How The Amarna Mysteries Are Made

Every novel involves tools. A pen. A typewriter. A word processor. Research databases. Style guides. The question has always been how writers use them and whether they are honest about it.

The Amarna Mysteries is written with the assistance of artificial intelligence. This page explains what that means in practice.

A.J. Tilke functions as creative director of a managed project team. The team includes AI tools alongside traditional research methods. Every creative decision, every historical judgment, every narrative choice, and every sentence in the final manuscripts passes through human evaluation. Nothing is published that has not been read, considered, and approved by the author.

A creative director defines the vision, sets the standards, evaluates the output, and takes responsibility for the result. The tools execute tasks within that framework. This is the method that produced four novels of historically grounded fiction.

The AI workflow supports this model at every stage. AI assists with initial drafting. AI verifies historical claims against the published scholarship. It audits manuscripts for consistency, anachronisms, and compliance with a comprehensive Series Bible. It tests narrative structure against dramatic theory. The creative authority, the historical judgment, and the final voice belong to the author, who reads, evaluates, and revises every sentence before it is published.

AI assistance begins at the drafting stage. Initial drafts are generated with AI assistance, then subjected to extensive revision: structural rewriting, historical verification, forensic accuracy corrections, language audits, and iterative polishing.

For Book 1 alone, this process produced 239 verified manuscript changes by the final pre-publication audit. The published novels are the product of that sustained analytical process. Every sentence in the final manuscripts has been evaluated, revised where necessary, and approved by the author.

The distinction matters. Some AI writing tools treat generation as the endpoint. This project treats generation as a starting point. What follows is the real work: the disciplined application of historical knowledge, narrative craft, and editorial judgment that transforms raw material into serious fiction.

Four tools play named roles in the production of The Amarna Mysteries.

Claude (Anthropic) serves as the primary AI partner across drafting, research, and editorial work. Its role includes generating initial draft material under the author's direction, historical verification, consistency checking across the series, and editorial discussion at every stage from structural planning to final polish. It informs, challenges, and refines the author's work.

Logically serves as the research platform. It provides access to Egyptological scholarship, PDF annotation for close reading of sources, and reference management. When the series makes a historical claim, Logically is where that claim is checked against the published literature.

Sudowrite is, at the platform level, a generative writing tool. In this project its use is constrained to analytical work through a custom-built plugin. The manuscripts are checked against a comprehensive Series Bible; the plugin flags compliance issues, catches anachronisms, and preserves formatting integrity. It operates as an advisory system that identifies problems and quotes their exact locations. Every correction is made by the author.

Dramatica Narrova provides structural analysis of narrative throughlines. It is used to audit the thematic architecture of the series, ensuring that the four-book arc maintains coherence across its moral and dramatic arguments.

Other tools handle formatting, distribution, and mechanical analysis, but these four represent the core of the AI-assisted workflow.

The cover artwork for The Amarna Mysteries series was generated using Claude under the author's direction. The design system pairs each book with a single mineral surface — Gold Leaf, Carnelian, Malachite, Lapis Lazuli, and Basalt — rendered as a layered CSS gradient with embedded SVG noise for grain. The covers carry no figurative imagery and no symbols: typography alone, Cinzel for the title and series mark and Cormorant Garamond for the author byline, sits on the mineral field. Cinzel and Cormorant Garamond are both freely licensed Google Fonts. Claude executed the vector layout, gradient stack, and type setting in HTML and SVG to the author's specifications. As with the prose, the author is responsible for the design and approved every element of the final artwork.

Characters, plot, and voice originate with the author. The choice of which scholarly interpretation to follow when the evidence is disputed is a human decision. What Nefertiti would say, think, or do in a given scene is a human decision. Every scene, chapter, and line of dialogue in the published books reflects human authorial judgment.

The author claims authorship of the original plot and narrative structure of each novel; the characterization and arcs of all named characters; the selection, coordination, and arrangement of all textual content; the integration of historical and archaeological research into fiction; the chapter architecture and scene sequencing; and all substantive creative revisions throughout the manuscripts. Individual AI-generated sentences and paragraphs, as such, are not claimed as protectable expression; the human-authored selection, coordination, arrangement, and creative modification of the work as a whole is.

When the evidence is ambiguous, the author decides which reading to adopt and marks it explicitly as a series canon choice. When a scene requires creative invention to fill a genuine gap in the record, the author weighs plausibility, consistency, and narrative purpose before proceeding.

AI assists with initial drafting under the author's direction, retrieves information, flags inconsistencies, and accelerates certain mechanical tasks. The judgment that determines whether a piece of historical fiction is honest belongs to the author.

