Direct Answer
Claude Opus 5 connected to DataForSEO can replace five common SEO analysis workflows, but it does not make Semrush or Ahrefs obsolete. The useful change is architectural: live search data becomes an input to a reasoning agent that knows the business, prioritizes actions, and returns a readable report instead of another export to interpret.
For solo consultants, founders, and small marketing teams, that can remove a meaningful amount of software and reporting work. For agencies and larger teams, mature platforms still provide valuable proprietary indexes, history, rank tracking, alerts, shared dashboards, stable exports, and governance.
Watch Nico's Walkthrough
Video, transcript, demonstration, and original prompt-pack credit: Nico | AI Ranking. Nico's free source pack supplied the five Markdown prompts. This article organizes and adapts them for publication, checks the commercial claims against current official sources, and adds cost, permission, and verification controls.
The $1.9 Billion Claim, Checked
The headline is directionally correct. Adobe announced an all-cash Semrush acquisition valued at approximately $1.9 billion. Adobe's later quarterly filing records the completed acquisition on 28 April 2026 at $1.87 billion, primarily cash.
That valuation is not the price of five screens. It includes Semrush's brand, customer base, distribution, proprietary datasets, crawling and keyword infrastructure, product history, integrations, employees, recurring revenue, and enterprise workflows. Claude can rebuild selected outcomes without recreating the company behind them.
| Video claim | Source-checked reading | Practical conclusion |
|---|---|---|
| Adobe paid $1.9B cash for Semrush. | Adobe's filing records $1.87B, primarily cash, after closing. | The rounded headline is fair; the exact completed figure is $1.87B. |
| You can rebuild Semrush or Ahrefs. | You can rebuild five frequently used analytical workflows, not their complete data and operating systems. | Compare workflows and accepted outputs, not feature-count slogans. |
| The agent is better because it understands the business. | This is plausible when the brief contains real business context and the API evidence is preserved. | Business context improves prioritization; it does not repair missing or incorrect data. |
| The reports cost only cents. | Some bounded calls are inexpensive, but price varies by endpoint, method, priority, and rows. | Log actual usage before promising a recurring report price. |
What the Agentic SEO Stack Actually Replaces
Nico's workflow has five layers. Keeping them separate makes the system easier to debug and prevents a polished report from hiding weak evidence.
| Layer | Role | Failure to watch |
|---|---|---|
| Business brief | Defines audience, offers, markets, priorities, and constraints. | Generic context produces generic recommendations. |
| Claude Opus 5 | Plans tool calls, interprets evidence, clusters findings, and writes the report. | It may infer beyond the returned data or stop before verifying every claim. |
| DataForSEO MCP | Exposes live search, keyword, backlink, on-page, and domain data to the agent. | Paid calls, unsupported modules, asynchronous jobs, and incomplete rows need handling. |
| Report contract | Turns analysis into a consistent self-contained HTML decision brief. | Good visual design can create false confidence if sources and uncertainty are omitted. |
| Human review | Checks relevance, evidence, economics, and production changes. | Skipping review turns recommendations into ungoverned site edits. |
The five workflows are site exploration, keyword research, competitor content gaps, technical triage, and sampled AI search visibility. The agent's strongest contribution is not collecting rows. It is deciding which rows matter to this business and explaining what to do first.
Current Pricing Check
Pricing changes frequently. These figures were checked against official pages on 31 July 2026 and should be verified again before purchase.
| Product | Published position | Important detail |
|---|---|---|
| DataForSEO | Pay-as-you-go, $1 trial credit, $50 minimum deposit. | No separate MCP fee. API cost varies by endpoint, retrieval method, priority, and result volume. |
| Semrush AI Visibility | $99/month per domain on the annual Base plan. | Published plan includes daily tracking of 25 custom prompts and several AI platforms. Packaging can change. |
| Ahrefs Brand Radar | $199/month for one AI platform index; $699/month for all platforms. | Custom prompt tracking and regional pricing can add or change cost. Some plan entitlements differ. |
| Claude | Requires an account or API path with access to the selected model and connector workflow. | Your real cost includes the subscription or tokens plus DataForSEO calls and review time. |
The correct comparison is cost per accepted decision or completed change, not subscription price alone. A $20 report that prevents a week of low-value content is cheap. A $0.20 report built on the wrong country, stale rows, or invented reasoning is expensive.
Connect DataForSEO Safely
- Create a DataForSEO account. Nico's link is an affiliate link; DataForSEO advertises a small trial credit.
- In Claude Desktop, open Customize, then Connectors, and choose Add custom connector.
- Name it DataForSEO and paste the official MCP URL:
https://mcp.dataforseo.com/mcp. - Complete DataForSEO's OAuth authorization and start a fresh Claude conversation.
