Inbal Mechoresh

After 250+ hours researching AI agents, I realized most agencies are missing the biggest opportunity in GTM automation. Here are the 8 categories actually worth using: 1️⃣ Agent Builders Relevance AI, Relay.app, n8n, Dify Automate your manual processes. We've saved hundreds of hours. 2️⃣ Scraping Agents Clay's AI Agent, Browserbase, Gumloop, Apify Extract data your competitors can't access manually. 3️⃣ MCP Servers Connect Claude to Slack, Notion, Gmail, Stripe Control all your work through one chat interface. 4️⃣ AI SDRs Artisan, Valley, 11x, AiSDR Actually decent outputs. Still need manual review. 5️⃣ GTM Co-Pilots Instantly.ai, Unify, Copy.ai, Bardeen Build entire campaigns just by writing prompts. 6️⃣ Sales Agents Attention, Attio, Gong, Momentum.io Gather insights from CRM and call transcripts. 7️⃣ Research Agents Claygent, Linkup, Airtop, Exa Deep account research at scale. 8️⃣ List Building Instantly.ai, Openmart, Exa Build prospect lists automatically. Most agencies use these in isolation, but we connect everything into our GTM flywheel. Enjoy this? ♻️ Repost it to your network and follow for more AI GTM content.

273

‏134‏ ‏תגובות‏

We just launched 247 location-specific pages in 7 days, fully AI generated. 140 are already ranking Top 3 and getting AI citations. Here’s the exact system we built (and why 90% of programmatic content fails): Most location-specific programmatic content is generic fluff with city names swapped in. With this type of content, you won't get any rankings, or traffic, and you shouldn't even think about leads. Google (and AI search) can spot this from miles away. But for one of our clients, we saw a big programmatic content opportunity that we didn't want to miss. So we took a completely different approach, and launched this as a content engineering project based on AirOps. I’m documenting this experiment publicly, because I strongly believe in this new approach of content engineering. Will share conversion data, traffic growth, and ranking stability over the coming weeks. _________________ Here are the exact 6 steps we took: 1) Building our "one step before" strategy We researched content opportunities "one step before" our client’s service. In our specific case, users need to understand local regulations before using the product. What's great about this cluster is that there's not a lot of noise on the SERP (like Google Maps or local service ads). 2) Creating a gold standard page We created one perfect page manually with deep research, proper structure, and genuine value for the target topic. This acted as our gold standard page for the location-adjust content. 3) Setting up deep research per location For each of the 247 locations, we set up an AirOps workflow that does a Perplexity Deep Research around the individual location. Pro tip: Name specific types of sources you find trustworthy to ensure the input for later content generation prompts is top notch. In our case this was official government sources or local regulatory databases. 4) Creating a really thorough system prompt We built a ~3,000+ character prompt that combined the vision for this specific content, the tone of voice and quality guidelines, Perplexity Deep Research input, and the gold standard page content. 5) Integrating automatic asset sourcing/creation We added steps for stock image research per location, integrated Infographics from other content reused, generated custom meta titles and descriptions, and added structured FAQs for every page. 6) Refining with Human-in-the-Loop controls A lot of people get lazy with AI, but we did a lot of continuous prompt refinement and spot checks throughout the process. Think about it like a factory process that has to be tweaked and needs thorough QA processes. _________________ Early Results (72 hours live): → 140 pages ranking Top 3 → 96 pages ranking Top 4–10 → First features in AI search citations → Traffic spike already visible _________________ Questions around the experiment? Drop them in the comments. 👇 PS. See a similar opportunity for your startup, but need support in implementing such a system? Reach out.

70

‏24‏ ‏תגובות‏

Are you also interested in showing up first when people ask questions to ChatGPT, Claude, Gemini, and Perplexity about the best brands out there in (X) industry? These models don’t crawl the web in real-time like search engines. They reference (-structured, -high-quality, -citation-ready) content that already exists in the datasets they were trained on or in content they scrape and summarize post-training. To get mentioned by LLMs, your content needs to follow a different set of rules. It's not about keyword density or schema markup. It’s about format, clarity, and semantic value. Think structured “best of” lists, first-person reviews with measurable results, brand-versus-brand comparisons, and FAQ-style entries that answer real user questions concisely. Add in clearly captioned visuals, downloadable tools, and frameworks: things that are easy to summarize, hard to misinterpret, and valuable enough to quote. The implications are massive. You can gain visibility even if your domain has low authority or limited backlinks. AI citations drive brand familiarity without requiring a click. Users might not visit your site, but they still see your name, absorb your insights, and recall you later. LLM Seeding levels the playing field. According to our research at Semrush, nearly 90% of ChatGPT citations come from content ranked outside of Google’s top 20 (Rachel Handley). This means that content structure, not search position, increasingly determines visibility. We're entering a zero-click world, where users ask a question and trust the AI to provide a complete answer. This shift requires a new mindset. Traditional SEO chased clicks. Now? Look for mentions. Traditional content aimed to rank. LLM-ready content aims to be remembered, summarized, and cited by default. The question is no longer, “How do I get more traffic?” It’s “How do I stay top of mind in a world where AI does the talking?”