In the language of U.S. Copyright Office guidance, the human author retains copyright in the original expression of the work and in the creative selection, coordination, and arrangement of all material — whether that material was drafted from scratch, modified from an AI suggestion, or composed in dialogue with the AI tools described above. The framework set out in the Office's January 2025 report on copyrightability and AI is the framework this project operates within.

The publishing industry is working through how to think about AI-assisted writing. This project takes a straightforward approach: use the tools openly, explain what they contribute, and let readers evaluate the work on its merits.

Readers of serious historical fiction are intelligent, attentive, and invested in the integrity of what they read. They deserve to know how the work was made. When the method is visible, the quality of the result speaks for itself. The question becomes "is the work accurate, compelling, and honest?" rather than "was AI involved?" That is the more interesting question.

The current marketplace contains a wide range of practices. Some authors avoid AI in any form. Others use it without disclosure. The Amarna Mysteries uses AI, including for initial drafting, with full disclosure under the framework current U.S. Copyright Office guidance sets out — human authorship of the expression and the creative selection, coordination, and arrangement of all material — and adds voluntary transparency about methodology as a standard this project sets for itself.

Historical accuracy governs every scene in the series. Every interpretive choice is evaluated against an internal framework the project calls the Accuracy Decision Ladder — a four-tier hierarchy applied in strict order. First: does the archaeological or textual record attest this directly? If so, the series follows the evidence. Second: does the scholarship support a particular reading? If so, the series adopts it as the default. Third: where the record is silent but inference is plausible, the series proceeds with caution and marks the inference. Fourth: where none of the above can answer, invention is permitted, but only if it remains historically plausible and internally consistent.

The Series Bible codifies these decisions across all four books. The Series Bible was authored by A.J. Tilke and governs canon decisions, character continuity, timeline constraints, language rules, and hundreds of specific creative choices. It is the spine of the series and the source of the editorial standards against which every manuscript is checked. The custom Sudowrite plugin checks manuscripts against this Bible automatically, ensuring that a decision made in Book 1 is still honored in Book 4.

The accuracy standard extends into the language of the novels themselves. Certain modern terms are forbidden in the ancient narrative because they carry assumptions that postdate the period. When a pre-publication audit identified that jasmine, lavender, and aconite were not attested in 18th Dynasty Egypt, all three were replaced with plants documented in the archaeological record: lily, lotus oil, and mandrake.

The scholars whose work provides the foundation for these decisions are acknowledged in the Historical Notes section of this site and in the front matter of each book.

The ladder is the operating discipline behind every choice the series makes. Its four tiers are applied in order; each tier is consulted only when the one above it cannot settle the question.

Tier 1 — Direct attestation. Where the archaeological or textual record attests something directly — a regnal date, an inscription, a title, a name — that attestation is a hard constraint. Akhenaten's regnal years follow the published chronology. The Aten temples at Akhetaten follow the excavation record. The diplomatic correspondence preserved in the Amarna Letters speaks for itself. The series does not improve on the evidence.

Tier 2 — High-consensus scholarship. Where the evidence requires interpretation but a clear scholarly consensus exists, that consensus is the default. The broad outline of the Year 12 durbar; the function and form of Eighteenth Dynasty governmental offices; the identification of the major royal women of the period and their relationships to the king — for matters of this kind the series follows the published scholarship, with citations available in the Series Bible and acknowledgments in the front matter of each book.

Tier 3 — Plausible inference. Where the evidence is silent but careful inference is possible, the series proceeds with caution and marks the inference as such. The interior arrangement of a private royal apartment; the daily rhythm of a healer's practice; the precise wording of a private conversation between two figures who certainly met — these are inferences, not records, and the prose treats them accordingly. Where an inference touches a contested scholarly question, the series identifies its position explicitly. Two such positions are disclosed pre-launch in the Interpretive Choices section below.

Tier 4 — Historically plausible invention. Where no evidence, consensus, or careful inference can settle a question, invention is permitted — but only invention that remains internally consistent and historically plausible. Minor characters, domestic detail, and the specific texture of incident may be invented; every invention is checked against the period's material culture, botany, geography, and language before it is admitted to the manuscript. When a pre-publication audit identified that jasmine, lavender, and aconite were not attested in 18th Dynasty Egypt, all three were removed and replaced with plants documented in the archaeological record.

A note for readers visiting this page before the novels have published: each book in the series rests on one or more interpretive choices large enough to warrant their own scholarly defense. The full defenses — the specific evidence weighed, the scholars whose work supports each reading, and the alternative readings the series considered and chose not to adopt — will be added to this page as each book publishes, so that the choices made for any given book can be discussed openly without spoiling the reading for those who have not yet finished it.

