It was perhaps inevitable that artificial intelligence (AI) would find its way into religious spaces. On one end of the spectrum, the religious public is becoming increasingly aware of – and making use of – chatbots such as Virtual Jesus (n.d.), Jesus AI (miniapps.ai n.d.) and Text with Jesus^TM (n.d.). On the other end, more conservative Christians may remain wary of venturing into this unfamiliar terrain, particularly when confronted with controversial chatbots that ‘pose’ as Jesus. However, less contentious applications such as Faith GPT (n.d.), Anchor: Christian Affirmations (faith.tools n.d.) and Gloo AI Chat (n.d.), among many others, have already become companions in the spiritual growth of ordinary believers. As a result, the proliferation of AI applications designed to support church communities has begun to attract the attention of clergy as potential ministry resources, especially in the area of sermon creation.

Because sermon preparation is traditionally a time-intensive process, sermon-assisting applications hold significant potential for easing ministers’ time constraints. Trained on vast collections of Christian texts (including the Holy Scriptures), tradition and theological resources, chatbots such as Claude (Anthropic n.d.), Google Gemini (n.d.), Microsoft Copilot (n.d.) and Meta AI (n.d.) can generate extensive material within seconds, ranging from sermon ideas and outlines to fully developed sermons. Dedicated AI sermon tools, such as Sermon Spark (n.d.), iPreach.AI (n.d.) and Sermonly (n.d.), further provide tailor-made solutions for busy preachers or for those seeking new inspiration. At the same time, these solutions present attractive alternatives to traditional homiletical resources, such as dictionaries, commentaries and sermon collections, whether in print or digital form, because they are free, easily accessible and able to provide extensive resources that may surpass conventional tools.

However, early explorations of AI as a sermon resource have raised concerns about its use. These concerns are probably best understood against the longstanding view of sermon-making as a spiritual process – one that depends on personal interpretation and is deeply embedded in lived experience (Mannerfelt & Roitto 2025:127). During the sermon-making process, the preacher is ‘listening for a voice, looking for a presence, hoping for the claim of God to be encountered through the text’, a dynamic also described as ‘wrestling’ with Scripture (Long 2005:47). Reliance on AI to produce sermons may therefore risk de-spiritualising the process, potentially shifting the traditional notion of ‘so sayeth the Lord’ into ‘so sayeth AI’.

This article examines whether such concerns are warranted. It argues that the use of large language models (LLMs) does not necessarily pose a threat but may instead be seen as a new and innovative resource for preaching – one previously unavailable on such a scale. When considered in light of the historical role of sermon resources, it becomes apparent that these were not regarded as a definitive benchmark, as the spiritual nature of the preaching process ought to be understood more comprehensively. Subsequently, the inherent spiritual nature of the sermon-making process calls for renewed discernment and critical reflection on how preachers view preaching and approach the art of sermon-making in light of innovations such as AI.

To facilitate a dialogue between traditional sermon-making approaches and contemporary AI sermon-making tools, this section examines key historical developments and intricacies in the conceptualisation of preaching and the craft of sermon-making.

Augustine’s theology, and its influence on preaching and sermon-making, would prove to be enduring. Anderson (2005:25) suggests that De Doctrina Christiana went on to become ‘the most significant source for the development of doctrinal preaching in the West’. The expression ‘Augustinus totus noster’, used by Luther, Calvin and other Reformers, attests strongly to this lasting influence (Van Oort 2009:8). However, it is also necessary to consider how later developments shaped understandings of preaching and sermon-making in the post-Augustine period. Most notably, the continuing quest for a deeper understanding of preaching led to Homiletics becoming an academic discipline at some European universities as early as the mid-thirteenth century (Long 2011:223). This rhetorical understanding of preaching persisted well into the nineteenth century, as reflected in the designation of homiletics teachers as ‘professors of sacred rhetoric’ (Long 2011:223).

