Your content backlog is thin, and not everything on it is still worth building. Before you brief another creator, answer a smaller, sharper question: now that AI search already answers most of what your buyer typed in, what’s actually still worth commissioning?
That’s different from “should we keep doing content.” Of course you should. The real decision is which pieces earn a slot on a short list, and which ones you quietly stop asking creators to write. This guide gives you a framework for making that call without an SEO background, so your commissioning budget goes toward pages that still do work once the AI answer shows up.
What changed
Two things shifted the ground here, and neither is the tired “AI killed SEO” story.
First, community platforms are eating a growing share of the citations AI engines hand out. In a February 2026 analysis for G2, SEO researcher Kevin Indig found that across roughly 35,000 ChatGPT citations for SaaS-buying-journey prompts, user-generated platforms, Wikipedia, Reddit, and LinkedIn, accounted for 17.1% of non-vendor citations, more than four times publishers’ 4.0%. That’s a narrow finding: one commissioned analysis, one engine (ChatGPT), one query type (SaaS buyers), one month of data. Treat it as a signal worth watching, not a universal law of AI search. Still, it lines up with a broader, separately-sourced pattern: Profound’s research on citations across multiple AI engines found that community platforms, Reddit especially, are consistently among the most-cited sources, not just for SaaS-buying queries.
Second, you have no first-party way to measure your own exposure to that shift. Google Search Console’s July 2026 update added tracking for Instagram, TikTok, X, and YouTube as platform properties. It did not add Reddit or LinkedIn. As Indig put it plainly: “Reddit is not on the list. Neither is LinkedIn.” That means the platforms carrying outsized weight in AI answers are also the ones you can’t check inside your own analytics. As a result, you’re commissioning content partly blind, which is exactly why a clearer prioritization rule matters more now than it did two years ago.
A framework for what’s still worth building
SEO researcher Aleyda Solis published a content-prioritization framework in August 2026 built for exactly this moment: deciding what’s worth keeping or building when an AI answer already exists for most queries. It’s her framework, developed for auditing an existing site’s worth of pages with a spreadsheet, not TBM’s own research. We’re translating her six criteria into plain questions a founder can answer in one sitting, without the audit tooling.
Here’s what she asks, and how to ask it yourself:
- Click resilience. Would someone still visit your page after reading the AI’s answer? A page that requires verification, comparison, or a next action tends to hold up. A page that just restates a definition doesn’t.
- Citation potential. Does the page say something unique, verifiable, or well-documented enough that another site or AI answer would want to reference it? Generic advice rarely gets cited. Original data, named sources, and documented processes do.
- Brand mention potential. Does the topic give AI engines a real reason to associate your brand with it, based on expertise you actually have? Not aspirational association. Demonstrated expertise.
- Business value. Does the page support a credible path toward someone hiring you, not just reading you? If it doesn’t touch acquisition, conversion, retention, or support, its business case is weak.
- Proprietary advantage. Could a competitor recreate this page without your data, your access, or your operating context? If the answer is yes in an afternoon, the page has no moat.
- Expected effort. What does this actually cost to produce and keep current, in editorial time, subject-matter input, and maintenance?
Run each backlog idea through those six questions before you commission it. A page that scores well on citation potential and business value but poorly on proprietary advantage is still worth building, just not worth building as a commodity explainer anyone could write.
What clears the bar for AI search
Here’s Solis’s prioritized list in buyer-facing terms. These categories consistently clear the bar:
- Pages about your brand and what you actually do (not generic category pages)
- Pages that let someone complete a task or transaction, not just read about one
- Product documentation and policies stated plainly
- First-hand tests and performance reviews you ran yourself
- Original research on your market, audience, or usage patterns
- Live databases or reference hubs built on your own data
- Documented customer outcomes and case studies, with real specifics
- Troubleshooting and implementation guides tied to a specific product or process
- Evidence-led comparisons and selection guides that show your reasoning
- Personalized tools that pull from live or proprietary data
- Curated community and practitioner knowledge you’ve synthesized, not just aggregated
- Original reporting or source analysis nobody else has done
- Pages built at scale, but only when they’re grounded in genuine first-party data
Notice the pattern. Every item on that list requires something an AI answer can’t fully replicate: your data, your judgment, or your accountability for the outcome.
What to stop building
Just as important is what to pull off the backlog. These categories rarely earn their production cost anymore:
- Standalone definitions with no connection to your brand or your buyer’s journey
- Rehashed explainers that add nothing beyond what’s already well documented elsewhere
- Fragmented FAQ or keyword-variant pages built to catch search traffic, not to answer a real question
- Rewrites of someone else’s press release or news item
- High-volume, tangential topics that never touch a real business decision
- “Best of” roundups with no disclosed methodology or clear bias
- Generic calculators or quizzes with no distinctive inputs
- Programmatic pages assembled from public or competitor data with nothing original added
If an idea on your list matches one of these, it’s not worth a creator’s time right now, however easy it would be to produce.
Where this fits with your AI-vs-traditional call
Once you’ve settled who has final sign-off on AI-assisted versus traditional work, the harder question is what that team should actually spend its hours building. This framework answers that second question. It’s the same accountability logic applied one layer down: not just who approves the work, but which work is still worth approving in the first place.
If your backlog is full of commodity explainers because that’s what used to rank, this is the moment to prune it. Redirect that budget toward the handful of pages only you can credibly write, and let the rest go.
Talk to us about your backlog
If you’re staring at a content backlog and can’t tell which items are still worth commissioning, that’s a conversation worth having before you brief another creator.
We can also help you think through the AI-assisted versus traditional call this framework builds on. See our companion guide on AI creative agency vs traditional agency for that decision.