Why Synopsis Middles Collapse
An agent reading your synopsis is not reading for plot. They are reading for evidence that you understand how narrative causality works—that you know the difference between things that happen in your novel and things that happen because of your novel's logic. The middle third of a synopsis is where that distinction becomes brutally visible.
Most synopsis middles fail for the same reason: the author summarizes the manuscript the way they experienced writing it, which is chronologically. One event follows the next. The protagonist tries something, it doesn't work, she tries something else. This produces what screenwriters call "and then" sequencing—a chain of events connected by temporal proximity rather than causal necessity. An agent trained to spot structural problems reads "and then" sequencing as evidence that the author doesn't fully control their own story.
The alternative is cause-and-effect escalation. In a well-constructed synopsis middle, each beat exists because the preceding beat made it necessary—and simultaneously made the protagonist's position more precarious. The inciting incident creates a problem. The protagonist's first attempt to solve it produces an unintended consequence. That consequence forces a choice that costs the protagonist something. The cost of that choice eliminates one of their remaining options. By the time you reach the climax, the protagonist has been systematically stripped of easy exits, which is why the climax feels like the only possible confrontation rather than one of several plausible outcomes.
When novelists compress 90,000 words into two pages, they instinctively cut the connective tissue—the small decisions, reversals, and consequences that make each major beat feel earned. What remains looks episodic because it is episodic. The synopsis has preserved the events without preserving the causality, and a synopsis without causality is just a list of things that happened to your protagonist before the ending.
AI tools are particularly effective at diagnosing this problem because they can evaluate the logical relationship between sentences without the emotional investment you have in your own story. You know why your protagonist does what she does in chapter fourteen. The AI only knows what you wrote. If the reasoning isn't on the page, the AI will tell you—which is exactly the diagnostic function a good editorial eye provides.
The Causality Audit Prompt
The first intervention is a structural audit that treats your synopsis as a logical argument rather than a narrative summary. You are asking AI to read your synopsis the way a hostile reader would—looking for every moment where a plot beat depends on coincidence, convenience, or a character decision that hasn't been adequately motivated by what preceded it.
Before you run this prompt, paste your full synopsis into a fresh document and label the three sections: inciting incident through the first major reversal (Act One), the escalating complications through the darkest moment (Act Two, which is your middle), and the climax through resolution (Act Three). You don't need to be precise about where sections begin and end—you just need to isolate the middle third so you can direct the AI's attention to it specifically.
I am revising a novel synopsis before querying agents. Paste the middle section of my synopsis below. Read it as a causality auditor, not a copy editor. For every plot beat or character decision in this section, evaluate whether it is: (A) Causally necessary — it happens because of something established earlier and it makes the next beat necessary (B) Episodic — it happens, and then the next thing happens, but there is no demonstrated causal link between them (C) Convenient — it happens primarily because the plot needs it to happen, not because the characters or established circumstances demand it (D) Undermotivated — a character acts in a way that is not adequately explained by what we know about them or what situation they are currently facing Return a beat-by-beat list. For each beat, give me the label (A, B, C, or D), one sentence explaining your reasoning, and — for any beat that is not A — one specific question the reader is left asking that the synopsis does not answer. Do not rewrite anything. Only diagnose. [PASTE SYNOPSIS MIDDLE HERE]
The questions the AI generates in that final column are your revision targets. Each question represents a piece of causal connective tissue your synopsis is currently omitting. Some of those answers exist in your manuscript and simply need to be compressed into the synopsis. Others may reveal that the manuscript itself has a logic gap you hadn't recognized.
The Stakes Escalation Test
A synopsis middle that passes the causality audit can still fail if it doesn't demonstrate escalating cost. Agents are looking for evidence that the protagonist's situation becomes progressively worse—or more precisely, that the protagonist's attempts to improve her situation consistently produce consequences that narrow her options and raise the price of failure. A middle that stays at roughly the same level of tension from the first reversal to the climax reads as flat, regardless of how many events it contains.
The stakes escalation test asks AI to identify the exact moment in your synopsis middle where the protagonist's situation stops getting worse. That moment is where your middle goes slack.
I am preparing a novel synopsis for agent submission. I need to test whether my middle section demonstrates proper stakes escalation. Read the following synopsis section and complete these four tasks: 1. ESCALATION MAP: List every moment where the protagonist's situation measurably worsens. For each moment, name what specifically is now at risk that was not at risk before, or what resource/option has been lost that previously existed. 2. PLATEAU IDENTIFICATION: Identify the point — if one exists — where the protagonist's situation stops deteriorating. This is where tension flatlines. If the stakes remain static for more than one beat, mark the entire static sequence. 3. COST AUDIT: A well-constructed middle requires the protagonist to pay at least two distinct costs before reaching the climax — meaning she must lose or sacrifice something that genuinely matters and cannot be recovered. Identify whether this synopsis contains two such costs. If not, identify where a cost beat is missing. 4. MINIMUM ESCALATION DRAFT: Write one sentence — do not write more than one sentence — that could be inserted at the plateau point to raise stakes. This sentence should name a concrete new consequence, not a vague threat. [PASTE SYNOPSIS MIDDLE HERE]
The single-sentence escalation draft in task four is intentionally limited. You are not asking the AI to fix your synopsis—you are asking it to demonstrate the kind of specificity a stakes escalation beat requires, so you can draft your own version with full knowledge of your manuscript's details. Use the AI's sentence as a structural placeholder to remind yourself what kind of beat is missing, then replace it entirely with language drawn from your actual story.
