How I Use an AI Agent to Get Competitor Intel + A Strategy to Beat Them
Inside my Manus workflow for competitive intelligence
The AI world is all Antropic right now: Claude Code, Claude Cowork, and Clawdbot. And I’ll be covering use cases for all of these in future posts, but this week I want to cover Manus.
I’ve slept on Manus. I covered it on Marketing Against the Grain when it originally launched (fun fact, I was one of the first to cover Lovable, back then it was called GPT Engineer) - it was cool, but to me, it was another agent that overpromised.
Well, recently they announced a partnership with SimilarWeb, giving users access to their data right from within Manus, and I was blown away.
Manus is really an ‘action agent’. It can spin up a cloud computer, browse the web, write code, manipulate files, and do pretty incredible data visualization.
Manus was recently bought by Meta for a reported $2B-$3B.
Over the past week, I started perfecting some powerful use cases. And, I’m going to give you some of the best ones I’ve found. A quick note, all these were run with Manus PRO. You can get a free 7-day trial to test them out.
1. Competitor Research
This is one of the most common use cases for an AI agent like Manus among marketers/founders and growth professionals. Do a deep dive on a competitor and show me how they’re growing. What makes this better is that we have direct access to SimilarWeb data + we Manus is great at reasoning so we can turn this analysis into action e.g. how do we disrupt our competitor based on these trends.
First, here’s the output so you can see how good it is. Here’s a link to the PDF it created. I ran it for OpenAI. The purpose is not just to analyze a competitor’s digital marketing trends, but also to identify weaknesses you can exploit.
PROMPT:
You’re a seasoned growth strategist and data analyst. Our CMO has commissioned
deep competitive intelligence research on [COMPETITOR NAME] to reverse-engineer
their exact growth playbook and identify exploitable weaknesses.OBJECTIVE
Conduct forensic analysis of [COMPETITOR NAME]’s growth strategy using:
- Real-time traffic data from SimilarWeb API
- Public commentary, news, and strategic announcements
- SEO and content strategy analysis
- Partnership and integration ecosystem mapping
- Funding and M&A activityRESEARCH FRAMEWORK
1. TRAFFIC FORENSICS
- Pull 12-month SimilarWeb data (desktop + mobile traffic sources)
- Identify which channels are growing vs. declining
- Calculate growth rates and absolute changes
- Map traffic composition shifts over time
2. GROWTH TACTIC IDENTIFICATION
- For each growing channel, hypothesize the specific tactics driving growth
- Look for correlation with funding rounds, product launches, or market events
- Identify what they’re doing that competitors aren’t
3. STRATEGIC CONTEXT
- Recent funding rounds and capital deployment strategy
- Geographic expansion priorities
- Product evolution and positioning shifts
- M&A activity and integration strategy
- Stated strategic goals from leadership interviews4. VULNERABILITY MAPPING
- Which channels are underperforming or declining?
- Where are they overspending to compensate for weaknesses?
- What market segments are they neglecting?
- Where is their brand positioning creating blind spots?OUTPUT REQUIREMENTS
Deliver a 2-page executive report with:
1. EXECUTIVE SUMMARY
- 3-4 sentence overview of their core growth strategy
- Total traffic growth and key metrics
- Primary strategic focus areas2. GROWTH PLAYBOOKS (Clustered by Theme)
- Group related channels into strategic playbooks
- For each playbook:
* Channel performance data (growth %, absolute changes)
* Tactical commentary on HOW they’re driving growth
* Confidence score (High 90%+, Medium 70-89%, Low <70%)
- High: Backed by hard data + multiple sources
- Medium: Strong data + logical inference
- Low: Logical hypothesis + partial data
3. THE ANTI-[COMPETITOR] PLAYBOOK
- 3-5 specific strategies to disrupt their growth
- Focus on exploiting weaknesses, not outspending
- Prioritize asymmetric advantages
- Include quick wins and long-term playsFORMATTING
- Use direct, no-BS language for C-suite audience
- Lead with insights, not explanations
- Use tables for data comparison
- Include confidence scores on all assessments
- Bullet points for tactics, paragraphs for strategic narrative
- 2 pages maximum (excluding charts)TONE
Write like a top-tier strategy consultant presenting to a board. Be confident,
data-driven, and actionable. No fluff. No obvious statements. Every sentence
should deliver value.
