
All three tools understood my query very well. Let’s stack them up against one another and see some strengths and weaknesses, some pros and cons.
- Logos Study Assistant, in the brand-new research mode, was best for offering me solid follow-up options. It didn’t give me many, and they were a tiny bit random, but they drew from my own curated library, and that’s a huge plus for me. I need to cite sources in my academic work, and most of my work is in some measure academicky. Even when I preach, I feel some need to make certain that people I quote are clearly marked in my notes. And though I usually do not mention sources in preaching unless I actively quote them, I definitely do mention sources at times, and I leave source links in my notes for my later self. Plagiarism can get you in huge trouble. In any case, sources matter to me. When I first tested this tool earlier this year, its one weakness was sometimes giving me random resources instead of high-quality ones. But that problem, in my experience, has been fixed. And just recently, it got even better with research mode. Notice also that Logos gives me neat headings and subheadings—a perfect little article. I also get links out to the relevant passages. This is pretty much exactly what I needed.
- Lumenology gave a beautifully packaged summary with the high points and the important verses. It was visually appealing, more so than a year ago, which matters to me, with clear style differences helping me survey the information. That, not so incidentally, is one of the reasons type matters: well-used styles can quickly convey the order and importance of information. Lumenology also gave me references to less-valuable-to-me follow-up options like The Gospel Coalition (TGC), Got Questions, and the BibleProject. When I first ran this search a year ago, I wasn’t getting resources that good and reliable—or that Christian and evangelical. I was sent to the Jewish Encyclopedia and to a William Wrede quote in Wikipedia. This is what it said within Lumenology back in January: “Scholar William Wrede (1901) viewed this motif as Markan theology rather than historical fact.” That, to me, is open silliness. The only significant record we have of Jesus’ words and actions is the Gospels. I think they have to be taken or left. But Lumenology this week gave me something I can very, very comfortably recommend to people in my church. TGC’s free online commentary, a project that my friend Phil Thompson had a lot to do with. A Tyndale Bulletin article, which is awesome. BibleStudyTools.com’s CEB Study Bible, which is maybe slightly less awesome. The BibleProject, which is great. Bible Hub, which gave me a little rundown that was good but whose author was not named. I got an intro to Mark by Craig Blomberg from TGC. I got a Wikipedia link (which, by the way, is something you can turn off; I’ll get to this in a bit). I got a Ligonier link to a devotional. And I got a link to a Got Questions article. This is really quite good for a resource that points only to public-domain sources.
- ChatGPT-5 a year ago and Claude this week gave me the fullest and most elegant summary of the actual thing I was looking for: a passage list. ChatGPT even insightfully listed “Strategic Withdrawals and Avoidance of Publicity” as a category of the Messianic secret. I never thought of that. But apparently someone else in ChatGPT’s large language model has. That was helpful. And Claude gave me “counter-instances that complicate the pattern.” That, too, was super helpful.
After just one question, here’s where I’m at:
- Logos gave me what I wanted, plus literal footnotes to actual resources in my own library, to books I’ve purchased. This is the most helpful thing for me as a professional. I need information that is reliable in two ways: I need it to be accurate in itself, and I need it to be sourced in reliable resources that I can then explore and cite. I spent money over many years to build a list of reliable resources in Logos; I’m thrilled to see Logos helping me make better use of them. The Logos resources I have that tend to summarize, like certain introductory dictionaries, are probably getting sidelined by this tool. But I’ve really loved it when Logos finds relevant journal articles for me that I wouldn’t have thought to look for. This is very, very good. I am loving the Study Assistant, both in study mode for shorter answers and in research mode for more detailed ones.
- Lumenology gave me a little less information in its auto-generated article, but the resources it sent me to were good—and free. The normal Christian who doesn’t feel the need for professional tools could get real help here from solid sources, especially the TGC stuff, which is just phenomenal. What a world we live in. And with a tiny bit of coaching and customization, this can be even better, because you can use the nicely curated list of sources (those are the ones that I mentioned already) or you can add or remove web sources. That is kind of mind-blowing.
- ChatGPT and Claude gave me the bare passage list with helpful headings, but little else to go on. I can’t cite AI sources, and I don’t get a lot of pathways toward citation. The few sources it did mention were not linked; I just got last names: Cullmann, Hengel, Räisänen. I happen to know that the sources it sent me to are not evangelical. That doesn’t mean they’re all wrong or have nothing valuable to say. But it also means I wouldn’t want to send laypeople in my church to this tool. If they’re really dedicated and capable, I’ll send them to Logos. If they’re on their way to that destination but not there yet, I’ll send them to Lumenology.
2. What are some key biblical texts indicating a doctrine of original righteousness?