The reorganisation of university curricula, together with the influence of Friedrich Schleiermacher, ultimately afforded Homiletics a place within the broader field of Practical Theology. Homiletics could now begin to understand itself as a generative discipline, rather than merely an application of theories derived from other theological fields. Arguably, it was Karl Barth’s ‘hermeneutical move’ that began to dismantle the rhetorical stronghold on preaching. Long (2011) articulates this as follows:

For preachers, this shift carried an important implication. Whereas a rhetorical framework positioned the preacher primarily as an ‘arranger’, Barth’s hermeneutical approach invited preachers into the rich world of the text itself. Long’s (2011:223) reference to ‘encounters with the biblical texts’ is of particular significance, as it suggests a deeper engagement with the ‘eventfulness of biblical language’.

Various other developments further shaped Homiletics as a field of study. Among these, advances in literary, psychological and communication theory drew Homiletics into a range of theoretical frameworks in order to enhance its effectiveness and relevance in communicating the gospel to ever-changing audiences within evolving contexts (Long 2011:223).

Comparing earlier and later developments in preaching and sermon-making, it may be fair to suggest that, once the call to preach was entrusted to the church, sermon-making assumed a symbiotic character. While the divine call to proclaim the biblical message remained the foundational principle, effective proclamation came to rely on techniques such as rhetoric and other sources of knowledge. In this regard, the era of direct-revelatory preaching was unique, and arguably the only period during which the preaching event was divinely steered. Since then, preachers have continued to heed the godly call to preach, yet have acted as ‘miners’, drawing on a range of resources in order to construct sermons that communicate the Word of the Lord as effectively as possible.

In this section, I reflect on preaching from the perspective of ministerial training. In doing so, I examine how the church and academia have understood – and continue to understand – preaching, since what is taught as normative reveals how the field is conceptualised. As noted previously, I draw in an unstructured manner on my own experiences as a student, practitioner and teacher of homiletics in order to develop an understanding of preaching from this perspective. This reflection revolves around three central questions: What was I taught? What did I do? And what am I teaching?

Mindful of my recollections of what I was taught, my initial impression during training was that the sermon-making process unfolds dynamically through the interplay of three core skills we were expected to master: exegesis, hermeneutics and homiletics. This training was grounded in our prior knowledge of biblical Hebrew and Greek, as well as Old and New Testament theology. It was further shaped by the dogmatic insights of the Reformed tradition and the pastoral obligations of preaching.

Some of the main resources for homiletics included the 1985 edition of Fred B. Craddock’s Preaching, as well as Die Woord as Opdrag [The Word as Commission] (1976) by the revered South African theologian Willie Jonker.

Writing in the contemporary South African context of his time, Jonker’s work guided students back to the Reformed foundations of preaching. It was by no means a ‘conservative’ text, as it provided a comprehensive overview of the various functions of preaching, including the exposition of Scripture, its interpretative task and its kerygmatic purpose. The book also introduced readers to innovative forms of preaching, such as dialogical and free styles. Notably, the book arose from Jonker’s conviction that there was widespread dissatisfaction with preaching in the Dutch Reformed Church (Jonker 1976:9,11) and was therefore intended to enhance its quality.

Craddock’s Preaching was an immersive learning experience as it treated sermon-making as a comprehensive process. His work departed from the assumption that ‘learning to preach is difficult’ because of the ‘complexity’ thereof, but that ‘it can be done’ (Craddock 1985:16,19).

For me, the complexity to which Craddock referred resided primarily in the dual nature of sermon-making and preaching, as both spiritual and mechanical endeavours. The mechanical aspect involves the physical and intellectual labour required in exegesis, hermeneutics and homiletics – skills that are developed through regular practice, guided by a professor in homiletics or an ordained pastor and fellow students who provide critical feedback on preparation and delivery. From their second year of Baccalaureus Theologiae (BTh) studies –following a required Bachelor’s degree in Theology, thus, typically, in the fifth year of theological studies, seminary students were also given the opportunity to preach in congregations. While these experiences were instrumental in developing the skills necessary for sermon delivery, I remained fascinated by its spiritual dimension – specifically, the point in the process at which it became spiritual. What truly made my sermons a genuine account of ‘So sayeth the Lord’ displayed boldly on so many of the pulpits I entered?

