Eric Guy

With the rate of AI progression, it's hard to predict the future...but that didn't stop me. I asked Claude to predict how AI will impact marketers by the end of 2026. Here's what it gave me: 1) Martech stack gets a haircut When an AI agent can connect to your CRM, pull data, build segments, trigger campaigns, and optimize in real time... why do you need an automation platform? 2) "Build vs. buy" ends. Build wins. The speed at which you can create custom tools is even more disorienting. Claude Code will build internal marketing tools that would've taken weeks to implement. 3) Content strategy for robots (and humans) AI agents are already making recommendations and shortlisting options for decision-makers. If you're in B2B, a chunk of TOFU will be influenced by whether Claude, ChatGPT, or Perplexity recommends you. Your case studies, documentation, and product pages need to be structured so that LLMs can understand and cite them. 4) From "can you prompt" to "have you built" By end of year, the expectation won't be "can you write a decent prompt." It'll be "have you actually built and managed AI agent workflows that the needle" AI fluency will be the same baseline expectation as using Excel used to, except the gap between people who get it and people who don't is way wider. 5) Small teams embarrass large ones A 10-person agency with the right SaaS stack used to have a real advantage. Now 2 people with well-orchestrated AI agents will match the output and respond faster. BUT it's important to know > None of this fixes bad strategy. Faster execution of the wrong message to the wrong audience is still a waste

3

1 Comment

Most marketers hear "MCP" or "agentic workflow" and immediately tune out I get it, but you don't need to become an engineer overnight, you just need to know enough to be effective Here are some AI terms that ACTUALLY matter for marketers right now: MCP (Model Context Protocol) → Universal adapter that connects AI to your existing tools. Gmail, Sheets, your CRM. One standard connection instead of dozens of custom integrations. Agentic Workflows → AI that doesn't just answer questions, it takes actions. Research a competitor, draft the brief, build the landing page. Sequential steps, minimal hand-holding. Vibe Coding → Building functional tools by describing what you want in plain english. No "actual" coding. Plus, it sounds cool to say 😎 Context Vaults → Master doc you feed the AI so it understands your brand, audience, and history. This is the difference between generic AI output and something that sounds like it came from your team. Markdown Files (.md) → Text files that act as instruction manuals for your AI. Write your SOPs, guidelines, or briefs in markdown, and AI can reference them every time. Skills → Pre-built instruction sets you install into Claude Code to make it better at tasks. Ex: Copywriting, landing page design, competitive research. Training for AI instead of starting from scratch every time. Slash Commands → Custom shortcuts inside Claude Code to automate repetitive tasks. Type something like /weeklyreport and it summarizes performance and drafts the update. Cron Jobs → Automated tasks that run on a schedule. Ex: Eliminate the need to manually generate weekly reports. Self-Healing Loops → When an AI workflow hits an error, it diagnoses the issue, fixes the code, and updates its playbook. LLM (Large Language Model) → Brain behind tools like Claude and ChatGPT. When someone says "which LLM are you using?" they're asking which AI model is doing the thinking. Helps you understand why outputs vary between tools. API (Application Programming Interface) → The behind-the-scenes connection that lets two tools talk to each other. When your AI pulls data from Google Analytics or posts to your CMS automatically, that's an API doing the work. IDE (Integrated Development Environment) → The app where you interact with Claude Code. Tools like Cursor or VS Code. You don't need to be a developer to use one. Think of it as the workspace where you and the AI collaborate. You don't need to master all of this overnight. The more you experiment with #Claude Code the more these phrases will come up. ♻️ Repost if this is useful

5

1 Comment

EP 63: Zero Click AI Searches- The Reason Your Website Traffic is Down In this episode, Jenna Gaidusek dives deep into the growing crisis of zero-click search—a shift in how people find answers online that’s impacting small businesses, designers, and anyone relying on website traffic for visibility and leads. Jenna explains what zero-click search is, why it’s become a problem over the past two years, and how AI-generated summaries on platforms like Google, ChatGPT, Perplexity, Claude, and Gemini are changing the rules for SEO and business growth. What you’ll learn in this episode: Why it’s more important than ever to understand your website analytics—and how to get started even if you’re not “techy.” The reality behind declining organic traffic: 60% of searches now end without a click, and 25% of small business site visits have dried up. How designers can pivot with long-form, authority-building content, FAQs, and real world case studies that set you apart from generic AI content. The crucial role of backlinks, referrals, press releases, and partnerships for visibility in AI-driven search engines. The new ways clients search