The same discipline that governs historical accuracy shapes the narrative structure of the series. Structure is not decoration applied to a story; it is the story's argument made visible.

The Poisoner's Throne does not follow a traditional three-act structure, and the departure is deliberate. In a conventional mystery thriller, the investigation builds to a climax, the killer is caught, and the story resolves. This novel catches the killer at the midpoint. Nefertiti proves the case, obtains a confession, and assembles overwhelming evidence. Then the novel asks the question most thrillers never reach: what happens when the murderer is indispensable?

The structure that emerges is a two-movement pivot. The first movement is the investigation: hope as engine, evidence accumulating, the net closing. The pivot is the confession that changes nothing, because removing the murderer would collapse the government, starve the army, and leave a nine-year-old pharaoh exposed to foreign invasion. The second movement is the reckoning: Nefertiti transforms from investigator to keeper, from someone seeking justice to someone preserving the possibility of justice for a future she may not live to see.

The temporal architecture reinforces this. The first movement covers roughly 150 days across fourteen chapters; each chapter is days or weeks apart, dense with event. The second movement covers nearly three years across eleven chapters; the intervals stretch to months, then seasons, then years. The date headers in each chapter make this compression and expansion visible to any reader who notices them. That pacing shift is structural, not accidental. The story slows because Nefertiti's world contracts. The reader feels the confinement in the rhythm of the prose itself.

This architecture is not unique to Book 1. Each novel in the series asks a harder version of the same question, and each novel's structure reflects the specific form of that question. The Hittite Reckoning operates as an espionage thriller because the question is about loyalty and betrayal under external threat. The Restoration Murders follows a pattern investigation because the question is whether institutional violence differs from individual violence. The Dakhamunzu Affair reconstructs a cold case because the question is whether truth can survive the people who need it buried. Structure follows theme, not convention.

The series as a whole follows the principle of the inverted detective story: the antagonist is known, the evidence is clear, and the drama lives in the distance between proof and consequence. The reader experiences something that feels like a mystery but refuses to end where mysteries traditionally end. That refusal is the point. The series argues that catching the guilty is the beginning of the problem, not the resolution of it.

Historical fiction set in a period this distant requires interpretive decisions where the evidence supports more than one reading. The Amarna Period section of this site documents those decisions in detail, explains the competing scholarly positions, and identifies which reading the series follows and why.

A few choices are structural enough to mention here. The series identifies the pharaoh Ankhkheperure Neferneferuaten as Akhenaten's sister, the Younger Lady of KV35, rather than as Nefertiti — a minority reading of contested evidence, set out in detail in the linked essay. This gives Nefertiti a different and more dramatically productive role: not the ruler herself, but the woman who watches rulers fall and decides to document what she sees.

Ay is treated as Nefertiti's uncle rather than her father, a minority but defensible reading of the evidence surrounding the Yuya family connections. The series refers to Ay as "Regent" rather than by his historical title, an accurate description of his role governing on behalf of a child king.

These and other interpretive choices are made transparently. Where the series departs from majority scholarly opinion, it says so. Where it fills a genuine gap in the record, it marks the invention. Accuracy in the small things earns credibility for the large interpretive choices. That principle governs every page.

Each interpretive choice — including the identification of Ankhkheperure Neferneferuaten, the treatment of Ay's relationship to Nefertiti, the chronology of the post-Akhenaten succession, and dozens of other decisions where the evidence supports more than one reading — was made by the author after evaluation of the competing scholarly positions. These decisions are not derived from AI output; they are authorial judgments documented in the Series Bible and reflected consistently across all four novels.

This approach was developed through practice and refined across the production of a four-book series. AI tools, used within a disciplined framework and directed by experienced human judgment, contribute meaningfully to the production of serious historical fiction. The work itself is the evidence.

The series these methods produced: The Amarna Mysteries.

The scholarship that provides the foundation: The Amarna Period: Egypt's Most Volatile Decade.

The author is solely responsible for the content of the published novels and for the accuracy of this methodology disclosure. Questions regarding the methodology may be directed to aj@theamarnamysteries.com.

A.J. Tilke is the author of The Amarna Mysteries, a four-book historical fiction series set in ancient Egypt's 18th Dynasty. The Poisoner's Throne (Book 1) will publish in July 2026, followed by The Restoration Trilogy — The Hittite Reckoning, The Restoration Murders, and The Dakhamunzu Affair — later in 2026. A short-story anthology, The Twelve Hours of Night, is planned for 2027.