Alejandro Rioja.
Marketing SEO

Proven Ways To Generate New Content Ideas

Alejandro Rioja
Alejandro Rioja
7 min read
Free newsletter

Every Wednesday. 28,400+ operators. Zero fluff.

Table of contents

Open Table of contents

1. AI Brainstorming (Claude / ChatGPT)

This is now the starting point, not the last resort. I open a conversation with Claude or ChatGPT and give it a role plus a constraint:

“You’re a content strategist for [my niche]. List 20 questions my audience is typing into Google and asking AI assistants right now. Focus on questions where a well-sourced long-form answer would be more useful than whatever currently ranks.”

The output is rough — maybe 5 of 20 ideas are worth pursuing — but generating 20 takes 30 seconds. That’s a better ratio than staring at a blank page.

A follow-up prompt I use often: “Which of these would an AI like Perplexity or ChatGPT currently answer poorly, and why?” That surfaces topics where original research or a strong first-person perspective creates a real moat.

AI brainstorming works best as the first pass, followed by validation with the tools below.

2. AlsoAsked and AnswerThePublic

These two tools mine the “People Also Ask” boxes and related-question graphs that Google surfaces for any seed term.

AlsoAsked shows the tree of follow-up questions users ask after each query — useful for finding second- and third-order topics that don’t show up in basic keyword tools. The free tier lets you run a handful of searches per day; the paid plan (verify current pricing) removes that limit.

AnswerThePublic visualizes questions, prepositions, and comparisons around a keyword. It has been through several ownership changes (verify current status), but as of early 2026 it still works for basic question mapping.

Both tools are especially valuable for GEO (Generative Engine Optimization): AI answer engines pull heavily from “what questions does this page answer?” Use these tools to find the questions, then make sure your post explicitly answers them.

3. Reddit and Online Forums

Reddit remains one of the most reliable sources of real audience language. Find the subreddits where your target readers hang out and look for:

Sort by “Top” and “Hot” in your subreddit to see what resonates. The comment sections often contain the actual vocabulary your audience uses — match that language in your headline and intro.

Beyond Reddit, niche forums and Discord communities serve the same purpose. The signal quality is often higher in smaller, focused communities than in broad social feeds.

4. Competitor Research

I use two tools for this:

Semrush (still the industry standard as of 2026) — plug a competitor’s domain into the Organic Research report and sort by traffic. Their highest-traffic pages are proven topics. Your job isn’t to copy them; it’s to identify where you can write something more thorough, more current, or from a different angle.

Similarweb — useful for benchmarking content format distribution. Are competitors getting traction from long guides, short news posts, or video? That shapes format decisions as much as topic decisions.

One filter I apply: look specifically at competitors’ posts that rank for AI Overview snippets. Those tend to be well-structured, definitive answers — the format most likely to appear in AI-generated responses as well.

Still useful, still free. Type your seed keyword into Google and check:

This works best for validating ideas you already have, not for generating them from scratch. Combine it with AlsoAsked for a more complete picture.

X’s trending topics and search function surface fast-moving conversations. For evergreen content, I look less at what’s trending today and more at what gets bookmarked and reshared over weeks — that indicates durable interest.

On LinkedIn, the trending content section and comment threads on posts from industry figures often surface the questions professionals are actively wrestling with. Those make strong “practitioner perspective” posts.

One shift since 2023: the algorithmic feeds on both platforms have made organic discovery harder. I use them for idea mining, not for assuming my content will spread there organically.

7. Audience Surveys and Direct Questions

The most underused tactic. If you have any email list, community, or regular clients, ask them directly: “What’s the one thing you wish you understood better about [your topic]?”

A single question in a short survey, or even a direct email to 10–15 readers, often surfaces ideas that no tool would find — because they reflect what your specific audience is confused about right now.

Why GEO Now Matters as Much as SEO

In 2023, ranking in Google was the primary goal. In 2026, you also want your content to be cited in AI-generated answers — in ChatGPT, Perplexity, Google’s AI Overviews, and Gemini.

The practical difference for content ideation:

When I evaluate a content idea now, I ask: “Is there a clear, defensible answer to this question that I can give from direct experience?” If yes, it’s worth writing. If the answer is “it depends on too many variables,” I either narrow the scope or skip it.

Content Generating New Content Ideas — 2026 FAQ

Which AI tool is best for content ideation?

Claude and ChatGPT are roughly equivalent for brainstorming; the quality depends more on your prompt than the model. I find Claude slightly better at following structured constraints (“give me only questions where the current top result is thin”). Try both on the same prompt and compare. Perplexity is also useful for seeing what AI engines currently surface for a topic — that shows you the gap.

Is AnswerThePublic still worth using in 2026?

As of early 2026, yes for basic question mapping, though verify its current ownership and pricing before committing to a paid plan. AlsoAsked covers similar territory with a cleaner interface for the “People Also Ask” tree specifically.

How do I optimize for AI Overviews and AI search engines (GEO)?

Write posts that answer a specific question definitively, use clear heading structure (## question, then direct answer), cover the topic thoroughly, and include your own direct experience or original data where possible. AI engines favor authoritative, well-structured, factually grounded content — the same things that make a good long-form post.

How many ideas should I generate before picking one?

Generate at least 20 before committing to anything. Most ideas look better in isolation than when compared against alternatives. I use AI brainstorming to get 20 fast, validate the top 5 with keyword and question tools, then pick 1–2 to execute well rather than 5 to execute poorly.

Related reading:

The shorter version

If you’re reading this because the workflow it describes is eating your week, that’s the kind of loop I build AI agents for. Two build slots open at a time.

Updated for May 2026

A short note from May 2026: the workflow this post describes was checked against the current state of the underlying tools and platforms. Where specific tools, UIs, or features have evolved, the structural advice still holds — the implementation will look slightly different in 2026. If you hit a step that doesn’t match what you see on screen, that’s likely a UI refresh, not a fundamental change in approach. Drop a note via the contact form and I’ll patch it explicitly.

Keep reading

Get the AI playbook in your inbox

Every Wednesday. 28,400+ operators. Zero fluff.

⌕
↵ ↵ to see all results esc esc to close