My second question arose from something a student mentioned to me. I’m a bizzy guy—so bizzy that I didn’t have time to spell that word correctly but instead used two Zs in my script—and I saw an opportunity for AI to give me a little shortcut once again.
I asked: What are some key biblical texts indicating a doctrine of original righteousness?
- Logos gave me teach-ready paragraphs with utterly clear and careful uses of Scripture. It gave a summary paragraph that was nothing short of phenomenal. It gave me follow-up links to two classic systematics and one journal article. This is so crazy to me, because I spent months being paid a full-time salary to produce the Lexham Survey of Theology, which was dedicated to producing articles like this. And here AI produces it in seconds. Yes, I feel a little weird about that. Welcome to the brave new world we’re all living in. But the real focus of the little instant article the Logos Study Assistant produced was Bible passages, just what I asked for. And I have the training necessary to recognize that what I’m getting is straight-up classic Christian orthodoxy. If I’m handing this tool to laypeople in my church, I think I’m also going to be urging them to do a little follow-up reading in the systematics they are given, Charles Hodge being one of them. My next impulse would be to ask Logos to send me to answers from different traditions. I did that, and I liked what I got. A year ago I asked that question in study mode and got very brief but genuinely helpful responses. This week I asked again in research mode, and I was blown away:

I got truly excellent, detailed reading recommendations with easy links to reading more in my library pulled into a clear, concise summary with neat headings. This is indistinguishable from magic, as Arthur C. Clarke might point out. All my flabber is completely gasted. My mouth is agape.
- Lumenology answered the question well for a motivated layperson, just not quite as well as Logos did. It gave me a very clear summary citing specific Scripture passages—just fewer of them. It also pointed me to free online resources, the same curated ones I mentioned in the previous question. If you need to cite your sources, I can tell you that Logos will do better in an academic setting than Lumenology. If you don’t need to cite any sources, you should probably still want to know what they are, what perspective they come from. Lumenology does that just fine. I asked Lumenology the same follow-up question, and it was again helpful and clear, just briefer and less resource-heavy than the Logos Study Assistant in research mode. I really do like the suggested follow-up questions. I could see people in my church going down helpful rabbit holes through them.
- ChatGPT and Claude didn’t point me to resources, but they gave me excellent bullet points, and lots of them. ChatGPT even did some theologizing by picking up “Thematic Threads That Support Original Righteousness,” including “Edenic harmony” and “Loss-and-restoration logic.” That latter point is exactly what a friend and I were writing about when I first asked my AI Bible study tools this question. This is a PhD-level point being made. “Notes on Interpretation Across Traditions” is especially welcome from ChatGPT. Claude was even briefer and probably less insightful, though still genuinely helpful as a summary statement. And now is probably a good time to point out that Claude could be coached. Bible study could be a project. I, for example, have about 750 full books in PDF and EPUB in a folder on my hard drive. I’ve been collecting them for years. I could totally tell Claude to use those books to help come up with answers. I could tell it to use the same sites Lumenology is using, I suppose. Lumenology provides value by doing that agent-coaching for you and putting it all in a nice UI. But AI is changing so fast: someone could surely write a prompt that would repeat its work within the AI tool you’re already paying for. I kind of hate to say this, because I really like what Lumenology is doing. But we’re living in boomtown right now, and things are not settled.
Let me make a related point: all this is a good argument for padding out your Logos library with a wide range of resources so that Logos can offer good summaries and send you to good places. I feel like the AI era has made my Logos library more valuable, not less.
3. Where does Herman Bavinck discuss Christ’s descent into hell?
My third of three questions, I’m now realizing, should probably have been an exegetical one—something about word study. I’ll toss such a question in as a bonus for subscribers to my free weeklyish newsletter; see belopw. But here’s the question I actually had and actually needed help on: Where does Herman Bavinck discuss Christ’s descent into hell?
I needed a citation as I was working on the last edits for What Saith the Lord? KJV Words We’re Getting Wrong—and How to Get Them Right. My best friend gave me a tip that I might want to look to Dutch theologian Herman Bavinck for an authoritative summation on the famous statement in the creed that “Christ descended into hell.” I needed to know where Bavinck talked about Christ’s descent.

