By Marcel Tobien • Published 29 August 2026 • 7 min read
Writing a novel can seem like one enormous task. In practice, you can treat it as a sequence of decisions: choosing a story problem, developing characters who want something, arranging events, writing scenes, and revising what does not yet work. You do not need to know every detail before you begin. You do need enough direction to decide what to write next.
There is no single required method for starting a novel. Some writers plan extensively before drafting, while others discover the story through successive scenes. An outline can organize ideas and provide direction, but it can also remain flexible. You can change it when the characters, conflict, or logic of the story suggest a different path. The useful question is not whether you are an “outliner” or a “discovery writer.” It is which amount of preparation helps you write the next worthwhile scene. The Center for Fiction also presents outlining as a flexible tool rather than a fixed contract.
Start with a story problem, not just a premise
A premise describes a situation. A novel also needs pressure. “A detective moves to a remote island” gives you a setting and a character role, but it does not yet explain why the story must unfold. A stronger starting point identifies a problem that cannot be ignored. Perhaps the detective discovers that the island’s only unsolved case is connected to their own past, while the community wants the investigation stopped.
To develop your idea, ask what the main character wants, what prevents them from getting it, and what may happen if they fail. These questions do not need polished answers. They create a working direction. You can also write a short paragraph describing the protagonist, the central conflict, the setting, and the change the character may undergo. If the paragraph contains several unrelated stories, decide which one deserves to become the novel.
Do not wait for a perfect concept before testing the idea in scenes. If you are unsure where to begin, choose the moment with the greatest tension or emotional necessity. You do not have to start with the chronological beginning of the story. Beginning with an urgent scene is one possible way to generate momentum. You can then return to earlier events and provide the context readers need. This approach is discussed in the Center for Fiction's guide to writing to the tension.
Build a flexible outline
An outline makes your story visible. It can be a page of turning points, a sequence of scene summaries, or a set of questions you want the draft to answer. Begin with the broad movement of the novel. What changes from the opening to the ending? What decision forces the protagonist into the main conflict? What complications make the original goal more difficult? What final choice shows how the character has been affected by the events?
Once you have those answers, sketch the major sections of the story. Each section should alter the situation in some way. A character may gain information, lose an ally, make a damaging choice, or discover that the original goal was based on a false assumption. If a scene leaves the characters and the reader in exactly the same position, give it a stronger conflict, a new piece of information, or a clearer purpose.
Keep uncertain sections as questions rather than pretending that you have solved them. You might write, “The protagonist confronts the missing person’s family, but what do they learn?” This gives you a direction while leaving room for discovery.
A practical process also starts with an honest assessment of your current skills. You may understand dialogue but need to study structure, or have strong characters but struggle to build scenes. The Center for Fiction's short guide frames novel writing as a set of skills and encourages writers to identify what they can already do, what may not be necessary for the current project, and what they still need to learn. Choose one relevant skill to practise during the manuscript instead of trying to master every aspect of fiction before drafting.
Write the first draft for movement
The first draft is where you find out what the novel actually is. Your outline may change, your characters may become more complicated, and scenes you expected to use may prove unnecessary. The purpose of a first draft is to create enough material to evaluate the story as a whole. It is not necessary to produce finished prose on every page immediately.
Set a writing routine you can repeat. The ideal schedule is less important than having a clear point at which you return to the manuscript. Before stopping, leave yourself a brief note about the next scene or unresolved question. This gives you a specific place to begin again.
When a scene feels difficult, identify the immediate objective. Who wants what in this moment, and what stands in the way? A concrete objective can give an abstract chapter a workable shape. If you cannot answer the question, consider whether the scene belongs in the draft, whether the character needs a clearer reason to act, or whether the conflict should begin later.
Research when the story depends on accurate details, but give research a defined purpose. Record information that affects the plot, setting, or character decisions, then return to the manuscript. You can verify and deepen details during revision. The goal is to keep research connected to the novel rather than letting it replace drafting.
Use AI as support, not as the author
AI can support brainstorming, outlining, drafting, editing, and refinement. Its role should remain connected to your process and your decisions. You might ask for alternative complications, questions about a character’s motivation, or a diagnosis of where a chapter loses tension. You decide which suggestions belong in the book.
Give AI a defined task and a clear boundary. Instead of asking it to write an entire novel, provide a scene objective and ask for possible obstacles, or share a short passage and ask which motivations are unclear. You can also ask for several structural options without accepting any of them. Narrow requests make the response easier to assess and keep the creative decisions connected to your intentions.
Maintain your own working record of character facts, relationships, timelines, setting rules, and unresolved questions. Use it when reviewing any AI-assisted suggestion. This gives you a consistent reference point for decisions about the manuscript. Keep a separate note of which suggestions you accepted, changed, or rejected if that helps you retain control of the process. For a broader comparison of workflows and tools, see this guide to AI novel writing tools and the practical discussion of writing a novel with AI.
AI should not replace your judgment about voice, meaning, or lived experience. Rewrite suggestions in your own language, remove anything that feels false, and preserve the unusual details that make the work yours. Treat generated text as material to evaluate, not as an authority on what the novel should say. Human revision remains the step that shapes the work into an individual novel.
Revise in passes
When the first draft is complete, step away from it briefly if you can. Then revise from the largest questions to the smallest. First examine the story's foundation. Does the protagonist pursue a clear goal? Does the conflict develop? Do the major events follow from character choices, or do they happen only because the plot requires them? Is the ending prepared by what comes before it?
Next, revise the chapters and scenes. Look for openings that begin too early, conversations that repeat information, and scenes that do not change the situation. Strengthen cause and effect. If a character makes a major decision, show the pressure and knowledge that led to it. If a revelation changes the story, make sure the reader can understand why it matters.
Only after these structural passes should you focus heavily on sentences. Read for clarity, rhythm, specificity, and consistency of voice. Replace vague description with details that reveal the viewpoint character’s attention. Cut repetition, but do not spend time polishing passages that may disappear in the next structural revision. Editing is not merely correcting errors. It is deciding what the novel should emphasize and removing whatever obscures it.
Choose the next useful step
Writing a novel becomes more manageable when you stop treating it as a single act of inspiration. Start with a problem that creates pressure, outline enough to find direction, draft before you feel completely ready, and revise with a deliberate sequence of questions. If you use AI, keep the authorial decisions with you and give the tool limited, reviewable tasks.
Your next step might be writing a one-paragraph premise, outlining three turning points, or drafting the scene in which the central conflict becomes unavoidable. Choose one action that produces pages or clarifies the story. Then use what you learn from that action to decide what the novel needs next.