AI & Automation
Personas
SaaS & Startup
Time to ROI
Short-term (< 3 months)
Last month, I was working on a B2B startup website project as a freelancer. The client needed a complete SEO strategy overhaul, and the first critical step was obvious: build a comprehensive keyword list that would actually drive qualified traffic.
I started where every SEO professional begins—firing up SEMrush, diving into Ahrefs, and cross-referencing with Google autocomplete. After hours of clicking through expensive subscription interfaces and drowning in overwhelming data exports, I had a decent list. But something felt off.
The process was expensive (multiple tool subscriptions adding up), time-consuming (endless manual filtering), and overkill (thousands of irrelevant keywords to sort through). That's when I remembered I had a dormant Perplexity Pro account somewhere.
What happened next completely changed how I approach keyword research. Using Perplexity's research tool, I built my entire keyword strategy in a fraction of the time. The platform didn't just spit out generic keywords—it understood context, search intent, and competitive landscape.
Here's what you'll learn from my experiment:
Why traditional SEO tools are becoming obsolete for keyword research
How AI research tools outperform expensive subscriptions
My exact workflow for building comprehensive keyword lists with AI
The critical mistakes most people make when using AI for SEO
How to validate AI-generated keywords for actual search volume
Before we dive in, check out my AI content automation guide and SEO audit framework for more context.
Industry Reality
What every marketer has already tried
The traditional keyword research process follows the same playbook everyone knows. You start with a seed keyword, plug it into your favorite SEO tool, and export thousands of suggestions. Then comes the manual work: filtering by search volume, competition, and relevance.
Here's what the industry typically recommends:
Use multiple paid tools - SEMrush for competition analysis, Ahrefs for backlink context, Google Keyword Planner for search volume
Analyze competitor keywords - See what's working for others in your space
Focus on search volume and difficulty scores - Target keywords with decent volume but manageable competition
Group keywords by intent - Separate informational, commercial, and transactional queries
Manual verification - Double-check everything by actually searching on Google
This conventional wisdom exists because it worked when search was simpler. Before AI, we needed these tools to understand search behavior patterns and competitive landscapes. The process was methodical, data-driven, and reliable.
But here's where it falls short in 2025: these tools show you what everyone else is already targeting. They're great at revealing obvious opportunities but terrible at uncovering unique angles or understanding nuanced search intent. Plus, the data is often months old, and the volume numbers? Completely unreliable.
Most importantly, traditional tools treat keywords as isolated data points rather than understanding the actual user problems behind the searches. They'll tell you "SaaS pricing models" has 2,400 monthly searches, but they won't help you understand why someone searches for that or what they're really trying to solve.
That's exactly where AI research changes the game completely.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
When this B2B startup came to me, they were in a typical situation: smart product, decent traction, but zero organic visibility. Their current website was getting maybe 300 visitors per month, mostly from direct traffic and a few referrals.
The client was a SaaS company building workflow automation tools for mid-market businesses. They had a solid product but were stuck in the classic startup trap—building features instead of building an audience. Their previous marketing attempts had been scattered: some Google Ads here, a few blog posts there, but no cohesive content strategy.
My first move was the traditional approach. I signed up for a SEMrush trial, pulled competitor analysis from Ahrefs, and started building keyword lists the "right" way. I spent three days diving deep into search volumes, competition scores, and difficulty ratings.
The results? A spreadsheet with 1,200+ keywords that looked impressive but felt generic. Most were obvious variations like "workflow automation software," "business process automation," "workflow tools for teams"—the same keywords every competitor was already targeting.
Worse, when I tried to understand the real search intent behind these keywords, the tools gave me surface-level categorizations. "Workflow automation" was labeled as "commercial intent," but what did that actually mean? Were people looking for comparisons? Implementation guides? Pricing information?
I realized I was following the same playbook as every other SEO consultant. I was optimizing for keywords that tools said were valuable, not keywords that would actually help my client connect with their ideal customers.
That's when I decided to try a completely different approach. Instead of starting with tools, I started with questions: What problems do my client's customers actually face? How do they describe these problems when they're searching? What related topics do they care about?
And that's where AI research tools became a game-changer.
Here's my playbook
What I ended up doing and the results.
Instead of opening SEMrush, I opened Perplexity Pro and started with a simple prompt: "What are the main workflow automation challenges that mid-market B2B companies face, and how do they typically search for solutions?"
