Challenge
The research behind every brief was eating an hour
Eleken had a real content operation before any of this. A UI/UX design agency built for SaaS – ten years in, 100+ designers, a subscription model, and a content program aimed squarely at the founders and product teams they design for. The brand publishes, the content ranks, and the knowledge base behind it is genuinely deep. The system worked.
The problem wasn't the writing. It was everything that came before it – the research behind every brief.
For a design agency writing for SaaS buyers, a good brief needs three separate kinds of research, and none of them is quick. There's the SEO layer: what's actually ranking, which keywords matter, what search intent the piece has to satisfy, and how long the top competitors' articles run. There's the audience layer: what real people say about the topic, in their own words, in the places they actually discuss it. And there's the design layer – the angle only a design agency can bring, so the article reads like Eleken wrote it and not like generic marketing filler.
A strategist doing all three by hand – Ahrefs, a SERP scan, competitor skims, a Reddit dig, and then shaping an outline that fits Eleken's positioning – spends the better part of an hour per brief before a writer types a word. Multiply that across a content calendar and the research becomes the bottleneck the whole operation waits on.
Three ways to fix it, each with a familiar failure mode:
- Skip the research. Hand writers a keyword and hope. The result misses search intent, misses the audience's actual language, and misses Eleken's design-expert angle entirely – content that reads generic and doesn't rank.
- Put more strategist hours on it. A real fix, but an hour-plus of senior research per brief is expensive and doesn't scale with the calendar. The cost lands exactly on the people whose judgment is most valuable elsewhere.
- Use an off-the-shelf AI brief tool. Generic prompts produce generic briefs – no real SEO depth, no community insight, and definitely no design-agency lens. Writers either ignore them or rebuild them, so nothing is saved.
The fourth option – the one that worked – was three connected custom automations, each doing one research job properly, then assembling into a single publish-ready brief grounded in Eleken's own knowledge base. This case study walks through how the three fit together, what a brief now contains, and where the approach still needs a human.
Solution
What we built
The engagement produced three connected n8n automations that run as one pipeline: SEO research, design research, and brief generation. Each does a distinct research job well; the third stitches everything into a full brief. A strategist kicks it off from a Google Sheet and comes back to a finished brief about five minutes later. The Sheet is the only interface anyone at Eleken touches – no one has to learn n8n.
Automation 1: SEO research that goes past keywords
Most "SEO research" stops at a keyword list. This one goes several steps further. For each topic, the workflow pulls the Ahrefs SERP and keyword data, assigns the topic to the right content cluster, and scrapes the top-ranking pages to count words and compute an average – so the brief tells a writer the target length that actually competes, instead of guessing. It classifies the search intent, then selects the keywords worth including from the full candidate set rather than dumping all of them in.
The part that makes it different is the community layer. The workflow generates seed queries, searches Reddit through the Arctic Shift API, picks the threads that actually discuss the topic, and extracts real insights and verbatim quotes from the comments. That's the audience's own language – the questions they ask, the frustrations they voice – pulled straight into the brief. A writer opens the brief and sees not just "here are the keywords" but "here is how real people talk about this," which is the difference between an article that ranks and one that also resonates.
Everything the workflow finds is written back to a Google Sheet: SERP data, keyword selection, target word count, intent, and the community insight and quotes.
Automation 2: design research only a design agency would run
This is the piece an off-the-shelf tool could never produce. Eleken's edge is that they're a design agency, so their content has to carry a design-expert point of view – not the generic take any marketing blog could write.
A dedicated research agent, equipped with web-fetch tooling and memory, gathers the design-specific angle for each topic: the patterns, examples, and product-design perspective relevant to Eleken's positioning as a SaaS UI/UX specialist. It drafts the research into a Google Doc, cleans and structures it, assembles the final document, and saves the result back to the operational sheet – then cleans up its own working files. The output is a design lens on the topic that the brief generator can build the outline around, so the finished article sounds like Eleken and not like everyone else.
Automation 3: the brief generator that assembles it all
The third automation is the heavy one – it turns all that upstream research into a structured, publish-ready brief. It takes the SEO research, the design research, the content plan, and Eleken's knowledge base, and works through the brief piece by piece: it creates the slug, decides the funnel stage and the content type, drafts meta titles, and defines the target audience. It formats the keywords, the competitors, and the community wisdom into usable brief inputs, and extracts the real questions the piece should answer.