- I have multiple Bavinck volumes in Logos. That helped. But the key ones are in my Print Library Catalog. So they didn’t show up. That’s because (and now is a good time to mention this) I quite purposefully set the Logos Study Assistant to search only my library. I did that because the whole point here is not necessarily getting answers but getting them in places I can cite. But Logos did absolutely send me to the relevant portions of Bavinck that I have in my library, and maybe I missed a setting that would allow me to search only my library and my Print Library Catalog. That would be mondo awesome. Logos summarized Bavinck’s views and gave me exactly what I needed.
- Lumenology didn’t give me any resources at all, but it did a little web search for me that pointed me in the generally right direction. It said, “I can’t verify the exact location in Bavinck from the references available here. The search results point to Reformed Dogmatics, volume 3, Sin and Salvation in Christ, as the likely volume to check, but they do not confirm a section or page. I’d rather not give you an unverified locator.” This is not the kind of question Lumenology is built to answer, unless you give it more web resources, I suppose. I’m not criticizing it. I’m just asking it professional questions. And it’s showing its perfectly understandable limitations.
- ChatGPT a year ago was equally brief and pointed me in the same direction as Lumenology. Claude this week was more detailed than Logos and did a pretty phenomenal job pointing me in the right direction. It even found a public-domain AI translation of the Reformed Dogmatics, done by Monergism. I have yet to read a full AI translation of any book, and here we go into mind-blown territory again. I’m a little nervous about this, and I’m not totally happy with the brief AI translations I have read. BUT a diligent student without the funds to buy Bavinck could get a lot out of this, I truly believe. In any case, Claude did give me an excellent summary, probably just a little better than Logos, since it could read more sources. I’d have to give Claude the win here by a nose, but then I’d waffle a bit, because could I really cite an AI-generated translation of Bavinck in a Baker Book House book? I don’t think so. So I’d land back on Logos.
This is me trying to be fair here, guys! My sponsors don’t get off easy!
A few more examples…
I have asked my AI Study Assistant in Logos many questions over the past year. I really, really love this tool. I asked, “What has the word ‘Palestine’ meant in various languages over the centuries?” Logos nailed it and sent me to more resources in my library. Lumenology did a fine job but was briefer and more lay-level, and it sent me mainly to Wikipedia. Claude was similar in detail to Logos and sent me to better online sources than Lumenology.

I asked, “What are the leading views on the meaning of the phrase ‘the lower parts of the earth’ in Ephesians 4:9?” Logos nailed it and sent me to more resources in my library; Lumenology did a fine job but was briefer and more lay-level, and sent me to resources that included nineteenth-century public-domain stuff that’s not as useful for me to cite; Claude was quite good and sent me to good resources, but not as good and useful for citing as those in my Logos library.

I asked: “Can you survey the writings of St. Augustine and mention places in those writings where he argues for something like an evangelical doctrine of biblical inerrancy?” I knew I had read such a quotation in Augustine, and that Gavin Ortlund had cited the same or similar passages. Logos nailed it and sent me right to further resources in my library; Lumenology did an unexpectedly good job, mostly because TGC has already picked up on this argument in some of its resources; Claude was quite mature in its listing of core texts, related passages that qualify the picture, and its summation. All three tools cautioned that Augustine’s bibliology wasn’t straight-up evangelical, and all were helpful. For my needs, Logos did the best.

