This week’s #SEOForLunch sponsors are Profound and Jolly Search.
I lost a potential client to a $99 AI subscription. Here’s what the dashboard could not replace.
When you’ve spent years in SEO, you get used to losing clients.
You lose them to budget cuts, shifting internal priorities, redesigned sites that tank traffic before you can save them, or a new CMO who “has a guy.”
It hurts, but it’s business.
What hits differently is the first time you get replaced not by a competitor, but by a software subscription.
Thank you to Profound for sponsoring this week’s #SEOForLunch
Do Claude and Claude Code Cite Different Sources?
To better understand the similarities and differences between Claude and Claude Code, we analyze two separate datasets:
24,135 Claude and Claude Code responses across a set of prompts from 11 randomly sampled categories, including a coding subset
The top 1,000 webpages visited by Claude and Claude Code agents over a 30-day period, among a large collection of internally tracked domains
We found that Claude and Claude Code behave like different Answer Engines. In our sample where web search is enabled, Claude Code searches in 13% of responses, compared to Claude’s 93%. Even though they mention a similar number of brands in each response, those brands overlap by only 20% on average.
The Time I Lost a Client to AI Software
Earlier this year, a prospective client reached out to discuss working together.
We hopped on a call to talk through their goals, and they checked all the boxes: strong fit, clear expectations, and zero pushback on my pricing or process.
Before we hung up, they let me know they were speaking with a few other providers. I applauded them for doing their due diligence.
My approach to pitching is super direct (shocker, I know): no fluff, no games, no wild promises. That transparency makes me a great fit for some companies, but others simply prefer to be wined and dined.
After a second call to address a few follow-up questions, I received the dreaded email: “Thanks, but we went in another direction.”
Bummer, but it happens. That’s just business.
Except the email did not end there.
The “provider” they chose was not another consultant or agency. It was an AI tool that promised it all: automated keyword research, content briefs, technical audits, and execution for a fraction of my monthly retainer.
I won’t lie. That one stung.
But I wished them well and told them to come back if they needed further help down the road.
It was not just a lost contract. It was a front-row seat to an industry-wide shift: we are rapidly trading human expertise, nuance, and strategic context for the illusion of easy-button efficiency.
And as I soon realized, once an organization decides human judgment is an unnecessary expense, the way it treats human relationships can change just as quickly.
Breaking down the “AI promise”
To be clear, I am not anti-AI tooling. I use it, and I expect other good marketers to use it too.
It can accelerate research, organize data, build a first-draft brief, identify patterns across many pages, and reduce the manual work everyone hates.
That is the point: AI should make skilled people more efficient. It should not be treated as a replacement for the skilled person responsible for deciding what deserves to happen in the first place.
The promise made to this prospective client was not entirely wrong. The platform could likely help them find keywords, generate briefs, run technical checks, and produce more content than a consultant could manually create in the same amount of time.
But SEO has never been a race to complete the most tasks.
A platform can tell you a keyword has search volume. It cannot tell you whether that traffic will attract the right customer, support the company’s actual revenue goals, or divert attention from a much better opportunity.
It can suggest creating another page because the query looks distinct in a spreadsheet. It may not recognize that the site already has several pages competing for the same intent—or that the right answer is to consolidate existing pages rather than publish one more URL nobody needs.
It can flag 100 technical issues in an audit. It cannot walk into an engineering planning meeting and explain why five of those issues are worth prioritizing, 90 can wait, and the remaining five are not meaningful problems at all.
And it can produce content at scale. What it cannot do is understand the political, legal, product, sales, brand, or customer-service context that may make an otherwise “SEO-friendly” recommendation a terrible business decision.
That is where the easy-button pitch breaks down.
The tool can create activity. It can make a dashboard look busy. It can even generate pages that rank. But none of that means it understands the business, owns the downside, or knows when the standard SEO recommendation should be ignored.
The True Cost of Commoditizing Expertise
The ideal setup is a skilled person using AI to maximize efficiency, NOT using AI as an excuse to eliminate the person accountable for the work.
For whatever reason, the reasonable middle ground has become hard to discuss. Raise concerns about quality, risk, or strategy and you are labeled “anti-AI.” Suggest automating everything possible, and you are innovative. (but you might also be out of a job in 6 months.)
Neither label helps the P&L.
The question is not whether AI can do work faster. Of course it can. The question is whether doing more work, more quickly, actually creates a better outcome for the business.
Ten years ago, publishing more content may have been a viable growth strategy (long live programmatic SEO). In 2026, turning five thoughtful articles into 50 mediocre ones at half the cost is not automatically an efficiency win. It may just be a faster way to fill your site with content that does not earn attention, trust, links, qualified traffic, or revenue. (Start with getting indexed for longer than a few weeks; that will be harder than you think.)
That is the easy-button trap: confusing lower production costs with better business results.
When companies reduce SEO to tasks like keyword research, briefs, audits, reports, and the creation of pages per month—they make it easy to compare a consultant against software. Software will almost always win that comparison on price and volume.
But you were never paying the consultant to open Semrush or Ahrefs, build a spreadsheet, or hand you a technical audit PDF.
You were paying for the judgment behind it all: what to prioritize, what to ignore, what creates risk, and what needs to change when the original plan is not working. You were paying for someone to own the recommendation, not simply check a box confirming the work was completed.
The Transition: From Code to Culture
Replacing a consultant with an AI tool is not automatically a bad decision. In the right situation, a good tool can absolutely help an internal team move faster.
But when every person becomes a cost center, and every relationship becomes an input-output transaction, something much bigger gets lost.
Once human expertise becomes just another expense to trim, it becomes a lot easier to treat the human relationship behind it as disposable too.
That mindset does not stay contained to SEO strategy or marketing budgets. It shows up in how companies communicate with agencies, freelancers, employees, and even candidates looking to join the organization.
Or, more often, how they avoid communicating with them at all!
Keep an eye on your inbox this Thursday.
In Part 2: The Ghosting Epidemic, I’ll cover more about what happens when this same automated, transactional mindset extends beyond the work itself….
… Think: clients disappearing mid-contract. Prospects going silent after asking for a proposal. Job seekers getting pushed through ATS pipelines only to be met with absolutely nothing.
Because apparently, basic professional courtesy is now optional.
Turns out keeping a “human in the loop” is only important when someone needs to take the blame. Everything else, though? AI it!
~Nick
Thank you to Jolly for sponsoring this week’s #SEOForLunch
Perplexity Cited Our Client on Day 1. Google Followed on Day 3.
Brought to you by Jolly, the team behind Own A Prompt
The most persistent myth in AI search: that it’s slow — that getting cited takes months of “authority building” before anything moves.
So we timed a real campaign, for a real client, and published every screenshot:
Day 1: our Reddit thread goes live — Perplexity starts citing it
Day 3: #1 organic ranking in Google, and AI Mode starts recommending the client by name
Day 19: all 7 major AI surfaces — ChatGPT, Gemini, Claude, Perplexity, Grok, AI Overviews, AI Mode — are recommending them
Total content required: one Reddit thread and one onsite blog post.
AI search doesn’t reward patience. It rewards the right evidence in the right places.
Dashboards don't answer for results.
The tool can't own the outcome.