What’s even cooler is that you can instantly turn a report like this into an interactive dashboard. See here. It’s pretty incredible. Imagine how much more your CEO will be impressed by this vs. a PowerPoint deck.
You can build on this competitor research in really interesting ways, for example:
Paid Ads Audit: “Pull their last 30 days of active ads from Meta Ad Library and LinkedIn Ad Library. Categorize by creative type (static/video/carousel), hook pattern, and CTA. Flag any new campaigns or messaging shifts.”
I did this for Monday.com because OpenAI isn’t running any real paid ads. It’s pretty wild how good it is. Report here.
Lastly, to continue the theme of competitor analysis, one way to understand where your competitors are investing across their marketing/growth strategy is to look at who they are hiring:
Hiring Audit: “Check their careers page and LinkedIn jobs. List open roles by function. Flag any unusual hiring clusters — these signal where they’re investing ahead of announcements.”
I ran this for OpenAI; it’s actually an incredible way to break out a company’s business strategy. See here, incredibly useful.
2. Digital Marketing & Growth Trends
If you know what trends you’re looking for, it’s incredibly powerful.
a. Organic Disruption and Paid Supplement
We know SEO is being disrupted. I want to know who in my vertical is losing organic traffic and if they are supplementing it with an increase in paid spend. If they are; I’ll know paid advertising CAC is going to increase.
PROMPT:
Here’s a list of HubSpot competitors. For each one, pull 12-month organic traffic trends from SimilarWeb. Flag any that dropped organic traffic by more than 15%. For those, check if paid traffic or ad spend increased in the same period. Output a table: Company, Organic Trend, Paid Trend, Verdict (Compensating Y/N), and a one-line takeaway on what it signals.
Competitors:
salesforce.com
activecampaign.com
pipedrive.com
zoho.com/crm
marketo.com
klaviyo.com
freshworks.com
keap.com
Look how incredible this interactive dashboard is:
Provides a detailed breakout of individual competitors.
These kind of custom interactive dashboards make it much easier to collaborate with your team. It’s easy to ask Manus to allow you to add annotations that you can save to the dashboard.
b. Find Partnerships
With the disruption in SEO and rising Paid Advertising CAC, partnerships are a realistic growth channel.
OK, the credits are real on Manus, I’ve blown through an insane amount of credits, and I had to upgrade twice already for this post :) So, unfortunately, I can’t show the remaining use case, or I’d have to sell my house. The credits are a REAL pain, but the research quality is pretty incredible. For in-depth research, which needs to be high-quality, it’s worth using Manus. For everything else, it’s too expensive. Below is the prompt I’d use for this use case.
PROMPT:
Here’s a list of AI-native CRM and GTM competitors. For each one, pull 12-month referral traffic sources from SimilarWeb. Flag any that show a new referral source appearing in the last 6 months that now drives 5%+ of traffic. Output a table: Company, New Referral Source, Traffic Share %, First Appeared, and a one-line hypothesis on what it signals (integration launch, co-marketing deal, acquisition channel, affiliate program).
AI-Native Competitors:
attio.com
clay.com
apollo.io
instantly.ai
folk.app
breakcold.com
warmly.ai
That’s a wrap. Yes, Manus is much more expensive vs. ChatGPT/Claude, but worth it for in-depth research use cases.
Until Next Time,
Happy AI;fying
Kieran
P.S - for those who keep asking, yes, Claude Code is coming for Marketers and Growth professionals. Working on some great use cases.
This. AI done right turns hours of research into minutes with better insights.
This is wild. I'll enjoy exploring!