- Ask Claude to list the available DataForSEO tools before running a report. Confirm that the module needed by the chosen prompt is available.
- Keep tool approval on ask each time for the first runs. Do not switch immediately to blanket approval for paid API calls.
- Set a ceiling in the prompt: Before paid calls, state the expected endpoints and stop for approval if estimated cost exceeds $2.
- Run one domain, country, and language first. Check the API usage log and the report's source footer before increasing scope.
Shared Report Contract
The supplied pack repeats one design system across all five prompts. Use it once as a reusable Claude Project instruction or append it to each prompt. This version preserves Nico's structure and adds evidence, cost, accessibility, and decision controls.
| Supplied Markdown file | Published section |
|---|---|
README.md | Connector setup, operating notes, and limitations. |
00-DESIGN-SYSTEM.md | The shared report contract below. |
01-site-explorer.md | Prompt 1: organic visibility and page priorities. |
02-keyword-research.md | Prompt 2: intent, clusters, and build order. |
03-content-gap.md | Prompt 3: commercially filtered competitor gaps. |
04-technical-audit.md | Prompt 4: asynchronous crawl and action triage. |
05-ai-visibility.md | Prompt 5: sampled mentions, citations, and source gaps. |
SHARED REPORT CONTRACT
Return one self-contained HTML file with inline CSS.
Do not use external fonts, CDNs, images, JavaScript libraries, or gradients.
VISUAL SYSTEM
- Page: #1a1a19
- Cards: #232322
- Primary text: #ffffff
- Secondary text: #c3c2b7
- Muted text: #83827a
- Rules: #33332f
- Accent: #3987e5
- Warning: #e0a03c
- High risk: #e66767
- Positive: #199e70
- Use a clean sans-serif system font stack.
- No pie charts, donut charts, 3D charts, decorative gradients, or tiny labels.
- Tables: maximum 25 rows, no vertical gridlines, readable on mobile.
- Every severity state must include a text label, not color alone.
REPORT STRUCTURE
1. Header: report title, domain/topic, market, and generated date.
2. Three to five hero metrics with exact definitions.
3. Main findings with evidence, compact tables, and simple horizontal bars.
4. "The 3 things I would do first" with owner, effort, expected effect,
confidence, and verification step.
5. Method footer: APIs/endpoints used, row counts, time range, country,
language, assumptions, missing data, and estimated/actual API cost.
EVIDENCE RULES
- Every number must come from returned API data.
- Preserve the source endpoint beside consequential claims.
- If data is unavailable, say "not available"; do not estimate silently.
- Label estimates and low-confidence inferences.
- Do not pad the report with generic SEO advice.
- Do not recommend a production change without a verification step.
- Before paid calls, state the intended endpoints.
- Stop for approval if estimated API cost exceeds $2.
- Do not publish or modify the live site in this task.
Save as: seo-report-[workflow]-[domain-or-topic]-[YYYY-MM-DD].htmlPrompt 1: Site Explorer
Use this for a prioritized view of current organic visibility. It is strongest when you need to choose pages to improve, not when you need a complete proprietary backlink history.
Build a source-backed organic site explorer report.
INPUTS
DOMAIN: [example.com]
COUNTRY: [United States]
LANGUAGE: [English]
TASK
1. Pull every available ranking keyword for the domain with:
keyword, current position, monthly search volume, and exact ranking URL.
2. Summarize keyword counts in positions 1-3, 4-10, 11-20, and 21-100.
3. Identify the top 20 pages by estimated organic traffic.
4. Choose the three pages with the highest realistic 90-day upside.
5. For each priority page, show:
- target keyword
- current position
- monthly volume
- ranking URL
- evidence-backed reason it may be stuck
- the single first change to test
- success metric and review date
6. Rank opportunities by expected traffic gain, not volume alone.
7. Show the calculation or method used for estimated traffic gain.
8. Flag uncertainty and missing values explicitly.
Use the SHARED REPORT CONTRACT.Review before acting: a ranking decline can come from intent change, seasonality, SERP features, cannibalization, indexing, links, competitors, or measurement changes. The agent should inspect the relevant SERP and page before claiming one cause.
Prompt 2: Keyword Research
This workflow uses live keyword and SERP data, then adds business relevance and page clustering. That is more useful than handing a writer a thousand loosely related terms.
Build a business-relevant keyword research and content architecture report.
INPUTS
SEED TOPIC: [topic]
BUSINESS: [one sentence describing the offer and target customer]
COUNTRY: [United States]
LANGUAGE: [English]
TASK
1. Expand a relevant keyword set using live API data.
2. Include monthly volume, difficulty/competition metric, and source endpoint.
3. Classify each term by intent:
informational, commercial investigation, transactional, or navigational.