The Compression Trap
Compression is not the same as omission. When you cut a turning point from your synopsis entirely, that beat simply disappears—the synopsis moves from beat three to beat five with no acknowledgment that beat four existed. When you compress a turning point too aggressively, beat four is technically present, but it has been summarized so tightly that the causal information it carries has been lost. The synopsis says that something happened; it does not say why it happened or what it changed.
Both problems produce an episodic middle, but they require different solutions. Omitted beats need to be written into the synopsis from scratch. Over-compressed beats need to be unpacked—typically by adding one or two sentences that restore the character motivation and the specific consequence that were stripped out during compression.
AI can distinguish between these two failure modes because it can evaluate whether a transition between beats contains enough information to justify the logical leap it's asking the reader to make.
I am revising a novel synopsis and need to identify where I have compressed turning points so tightly that the causal logic has disappeared — what I'm calling "invisible" turning points. Read this synopsis section carefully. For each transition between major beats — meaning each moment where the story moves from one situation to a meaningfully different situation — evaluate whether the transition is: TRANSPARENT: The synopsis explains what changed, why it changed, and what the character did in response. A reader could reconstruct the narrative logic from the text alone. COMPRESSED: The synopsis acknowledges that a change occurred, but has stripped out the character motivation or the specific consequence, leaving the reader to assume the logic rather than follow it. INVISIBLE: The synopsis jumps from one situation to another with no acknowledgment that a turning point occurred. The reader cannot tell how the story moved from point A to point B. For each COMPRESSED or INVISIBLE transition, answer two questions: - What information would need to be present for this transition to become TRANSPARENT? - Is this likely a beat that exists in the manuscript but was over-edited in the synopsis, or does it feel like a structural gap in the story itself? Return your findings organized by transition, not by paragraph. [PASTE SYNOPSIS MIDDLE HERE]
The second question in this prompt is doing significant diagnostic work. If the AI consistently flags transitions as over-edited (present in the manuscript, lost in the synopsis), your revision task is primarily a drafting problem—you need to restore compressed information. If the AI flags transitions as potential structural gaps, you may need to return to the manuscript itself before the synopsis can be fixed. A synopsis cannot be made structurally sound if the underlying story is not.
Rebuilding the Spine
Once you have identified which paragraphs in your synopsis middle are failing and why, the most efficient revision method is to ask AI to rewrite a single sagging paragraph using explicit cause-effect language, then study the result to understand the structural logic you need to replicate in your own revision.
This works because cause-effect language has recognizable markers—words and constructions that signal that one event produced another: "which forces," "leaving her without," "because of this," "now that X has happened, Y is no longer possible." When you see those constructions applied to your own story's beats, you can identify which of your sentences need to be restructured to carry that same logical freight.
I am going to give you one paragraph from the middle of my novel synopsis. This paragraph has been flagged as episodic — it lists events without demonstrating how each event causes the next. Your task is to rewrite this paragraph using strict cause-effect language. Every sentence must connect to the next through demonstrated causality, not temporal sequence. The rewritten paragraph should make clear: - Why each action the protagonist takes is the direct result of what preceded it - What each action costs or forecloses - How the protagonist's situation at the end of the paragraph is measurably worse (or more constrained) than at the beginning Rules for your rewrite: - Do not invent new plot events. Use only the events present in the original paragraph. - Do not use the word "then" as a causal connector. - Do not use passive constructions that obscure agency ("a decision is made," "complications arise"). - Every character action must have an explicit motivation visible in the sentence or the immediately preceding sentence. - End the paragraph with a sentence that names the specific thing the protagonist has now lost or the specific constraint she is now operating under. After the rewrite, provide a brief structural annotation: for each sentence in your rewritten version, identify in brackets what narrative function it is performing (establishing motivation / showing attempt / naming consequence / closing option / raising cost). Here is the paragraph to rewrite: [PASTE SAGGING SYNOPSIS PARAGRAPH HERE]
The structural annotation is the most instructive part of this output. Read it before you read the rewritten paragraph. Understanding what narrative function each sentence is performing will teach you to evaluate your own synopsis sentences the same way—not asking "does this sentence summarize what happened?" but "what is this sentence doing in the causal sequence?"
A Note on What AI Cannot Fix
These prompts are diagnostic and structural tools. They work by evaluating the logic of what you have written and identifying where that logic breaks down. They cannot supply the specific details—the character history, the thematic resonance, the particular cost that will devastate your protagonist because of who she is—that make a synopsis feel like it belongs to a specific and irreplaceable novel.
Every revision you make after running these prompts should be written by you, in your own language, drawn from your manuscript's actual material. The AI identifies the gap. You fill it. This is not a limitation of the tool—it is the correct division of labor between diagnostic analysis and creative authorship.
Run the causality audit first, then the stakes escalation test. Use the compression trap prompt only after you understand which beats are genuinely missing versus which are present but illegible. Save the spine-rebuilding prompt for last, as a model to study rather than a solution to copy. By the time you have worked through all four interventions, your synopsis middle should be doing what agents need it to do: demonstrating that you understand why your story unfolds the way it does, and that the structure of your novel is not accidental.
No comments yet. Be the first to comment!