Turning to how I applied my training as a preacher, my uncertainty about what constitutes preaching as a spiritual event has deepened during my time serving as a minister in various congregations. Here, I encountered the full weight of ministry, with any idealised notions of a tranquil and pastoral life quickly dispelled by the realities of congregational demands. This primarily affected the time available for sermon preparation, often compelling me to forgo the exegetical and hermeneutical stages in order to produce material for Sunday sermons (homiletics), weekly prayer meetings, funerals, weddings and other occasions. Although Augustine’s justification for preaching sermons authored by others provided some reassurance, my unease persisted.

Soon, my collection of sermon aids exceeded my compilation of commentaries. Numerous sermon collections – comprising ‘ready to preach’ sermons – proved invaluable in the early hours of Sunday mornings (cf. Burger 1989; Cilliers 2003; Smit 1987). Sermon guidelines (books that provided exegetical, hermeneutical and homiletic frameworks) supported my sermon preparation when I had more time available during the week.

Of particular benefit was the Woord teen die Lig [Word in the Light] series of books edited by Coenie Burger, Bethel Müller and Dirkie Smit, which appeared between 1981 and 2002. According to the editors, the series was intended as exegetical, theological-hermeneutical and homiletical assistance for pastors (Burger, Müller & Smit 1981:1–2). As far as I recall, this series was warmly received by ministers of (at least) the Dutch Reformed Church, as it drew on contributions from well-known and trusted South African theologians. On many occasions, I noticed the characteristic burgundy, blue, white and green covers on the bookshelves of colleagues from other churches too. This suggested that my rather mechanical approach to sermon preparation – attempting to complete it in the shortest possible time – was not unique to my own ministry. Indeed, ‘mechanical’ may be an apt way of describing how I most often applied homiletical theory.

Since leaving full-time ministry to serve as an academic at a public university, my relationship with preaching has expanded to include the teaching of homiletics. When I was handed the baton by retiring colleagues, I found that the same guiding principles that shaped my student years largely remained in place. The central tasks still revolve around the familiar steps of exegesis, hermeneutics and homiletics. The fundamental recognition of the sermon-making process as spiritual in nature has also endured, often expressed in terms of the ‘depth’ dimension of preaching.

How, then, can we understand what makes the sermon-making journey spiritual at its core? One starting point is the assumption that those who respond to the calling to preach have a personal relationship with God, as it is God who calls them to this task. Many, if not most, works on homiletics therefore address the spiritual life of the preacher as an integral part of the homiletical process. In John Stott’s classic work, I Believe in Preaching (1982), considerable attention is given to the preacher’s life. Drawing on the thinking of Spurgeon, Wesley, Moody, Baxter and Lloyd-Jones, Stott presents a compelling case for preaching rooted in the religious and spiritual health of the preacher, demonstrated through virtues such as spiritual discipline, sincerity and earnestness.

While these virtues should be evident in the life of the preacher, they also form the foundation for the making of the sermon itself. Stott regarded this as central to effective preaching – or, as the early church argued, preaching that truly ‘moves’ the hearer. As Stott (1982) writes:

Consequently, De Klerk and Janse van Rensburg (2005:13) regard prayer as the first step in exegesis. They maintain that ‘the Holy Spirit…is working in the preacher during the whole of the exegetic and homiletic process’ (De Klerk & Janse van Rensburg 2005:13). Preachers should therefore pray for the guidance of the Holy Spirit at the beginning of sermon preparation and continue this practice throughout the process. Mindling (2006) elevates prayer to a hermeneutical function of sermon-making, while Luchetti (2008:66), following Childs (1970:219), suggests that the discipline of prayer is intrinsically interwoven with the preaching task, as ‘it anticipates the Spirit’s carrying its reader through the written page to God himself’.