4. Cluster keywords that should be answered on the same page.
5. Give every cluster:
- recommended page title
- suggested slug
- primary term
- supporting terms
- dominant intent
- recommended page format
6. Identify weak SERPs only after checking the actual result set.
7. Create a build order based on business relevance, demand,
realistic difficulty, existing authority, and conversion path.
8. Exclude irrelevant or unserviceable terms and explain why.
9. For the top five clusters, define one measurable 90-day outcome.
Use the SHARED REPORT CONTRACT.Keyword difficulty is a vendor-specific proxy, not a law of nature. Keep the metric's source visible and combine it with the actual ranking pages, domain capability, intent fit, and the quality of the proposed answer.
Prompt 3: Competitor Content Gap
A useful content gap is not every keyword a competitor ranks for. It is a commercially relevant question that at least two competitors answer, your site does not, and your business can answer distinctly.
Build a commercially filtered competitor content-gap report.
INPUTS
DOMAIN: [example.com]
COMPETITOR 1: [competitor-one.com]
COMPETITOR 2: [competitor-two.com]
COMPETITOR 3: [competitor-three.com]
COUNTRY: [United States]
LANGUAGE: [English]
TASK
1. Find keywords where at least two competitors rank in the top 20
and the target domain is absent.
2. Remove competitor-branded, irrelevant, and unserviceable terms.
3. For every retained gap, include:
- monthly search volume
- best competitor position
- ranking competitor URL
- observed page format
- dominant intent
4. Cluster gaps into pages rather than producing one page per keyword.
5. Prioritize clusters using:
traffic opportunity, commercial intent, business fit,
existing site authority, and SERP beatability.
6. For the top five pages, inspect the ranking results and propose
a specific differentiator based on evidence, not generic "better content."
7. Give the writer a build brief with:
audience, promise, outline, proof required, internal links,
conversion path, and acceptance criteria.
8. Separate new-page opportunities from improvements to existing pages.
Use the SHARED REPORT CONTRACT.A gap report should not become a copy plan. The differentiator may be original data, a calculator, a clearer process, expert evidence, first-party examples, better product integration, or a genuinely different point of view.
Prompt 4: Technical Audit
DataForSEO OnPage crawls can be asynchronous, so the agent must start the job, poll responsibly, and wait for completion before writing conclusions. The objective is triage, not the largest possible warning count.
Build an evidence-backed technical SEO triage report.
INPUT
DOMAIN: [example.com]
TASK
1. Run the required OnPage crawl and wait for the asynchronous job to finish.
2. Record crawl scope, robots behavior, user agent, pages attempted,
pages completed, and any crawl limitations.
3. Collect all technical findings, then filter out cosmetic warnings.
4. Keep only issues likely to affect crawling, indexing, rankings,
accessibility, trust, or conversion.
5. For each retained issue, provide:
- severity label
- evidence and affected URLs
- likely impact
- exact proposed change
- estimated fix time
- verification method
6. Group recommendations into:
fix this week, fix this month, and monitor/eventually.
7. Name the single highest-impact fix and explain the expected effect.
8. Explicitly note warnings that look serious but probably are not.
9. Do not present text-to-HTML ratio, missing llms.txt, or a generic
word-count threshold as a ranking issue without direct evidence.
10. Do not change robots.txt, canonicals, redirects, templates,
structured data, or production content in this task.
Use the SHARED REPORT CONTRACT.llms.txt file is optional and is not an established Google ranking requirement. Verify status codes, canonicals, indexing, internal links, crawl paths, structured data, and real user impact first.
Prompt 5: AI Search Visibility
This report measures a declared sample of buyer questions and AI systems. It is useful for discovering citation patterns and missing source material, but it is not a universal score for everything an assistant may say about a brand.
Build a sampled AI search visibility and citation report.
INPUTS
BRAND: [brand name]
DOMAIN: [example.com]
COMPETITOR 1: [competitor-one.com]
COMPETITOR 2: [competitor-two.com]
MARKET: [country and language]
PRE-FLIGHT
1. List the available DataForSEO AI Optimization / LLM tools.
2. If the required module is unavailable, stop and explain the missing access.
3. Declare the AI engines, query set, date, location, and retrieval method.
TASK
1. Build or validate a representative set of buyer questions.
2. Show where the brand is cited, mentioned without citation, or absent.
3. Identify competitor citations and the exact source pages used.
4. Compare cited competitor pages with the target site's relevant pages.
5. Explain what the cited sources contain that the target site lacks,
separating observation from inference.
6. Recommend the three site changes most likely to improve eligibility
for citation, with proof required and a measurement plan.