Bonus Exegetical question

I borrow this question from my friend Phil Gons of Logos, who will be on the channel in two weeks, Lord willing, for one more discussion about AI Bible study:
How do my top 15 commentaries handle what “in love” modifies in Ephesians 1:4–5? Organize them by their positions, and include the best arguments for each.
I asked this slightly differently to Lumenology, where I don’t have a “top 15 commentaries.” I also edited it a bit for Claude, and pointed it to scholarly resources.
This is actually the kind of question I’ve been wanting someone to answer with a commentary series for a long time now. Each commentator writes in the vein of Calvin nowadays, more or less: give your own views, and interact with other commentators only as desired and needed. Nothing wrong with that. But on a discrete number of exegetical questions, I want to run a plebiscite, a referendum. I want to see all careful opinions neatly laid out, along with their supporting arguments. I’ve wanted to produce such a series for publication, but it’s too close to the Lexham Research Commentary series, and I’m not sure I’m the guy for this. Maybe.
In any case, I got good answers from all three tools, but they fell into the same buckets they fell into for the previous trial balloons (mixed metaphor alert): Logos was the most helpful, because it gave me citable sources from my own library; Lumenology was elegant and simple but featured no names and few heavy-hitters; Claude was rigorous and named helpful names, but it is limited (right now, anyway) to free online resources, and that skews the results a bit—not so much on the substance but on the citability.
Conclusion
Some concluding remarks: I love Logos. I’m biased, yes, because I know and love and trust so many people there. Good people who make awesome Bible study tools. I worked there for nine years, of course. Nine of the best years of my life. And I’ve invested literally thousands of my own US dollars in Logos books.
But I determined while I was at Logos to say only what I truly believed about the software. I think there were two times in nine years and over 200 articles in which I let myself say nice things about Logos tools that didn’t truly come from my heart. In other words, I praised one Logos tool and one Logos book with a little more gusto than I’m now proud of. Sermon Editor and the Amplified Bible were those two things, for what it’s worth. The Sermon Editor truly is good, but I didn’t end up using it the way I believe I said I would. The Amplified Bible is not so good, and I pulled my punches against it because we sold it. I regretted that immediately. As far as I know, those are my only two regrets from myriads upon myriads of words that I wrote and edited while there.
I say all this to set up my conclusion to this piece, in which I say what I really think about the future of Logos. I think it’s in no more danger than it has ever been, though that danger is more evident in the age of AI. But Logos is making good on its Hopeful Entrepreneurial Pragmatism (a John Dyer phrase I use in this video) and using AI in an amazingly helpful way. Its own AI tools don’t quite make up for the danger, but they make Logos even more valuable to those who are fully invested in it.
Here’s what I mean: Logos was never able to reach that many people who land outside the “pastors-and-other-Bible-nerds” market that we knew how to reach. My wife is a seminary graduate with access to her husband’s Logos library. She isn’t a regular teacher in church, and she effectively never teaches adults. Now that our kids are old enough for her to breathe a little better, she has joined the biblical counseling team at our church. She has therefore used my Logos library a little bit, and only in the last year.
She took Greek. She took Systematic Theology. She got better grades than me. She got a master’s degree in biblical counseling. I believe she took New Testament Introduction and probably Old Testament Theology. She reads her Bible regularly and is a devout and active Christian. If she doesn’t need Logos, who does? It’s really mainly going to be the teachers and the nerds who have the same bents as the teachers.
Some of the people on the borderline between my wife and the nerds may just stick with ChatGPT/Claude/Gemini/etc. for their theology and biblical-studies questions. But the nerds and the teachers and writers are still going to want some level of personal authority in their theological writing: they’re going to want the comfort of knowing D. A. Carson’s opinion, the comfort of knowing where in the Christian tradition certain answers are coming from. They’re going to want some level of personal accountability for the writers of their theology books, some knowledge of the confessional background they bring. ChatGPT and Claude are weirdly competent and confident. It’s a little spooky for these tools to speak so much like a person and yet not be persons. It’s comforting to me that there is a kind of orthodox middle of the road that ChatGPT and Claude seem to hew to. But what subtle cues am I giving it that that’s what I want? I don’t know. I, too, prefer to get my news and views from real-live people when it’s important.
I want you to know that while I was there, the employees and leadership never stopped talking about how to get good Bible study tools to more Christians. And we never figured it out. The Logos Study Assistant is the best new tool for professionals to come out in Logos in many years, and it may not initially seem like the greatest tool for laypeople—not when Lumenology does so well, even with just the list of curated resources it comes with. But I could absolutely see AI Bible study in Logos making Logos more accessible to more of those laypeople we were always trying to reach. I need my sources to be citable. Laypeople need their sources to be reliable. I’d love to see Logos put together a collection of resources that will aim explicitly at that goal: building a base, a kind of personal LLM, from which the Study Assistant can draw for lay Christians. That could be done now, of course, by anyone who is up for buying some books.
In fact, you can check out my little Ward on Words Bible Study bundle for a small version of that base. Just go to logos.com/wardonwords. Again, you’ll get six months of Logos for free.
I do also want to give you a link to Lumenology, however, if that is more your speed. I’ve settled on Logos because of what I’ve said in this piece, but you must, of course, make your own decision.