How-to
A content story angle is the editorial framing the writer chooses for a story, the lens that determines whether the audience reads the piece as analytical, contrarian, instructional, or personal. Finding the right angle is the step that separates a generic take from a citable one. The method that works runs six steps: identify the story worth framing, name the audience's existing frame on the topic, propose three or four alternate framings, evaluate each against the brand and audience, pick the framing with the highest tension or differentiation, and draft from there. The pattern is mechanical, but the picks require judgment, which is why this step is the hardest to automate well.
Two writers can read the same source material, agree on the same story, and publish pieces with completely different reach because they framed the story differently. The framing is what determines whether the audience finishes the piece, shares it, or cites it later.
A weak angle takes the obvious framing everyone else takes. The piece reads as one of many takes on a topic, accurate but undifferentiated, and the audience moves on without remembering who wrote it. A strong angle takes a framing the audience didn't already have, lands a take they hadn't considered, and earns the position of the citable analysis. Over time, the writer whose angles consistently differentiate becomes the named authority in the beat.
The mechanical work (gathering sources, drafting paragraphs, formatting for platforms) is the easy part of content production. The angle work is where the editorial judgment compounds.
Start with a story that actually has multiple defensible framings. Some stories don't, a routine product update has one reasonable framing (here's the update, here's what changes). A story worth framing has tension: there are at least two reasonable ways a serious writer could approach it, the audience hasn't yet locked into one frame, and the writer's take adds information beyond what the source already provides.
Stories with high framing leverage: regulatory shifts (multiple stakeholders read them differently), industry-defining product launches (the framing battle determines positioning), policy decisions, primary-source filings that haven't been mass-covered yet, contrarian data releases.
Stories with low framing leverage: routine news, follow-ups to last week's coverage, anything where one framing has already become consensus.
Before proposing new framings, identify the frame the audience is currently using. What's the assumed narrative? What's the unspoken consensus? What would a generic write-up of this story say?
This step is often skipped, which is why writers default to the existing frame without realizing it. Naming the existing frame explicitly forces a question: do you want to reinforce it or push against it?
For example, if the topic is "AI in customer service" and the existing frame is "AI is replacing customer service agents," writing a generic piece reinforces that frame. Naming the frame opens alternatives: AI is changing what customer service agents specialize in; AI is creating new escalation paths that didn't exist; AI is exposing service operations that were always understaffed.
The frame the audience already has is usually wrong in some specific way. Finding the specific wrongness is where the angle work pays off.
For the chosen story, draft three or four framings the writer could plausibly take. The four most common framings worth considering:
The analytical framing. What does this story actually mean? Strip the noise; explain the underlying mechanism; surface the second-order consequences. Works when the audience is sophisticated and wants depth.
The contrarian framing. The audience assumes X; the case for Y is stronger. Works when the audience's default frame is genuinely missing something, not just when contrarianism would be fun to write.
The personal-experience framing. I've seen this pattern; here's what actually happens in practice. Works when the writer has operator credibility (not just opinion) on the topic.
The how-to framing. Given this story, here's what readers should do differently. Works for stories where the audience can take concrete action, not when the topic is purely observational.
Less common but useful framings: the historical framing (this has happened before; here's how it played out), the counterfactual framing (what would have happened if the story went the other way), the comparative framing (this looks like X situation; here's the relevant analogy).
Drafting multiple framings forces the writer to consider alternatives rather than defaulting to the first one. The first framing that occurs to the writer is usually the framing the rest of the beat will also take.
Score each candidate framing on four dimensions.
Brand fit. Does this framing match what the brand is known for? A brand built on analytical rigor takes the analytical framing more naturally than the contrarian one. A brand built on operator credibility takes the personal-experience framing more naturally than the how-to one.
Audience differentiation. Will this framing produce a piece the audience hasn't already seen? A framing five other writers in the beat will also take produces an undifferentiated piece. A framing nobody else is taking produces something the audience remembers.
Source-grounded defensibility. Does the source material actually support this framing? A contrarian framing that requires invented evidence is worse than no piece at all. A framing the source material supports clearly is publishable.