Within seconds, I had a comprehensive analysis that included not just keywords, but context, pain points, and user intent. Perplexity didn't just list "workflow automation software"—it explained that companies search for solutions when they're dealing with manual handoffs between departments, when they're scaling and processes are breaking down, or when compliance requires better documentation.
Step 1: Intent-Based Research
I used Perplexity to map out the entire customer journey. Instead of keyword lists, I got user scenarios:
"How to automate approval processes in HR departments"
"Why manual invoice processing is killing our finance team"
"Workflow automation for remote teams with multiple time zones"
Step 2: Competitive Context Without the Noise
I asked Perplexity to analyze what competitors were missing: "What workflow automation topics are underserved in the current content landscape?" The AI identified gaps like industry-specific implementations, integration challenges with legacy systems, and change management for automation rollouts.
Step 3: Long-Tail Discovery
Traditional tools excel at finding variations of head terms. AI excels at finding completely different ways people express the same need. Instead of "workflow automation tools," I discovered searches like "how to stop emails getting lost between departments" and "reduce manual handoffs in project management."
Step 4: Semantic Clustering
Rather than grouping keywords by exact match variations, I used AI to cluster them by actual user intent and business problems. This revealed content opportunities that traditional tools would never surface.
Step 5: Content Angle Generation
For each keyword cluster, I asked Perplexity to suggest unique content angles that weren't already covered by competitors. Instead of generic "Ultimate Guide to Workflow Automation," I got specific angles like "Why Your Finance Team Hates Your Current Workflow (And How to Fix It)."
The entire process took about 4 hours across two days—not two weeks of manual research. More importantly, I ended up with keywords that felt connected to real business problems, not just search volume numbers.
Research Method
How AI understands context better than traditional tools
Content Angles
AI reveals unique perspectives that tools miss
Validation Process
How to verify AI recommendations with real data
Implementation
From research to content calendar in one workflow
The results were honestly surprising, even to me. Using Perplexity's research capabilities, I ended up with a keyword strategy that was both more comprehensive and more targeted than anything I'd built with traditional tools.
Instead of 1,200 generic keywords, I had 180 highly specific keyword clusters organized around actual user problems. Each cluster included not just the keywords, but the context, search intent, and content angle.
More importantly, when I started creating content based on this research, the engagement was immediately higher. The first three blog posts I published based on AI-generated keyword insights got 3x more organic traffic in their first month compared to previous content optimized with traditional keyword research.
The client saw results within 8 weeks: organic traffic increased from 300 to 1,200 monthly visits, and more importantly, these were qualified visits. The content attracted people who were actually dealing with workflow automation challenges, not just researchers or competitors.
But the biggest win was time efficiency. What used to take weeks of research and validation now happens in days. I can pivot keyword strategies quickly, test new content angles, and adapt to changing search patterns without being locked into expensive tool subscriptions.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
This experiment taught me that keyword research isn't about finding keywords—it's about understanding people. Traditional SEO tools optimize for the algorithm, but AI research tools help you optimize for actual human needs.
Here are the key lessons that changed how I approach SEO entirely:
Search volume is a vanity metric - AI helped me find keywords with "low" volume that actually convert better because they match specific user intent
Context beats competition scores - Understanding why someone searches matters more than knowing how many sites target that keyword
Intent discovery is the real opportunity - AI reveals search intent patterns that keyword tools completely miss
Content angles matter more than keyword density - AI suggests unique approaches to common topics that naturally rank better
Speed enables iteration - Fast research means you can test and adapt strategies instead of committing to rigid keyword lists
Semantic understanding beats exact match - AI groups related concepts that traditional tools treat as separate keywords
Integration reduces tool fatigue - One AI research session replaces multiple tool subscriptions and manual processes
The biggest mistake I see other marketers making is using AI as a keyword generator instead of a research assistant. They ask for "keyword lists" when they should be asking for user insights, competitive gaps, and content opportunities.
If I were starting this project again, I'd spend even less time on traditional validation and more time on understanding the nuanced ways people express their problems. The search volume numbers from traditional tools are often wrong anyway—better to optimize for relevance and user value.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS startups looking to implement AI keyword research:
Start with user problem mapping, not seed keywords
Use AI to understand customer journey touchpoints
Focus on intent-based content clusters rather than individual keywords
Validate AI insights with actual customer conversations
For your Ecommerce store
For ecommerce stores implementing this approach:
Research product-related problems, not just product names
Use AI to discover seasonal and trending search patterns
Map keywords to specific customer personas and buying stages
Focus on long-tail commercial intent queries