32

‏18‏ ‏תגובות‏

SEO & AI Tip: Agentic Buying Is Now Open to Everyone, not just ChatGPT. The ability to complete purchases through AI agents just took a big step forward yesterday. OpenAI & Stripe announced yesterday the Agentic Commerce Protocol (ACP) and that it is now live and open to everyone to build with. That means Agentic checkout that OpenAI just announced for ChatGPT can be built into your own websites, applications, and agents. Think AI-driven transactions that actually complete purchases without needing to click through pages or rely on traditional UX flows. This changes what’s possible for e-commerce teams building AI into their products. Here’s what just became possible: 1 - You can add a shopping agent to your own site trained on your inventory, pricing, and order APIs. 2 - The user can interact via text, voice, or image and the agent can handle the transaction end to end. 3 - You don’t need to rebuild your store, ACP works with your existing infrastructure and a payment processor that supports it (Stripe is already supported). 3 - Any agent, from an on-site assistant to a mobile app, can initiate checkout in a standardized way with you still owning the customer and fulfillment Here is a potential example: Large E-commerce Site With a Visual Concierge Let’s say you run a major e-commerce site in fashion, furniture, or auto parts. A shopper lands on your homepage and uploads a photo of the item they want, maybe a jacket from Instagram, a sofa leg they broke, or a specific wrench that fits their mower. Your on-site AI concierge kicks in: - Converts the image to a vector embeddings to match it against your inventory - Verifies availability, size, and price - Suggests alternatives if the exact match is not in stock - Then calls your ACP checkout endpoint, passes a secure payment token, and places the order No need to browse, filter, add to cart and add payment information. Just ask for it and buy it. Why This Matters for SEO Optimizing your site content, product inventory, imagery, and anything that can help make Agents more easily find the right product for those searching via agents will directly impact how Agents find and recommend your products. Make sure its clear in your data, text and imagery who your ideal customers are for each product to help Agents better match them to the people they're acting on behalf of. Why This Matters for Brands Most brands think of AI assistants as a front-end tool for discovery or support. ACP flips that around. We know that speed to purchase has a huge impact on e-commerce transactions. This may make it even easier and faster for people to find what they want. With ACP: - You remain merchant of record - You control fulfillment, returns, and data - You open up your site to new shopping patterns: mobile agents, voice shopping, post-purchase reordering, concierge style product finders This protocol makes agentic commerce a real channel not a side project.

33

‏6‏ ‏תגובות‏

Remarketing spend hit $16.4B in 2024… 3.5× more than UA. During peak season, 85% of iOS budgets went to remarketing. Marketers are doubling down on LTV. But many still see influencer marketing as purely upper funnel. That’s a mistake. In the campaigns I run, influencers don’t just drive installs, they: ✅ Reactivate past users ✅ Trigger second purchases ✅ Reinforce brand value when it matters most Influencers can supercharge remarketing by: - Creating custom content for audiences who clicked an influencer’s affiliate link but didn’t purchase - Running Instagram Story ads targeting users who engaged with a sponsored post but didn’t follow or buy For example: If an influencer shares a skincare launch and people click through but don’t check out, you can later retarget them with a carousel ad featuring a time-limited discount or a relatable testimonial, right when they’re more likely to convert. 📣 My take: Creators don’t just drive discovery, they reinforce intent. If your remarketing mix doesn’t include influencer content, you’re leaving results on the table. 📊 Want to see the full picture? Get the latest data on UA trends, iOS loyalty, and why remarketing now dominates eCom budgets. Read AppsFlyer’s full 2025 State of eCommerce App Marketing report here: https://lnkd.in/e7_6kSQC

37

‏11‏ ‏תגובות‏

Where to place Google Ads tags when moving to server-side GTM? A client asked me this last week. Here's what I usually do 👇 Let's look at the most prevalent Google ads tags one-by-one. Conversion Linker tag ❌ Client-side container ✅ Server-side container Why? If you use both the client-side tag will be redundant. Especially since the Google tag is required for Google Ads since April. The server-side tag basically places a longer-lasting cookie which improves attribution accuracy. Remarketing tag ❌ Client-side container ✅ Server-side container Both still work only server-side gives you more control and robust attribution. Make sure you don't use both or you'll duplicate the data. This is often overlooked. Conversion event tags ✅ Client-side container ✅ Server-side container Usually, I would recommend using only server events. But for an event like a purchase, you can actually use a dual setup. Transaction IDs prevent double counting automatically. This ensures you measure most datapoints. When setting it up recommend creating new conversion in Google Ads and setting it to secondary until you know the data is accurate. This way it doesn't affect active campaigns. Happy analysis ✌ #GoogleAds #GTM #ServerSideTracking

82

‏15‏ ‏תגובות‏