Risk fit. Does this framing carry risk the brand can tolerate? A sharp contrarian take might land well or might burn a relationship; the brand has to decide whether the upside is worth the downside.
Rank the candidate framings by combined score. The top-ranked framing is usually the one to pick, but with a tiebreaker: prefer differentiation over consensus when the source material supports both. Citable analysis is rare; safe analysis is common.
The pick is the editorial decision. The highest-ranked framing on combined score is usually the right call, but two heuristics override the score.
Tension wins over comfort. A framing that creates productive tension (challenges the audience's assumption, frames a tradeoff, names a problem the audience has been avoiding) holds attention. A framing that resolves into a comfortable agreement loses attention by paragraph three.
Differentiation wins over completeness. A framing that says one sharp thing well beats a framing that says four reasonable things adequately. The sharp piece gets cited; the adequate piece gets skipped. If a candidate framing is more differentiated but less complete, prefer it.
The pick is the writer's call, not the system's. A tool that proposes framings is doing the mechanical work; choosing among them is editorial judgment, and judgment is the writer's leverage.
The drafting step starts with the framing locked. The hook flows from the framing's central tension; the body argues the case; the close reinforces the framing's takeaway. Drafting starts from "I'm taking the analytical framing on this story; the central tension is X; the takeaway is Y" rather than from a blank page.
Drafting before framing produces a piece that meanders or hedges across multiple framings; drafting after framing produces a piece that lands one argument cleanly. The order matters more than the writing speed.
If the draft reveals the framing was wrong (the source material doesn't actually support the angle, or the framing falls apart on close reading), go back to step 3 and pick a different framing rather than forcing the wrong one. Re-framing mid-draft is faster than publishing a confused piece.
Three properties of a serious angle-finding method.
The method considers multiple framings explicitly. A workflow that takes the first framing that occurs and runs with it is faster but produces undifferentiated work. A workflow that proposes three or four framings forces the writer to make a real editorial choice.
The method grounds framings in source material. A contrarian framing that requires invented evidence is a fabrication risk. A serious method checks each framing against what the sources actually support before drafting.
The method incorporates brand fit and audience differentiation. A framing that wins on cleverness but doesn't match the brand or doesn't differentiate against what the audience already reads is the wrong pick. Brand and audience are scoring dimensions, not afterthoughts.
A tool that ships all three is doing useful angle work. A tool that ships only the first (generates multiple framings without grounding or brand-fit scoring) is theatrical, the writer ends up doing the real work anyway.
The mechanical parts of the method are exactly the work AI is good at: reading the story material, proposing multiple framings against a defined methodology, scoring framings on brand and audience fit, surfacing the supporting evidence for each framing. The editorial judgment (picking the framing, deciding what trade-offs to accept) stays with the writer.
A serious AI angle-finding tool runs the propose-multiple-framings step structurally (not as a free-text "here's some ideas" output), scores each framing against the brand profile, surfaces per-framing supporting evidence from the source material, and lets the writer regenerate with alternate frame keywords if the first set didn't land.
A theatrical AI angle-finding tool produces one framing, presents it as the answer, and skips the editorial-judgment step entirely. The writer ends up with content that reads as generic-AI-output because the framing step was outsourced rather than supported.
Niche's niche_angle_propose runs steps 3 through 5 of this method: for a picked story, it proposes three frame-aware angles with hooks, tensions, and CTAs, scores each against the writer's bound brand profile, and surfaces a recommended_angle_id with reasoning. The writer picks; the pipeline drafts from the picked framing.
If the proposed framings don't land, niche_angle_explore regenerates with an alternate frame keyword, cheaper than a full re-scan when the first angle set isn't right. Every output carries a brand-fit score and supporting evidence from the source material.
The 21-tool agent surface means the same workflow runs from Claude Desktop, Claude Code, Cursor, or any MCP-compatible client. Pricing is credit-based with a three-day, 1,500-credit trial that requires no card; failed runs are free.
To go deeper: read what we mean by editorial intelligence, how AI picks news stories, or what a brand profile is and why it shapes angle selection.
Keep reading