Returning to Long’s notion of sermon-making as ‘listening for a voice’, the preparation of a sermon involves more than the mechanical application of skills and resources. It requires openness to, and dependence on, the guidance of the Holy Spirit in the exposition of Holy Scripture, thereby marking sermon-making as a profoundly spiritual process. In many respects, what occurs within this space – where preachers enter into this spiritual dimension – will remain a mystery, as it is deeply personal. This is particularly true because every preacher is shaped by a unique combination of factors, including religious tradition, personal perspectives on spirituality and levels of spiritual maturity. Nevertheless, preaching reaches the congregation through the distinct person and lived faith of the preacher, who, in obedience to God’s calling, bears witness to what was ‘heard’ while ‘listening for a voice’. As Craddock (1985:17) and his contemporaries have noted, God himself remains present in preaching through His Spirit, which ‘transforms’ preaching, both before and during delivery at the pulpit (my insertion), into an ‘event’.

One can therefore argue that the process does not become spiritual simply by mechanically adding communion with God through prayer and dependence on the Holy Spirit at either the beginning or end. Its spiritual nature lies in the fact that preparing and communicating the Word of God are accomplished by the Holy Spirit working through people, using resources that make the Word understandable in contemporary contexts.

Compiling a comprehensive and accurate catalogue of AI sermon-making tools available to preachers may be challenging. This is not only because of the sheer number of tools but also because of the rate at which new ones are emerging. As a result, a written synthesis of all available resources was not undertaken.

Instead, searches were carried out using three well-known AI platforms: ChatGPT, Claude and Perplexity, to obtain a general overview of the tools currently accessible (see Table 1). Using the prompt ‘Please recommend AI tools for sermon-making’, the following results were produced, each accompanied by a brief description. For brevity, three responses from each AI platform have been included as examples, with duplicate results across the platforms omitted. Applying the same prompt on other platforms yielded further suggestions.

Reflecting on the descriptions of each tool, it becomes evident that these platforms are designed as sermon-making resources intended to support preachers in sermon preparation, as reflected in the frequent use of terms such as ‘first draft sermons’, ‘outlines’, ‘tools’, ‘generating’, ‘assistant’ and similar jargon.

However, upon exploring any of these platforms, it becomes clear that, with refined prompting, complete sermons can be produced. While the tools listed above are trained on relevant datasets to provide sermon-making support, similar outcomes can often be achieved through standard AI platforms, such as those used to conduct the initial search for dedicated AI sermon-making tools. ChatGPT, for example, can be prompted to generate a sermon outline based on a particular text. The outcome may be further refined through additional prompts, such as specifying a theological tradition, the context of the intended audience, the style of the sermon and other relevant details, thereby enabling the development of a complete sermon. Throughout the process, the AI platform may suggest additional options, such as preparing a summary of the sermon for the church bulletin or generating a PowerPoint presentation to accompany it. It may also offer to develop a liturgy and worship design. In this way, an entire worship service can be generated by drawing on the most recent available resources. Preachers may furthermore tailor prompts to reflect specific traditions or denominations or refine search parameters in order to achieve the desired outcome.

What is notable is that all these platforms caution users about possible inaccuracies in their responses and advise that results should be verified if the user considers them dubious. It is therefore commonplace for platforms to include disclaimers, such as the familiar slogan at the bottom of the ChatGPT screen: ‘ChatGPT can make mistakes. Check important info’ (OpenAI n.d.).

In addition to general disclaimers, AI platforms also remind users that their functions are not intended to replace the preacher’s own role in the sermon-making process. ChatGPT, for example, states clearly that:

From an appreciative stance, and without veering into issues of bias or theological integrity, two significant advantages of AI sermon-making support become evident: open access to vast datasets of theological resources, together with the ability to synthesise information in response to user prompts, and substantial time savings. Anecdotally, it is as if the wishful thinking of time-strained pastors has suddenly become reality – sermons can be produced at the touch of a button. For older preachers, the possibilities of AI sermon-making tools may be reminiscent of earlier sermon collections that enabled them to draw on the sermon-writing abilities of others and their resources, albeit now on a vastly expanded scale. Relating this to the inherent spiritual character of sermon-making and preaching, it is worth noting that AI sermon-making tools can be used either as resources to assist in sermon preparation or as tools to generate complete, ready-to-preach sermons. This distinction may be compared to the difference between traditional resources that support sermon creation, such as Bible translations, commentaries and theological works and sermon collections. The former required the preacher’s labour to extract and synthesise material into a sermon, whereas the latter provided a complete end product.