For the outline, it scrapes the competitor articles, summarizes each one, and has an agent build a master outline from the combination – then grounds that outline in Eleken's knowledge base so the structure reflects Eleken's services, ICP, and positioning rather than a generic template. It even pulls a relevant YouTube video and builds the internal-linking section from Eleken's live sitemap, so every brief points writers at the right existing pages. The same content-brief-automation approach (linked in the sidebar) is the backbone – a chain of focused steps, each feeding the next, rather than one big prompt trying to do everything at once.
Grounded in Eleken, not generic AI
The reason the output is usable and not just plausible is grounding. Off-the-shelf brief tools produce briefs about a topic; this pipeline produces briefs for Eleken. The prompts encode Eleken's ICP, services, and positioning. The internal links come from Eleken's actual sitemap, not invented URLs. The outline is shaped by a design agency's perspective and checked against Eleken's knowledge base. The community insight is real, pulled from real threads. Every one of those is a decision a senior strategist used to make by hand, now made consistently on every brief.
What I delivered
- Three connected n8n automations – SEO research, design research, and brief generation – running as a single pipeline
- SEO research workflow: Ahrefs SERP + keyword analysis, competitor word-count benchmarking, search-intent classification, and Reddit community mining for real audience insight and quotes
- Design research workflow: a research agent that gathers a design-expert angle grounded in Eleken's positioning, delivered through Google Docs
- Brief generation workflow: slug, funnel stage, content type, meta titles, target audience, master outline, internal links, and pitch – assembled from all upstream research
- Knowledge-base grounding so every brief reflects Eleken's ICP, services, and voice
- Sitemap-driven internal linking built into every brief
- Google Sheets as the writer's interface, so no one at Eleken has to touch n8n
Impact
The numbers
- ~60 minutes → ~5 minutes per brief, including the SEO and design research
- 3 connected automations doing the work that used to be a manual, multi-tool research session
- 12× faster from a keyword to a full, publish-ready brief
- Every brief now carries SEO depth, a design-agency angle, real community insight, and internal links – consistently, not just when there's time
What changed
The hour of senior research that used to sit in front of every brief now runs in the background. The strategist time it freed goes where it's actually valuable – deciding which topics are worth briefing, and applying editorial judgment to the highest-stakes pieces – instead of running the same Ahrefs-plus-SERP-plus-Reddit loop by hand for the hundredth time.
Writers feel the difference too, because the briefs got more useful, not just faster. Every brief now arrives with the same structure: intent, target length, the keywords that matter, the questions to answer, the audience's real language, a design-led outline, and the internal links to include. There's no more "what kind of article is this" or "how long should it be" – the decisions are made up front, consistently, on every piece.
And the design research is the part no generic tool could have delivered. Because Eleken is a design agency, their content has to carry a design point of view, and now every brief does – automatically. The bottleneck moved from "we can't research fast enough" to "what should we brief next," which is the right problem for a content team that's working.
Honest caveats
- The pipeline doesn't replace keyword selection or editorial strategy. Someone still decides which topics are worth briefing; the automations assume that call has been made.
- Research is only as good as its sources. Ahrefs data, SERP scraping, and Reddit mining depend on what exists for a given topic – thin topics produce thinner research, and the output still benefits from a strategist's eye before it reaches a writer.
- It's tuned to Eleken. The prompts encode Eleken's ICP, services, and positioning, and the internal links come from Eleken's sitemap. Pointing it at another company means re-grounding the knowledge base and re-tuning the prompts – the pipeline shape carries over, the editorial intelligence is custom every time.
- The first weeks need tuning. Like any brief system, the prompts get refined against the briefs writers actually use – which parts they accept versus rewrite – before it runs clean.
Want the same pipeline?
Book a 30-min call. Bring one keyword you've been meaning to publish on. We'll run it through the pipeline live and you'll watch the full brief – SEO, design angle, outline, internal links – come out the other end. You decide on the spot whether it's what you'd hand a writer.
Results
- Cut brief creation and SEO research from around 60 minutes to about 5 minutes per brief
- Three connected n8n automations – SEO research, design research, and brief generation – running as one pipeline
- SEO research per topic: Ahrefs SERP + keyword analysis, competitor word-count benchmarking, search intent, and real Reddit community insight
- Design research grounded in Eleken's own expertise as a SaaS design agency – not generic AI filler
- Full briefs auto-built: slug, funnel stage, content type, meta title, target audience, master outline, and internal links
- Every brief pulls internal links from the live sitemap and grounds the outline in Eleken's knowledge base
Built with
Reusable systems behind this build – same automations are available for your team.