7. Report:
- number of prompts tested
- engines tested
- mentions and citations by engine
- source domains and pages
- date and market
8. State "no data" when a query or engine returns no usable result.
9. Do not convert a small sample into a universal visibility percentage.
Use the SHARED REPORT CONTRACT.The best actions usually improve the underlying source, not merely its AI wording: clear answers, verifiable claims, first-party evidence, stable entity information, useful tools, strong author or company provenance, and crawlable pages that other sources can reference.
What This Does Not Replace
| Capability | Agentic stack | Mature SEO platform |
|---|---|---|
| One-off decision brief | Excellent when the prompt, data, and report contract are strong. | Usually exports data that still needs interpretation. |
| Business-specific prioritization | Strong with a detailed brief and human correction. | Often generalized unless custom workflows are built. |
| Deep proprietary backlink index | Depends on the connected provider and endpoint. | A core advantage of established platforms. |
| Years of normalized history | Must be stored and maintained separately. | Built into the product and account. |
| Scheduled rank tracking | Possible, but requires automation, storage, retries, and monitoring. | Ready-made with alerts and comparisons. |
| Shared client dashboards | Requires a reporting layer and access controls. | Designed for teams, seats, exports, and recurring delivery. |
| Repeatable governance | You own the permissions, logs, schemas, and QA burden. | More standardized, though still not error-free. |
| Taking action | Can draft content and changes after review. | Usually stops at analysis or integrates with separate tools. |
Nico's original pack is unusually honest about these limits: it does not claim to reproduce deep backlink history, years of stored data, scheduled tracking, or a shared team dashboard. Keep that honesty in the buying decision.
Who Should Use Which Approach?
| User | Best starting point | Reason |
|---|---|---|
| Founder with one site | Claude + DataForSEO, one workflow at a time. | Low software overhead and direct prioritization can beat a complex dashboard. |
| Solo SEO consultant | Hybrid stack. | Use agents for analysis and reports; retain a platform when index depth or history matters. |
| Small content team | Keyword, gap, and page-priority prompts first. | These workflows turn research into a manageable publishing queue. |
| Agency with many clients | Platform plus controlled agent layer. | Recurring tracking, permissions, templates, and client delivery justify stable infrastructure. |
| Enterprise SEO team | Existing data warehouse and tools, with agents above them. | Governance, procurement, history, auditability, and cross-team ownership dominate. |
A Seven-Day Rollout
- Day 1: connect DataForSEO with per-call approvals, list available tools, and set a $2 run ceiling.
- Day 2: run Site Explorer on one domain and manually verify ten consequential rows.
- Day 3: run Keyword Research for one commercial topic; remove irrelevant clusters by hand.
- Day 4: compare three true search competitors, not merely three famous brands.
- Day 5: run the technical crawl and validate the top issue in Search Console, server logs, or the site itself.
- Day 6: test 10 to 25 buyer questions for AI visibility and preserve the exact query set.
- Day 7: compare total API cost, review time, accepted recommendations, and decisions made against the current tool.
pilot value =
(hours of analysis avoided + value of better decisions)
- (API cost + review time + implementation risk)Continue only when the agent produces reproducible evidence and useful decisions. A beautiful report is not the success metric.
Video Map
| Time | Chapter | What to watch for |
|---|---|---|
| 00:00 | Adobe paid $1.9B for this | The replacement claim and the five selected workflows. |
| 01:40 | Connect DataForSEO | Account, connector URL, OAuth, and permissions. |
| 04:06 | Site explorer | Rankings, pages, and three prioritized changes. |
| 06:42 | Keyword research | Intent, relevance, clusters, and a visual report. |
| 09:03 | Content gap | Competitor opportunities filtered through the business. |
| 11:25 | Technical audit | Action triage rather than an unfiltered warning list. |
| 12:40 | AI visibility | Mentions, citations, competitor sources, and next actions. |
Bottom Line
The strongest idea in Nico's video is not that a Claude subscription destroys a $1.9 billion SEO company. It is that an SEO workflow no longer has to end in a spreadsheet. Live data can flow into an agent that understands the offer, explains uncertainty, prioritizes three actions, and prepares the next artifact.
That is a real improvement. It becomes durable only when teams preserve the source rows, control spend, verify recommendations, separate reports from production changes, and retain specialized platforms where history, index depth, monitoring, or collaboration justify them.
Sources and Useful Links
- Nico | AI Ranking: Opus 5 Just Destroyed a 1.9 Billion Dollar SEO Tool
- Nico | AI Ranking on YouTube
- Nico's original five-prompt download page
- DataForSEO: official MCP setup
- DataForSEO: current API pricing
- DataForSEO: AI Optimization API overview
- Adobe Q2 2026 filing: completed Semrush acquisition
- Semrush: AI Visibility pricing
- Ahrefs: Brand Radar pricing and coverage