This article began by acknowledging that AI has permeated religious life and the ministerial domain. The ability of AI platforms and dedicated AI sermon-making tools has raised concerns that their use may de-spiritualise the sermon-making process and challenge traditional views of sermon-making and preaching.

Adopting an appreciative stance and drawing on relevant literature, AI platforms and unstructured personal experiences, this article revisited preaching and sermon-making to recover essential aspects of their nature. The reflection was limited to two broad eras in preaching. During Old Testament times, numerous examples of direct revelation from God are evident, as leaders and prophets were instructed to proclaim His will to different groups of people. With the emergence of the Christian church, preaching increasingly drew on the contributions of preachers such as Paul, who utilised rhetoric to reach listeners, while relying on existing theological knowledge and lived faith in the resurrected Jesus as a gift of the Holy Spirit.

When the early church began to explore the meaning of preaching and the art of sermon-making, knowledge of the Scriptures, rhetoric and other sources was considered essential for creating eloquent sermons capable of both instructing and inspiring listeners. The widespread availability of sermon material and the systematic organisation of preaching resources indicated that preachers were not expected to rely solely on their own abilities but were, in fact, encouraged to draw upon available resources to improve their sermons, even if this involved delivering someone else’s sermon.

Following Augustine’s monumental influence on preaching, the ‘hermeneutical move’ was recognised as a return to the biblical text itself. This shift proved significant for modern understandings of sermon-making, framing it as a struggle with the text rather than merely the rhetorical organisation of Scripture.

Interrogating the training of preachers reveals that sermon-making depends on proficiency in three key skills: exegesis, hermeneutics and homiletics. It was noted that these processes are situated within a spiritual framework, involving prayerful dependence on the guidance of the Holy Spirit during both the preparation and delivery of sermons, transforming preaching into a spiritual event. Alongside the central roles of exegesis, hermeneutics and homiletics is the continual availability of resources that have, throughout history, supported preachers in sermon preparation – and, as mentioned in the author’s personal reflection, have sometimes replaced the process entirely when preachers turn to ready-to-preach sermon collections.

A brief overview of AI sermon-making tools reveals a growing array of resources that provide open access to LLMs with remarkable capabilities. Effective use of both dedicated AI sermon-making tools and general AI platforms depends on formulating clear prompts that specify a particular theological tradition while maintaining awareness of the primary purpose of these tools – namely, to function as an aid rather than as a replacement for authentic sermon creation.

From an appreciative reflective stance, and without further argument, I propose that AI sermon-making tools can largely be seen as a contemporary expression of past sermon-making resources. However, traditional resources such as dictionaries, commentaries and interlinear Bibles differ from AI tools in two significant ways. Firstly, written resources were produced by theologians who applied their theological knowledge and discernment in the development of their works, whereas AI tools are the product of machine training. Even if AI is trained within specific theological traditions, it possesses neither faith nor conscience. Accordingly, these tools remind users that the spiritual discernment required for sermon-making remains their responsibility, as reflected in the numerous disclaimers accompanying them. Secondly, AI tools offer access to vast amounts of data, speed and current information – qualities that contrast sharply with traditional resources, which were limited by the preacher’s budget, remained comparatively restricted in scope and inevitably became outdated.

AI platforms and AI sermon-making tools offer a new and dynamic resource for preachers, which alert users to their internal vulnerability that requires constant verification. Exploring them need not pose a threat to the integrity of the sermon-making process, as its authenticity does not depend on the resources preachers consult, but on how these resources are discerned and applied within a task inherently understood as spiritual. Of course, AI can be used in much the same way as the ready-to-preach sermon collections of earlier times, insofar as it is capable of generating complete sermons with minimal input from the user. Preachers should therefore reflect carefully on whether such use ought to become common practice or whether it should remain the exception.