6

Restaurant Money: Upselling How to make A LOT more money from customers you already have. How often do you track average spend per person? If you serve 1000 diners per week, and you can increase the average spend by just $2, then you would grow revenue by over $100k per year. If spend per person decreases by $2, you'll lose $100k a year in revenue. Ouch. Given its importance, we recommend you track average spend monthly and set KPIs. So how can you increase it? TLDR: Ask your customers if they want more stuff Long version: Involves maths, see below: A restaurant wants to increase average spend by $2pp. They create an upsell item which is valued at $20pp. (2 / 20) x 100 = 10% This only needs to work 10% of the time to achieve KPI. or They create an upsell item which is valued at $25pp. (2 / 25) x 100 = 8% This only needs to work 8% of the time to achieve KPI. Takeaway #1 The greater the value of the upsell, the less frequently it needs to work. What happens with multiple upsell opportunities? The same restaurant creates 2 upsell opportunities. Both are $20pp. The goal is still to increase average spend by $2pp. (2 / (20 + 20)) x 100 = 5% Each upsell now only needs to work 5% of the time on average to achieve the same result. Takeaway #2 The more upsells you have, the less often they need to work. Take this to the natural extreme, with 10 upsell opportunities, with an average of $20 per upsell. (2 / (10 x 20)) x 100 = 1% In this scenario, each upsell would only need to work 1% of the time, on average, to achieve an overall $2pp increase. I'm betting you can create upsell opportunities that work more frequently than just 1% Need an idea? Add your set menu, and premium version of the set menu, to the booking widget. Your customers can upsell themselves at the time of booking. No training required. Pro Tip: The SevenRooms widget has heaps of upselling features. Just ask Haidee Torning If you build a list of 10 upsell opportunities, implement this with your team, and track which upsells work the best, it would be, mathematically speaking, improbable NOT to increase average spend per person. Need more ideas? We have just added 20 upselling ideas to the SERV. Academy. https://lnkd.in/gThv6TM4 Tag your restaurant manager if they need extra support building their sequence of service around upselling opportunities.

11

What's been our winning B2B combo lately? Here's a playbook we've execute multiple 6 & 7-figs in pipeline via campaigns our omni Channel B2B campaign playbooks. We build a custom fractional marketing & rev ops team for the company; & work directly with their internal team members (i.e. their sales lead, head of Rev, marketing lead, etc) This works well bc it can be sprint focused & modular based on how your team is constructed & the size of your TAM. We've done this for Cannabis POS systems, expensive machinery, and CPG products in a high-risk space. This works best for teams that have a small (<10) sales team; where many are full cycle AEs; so having to do outbound / demand gen can be challenging. 1. Align on cohorts & message for your key buyer segments (who / why / how). You need an offer & it can't be how game-changing your software is. You have to add value in some way. Tech audit, no-risk 30 day consignment, HR fee audit, etc. 2. Stand up 15-20 inboxes (domains + addresses) & build 7-10 variations of outbound copy; you need a simple landing page & scale outbound email. This is used to get in front of customers. You need a TAM that is 10K+ here. We stand up shared slacks / sheets to keep comms flowing & jump on opportunities quickly (this is key; can't linger on cold leads). We'll use Clay to drive a fully enriched data list that aligns w/ these cohorts. 3. Stand up digital touches (Google + LinkedIn + Meta LinkedIn + Meta; or LinkedIn only); align campaigns using cohort data. Lot of creative early on; change early & often. Don't focus on conversions initially; focus on inbound lift. We use Fibbler for LinkedIn influenced attribution. For example; we can tie paid clicks back to deal creation date & show high degree (based on engagement levels) to deal creation to help tell story. This then ties back to CRM (Hubspot / SF). Depending on channel mix; more downstream variations can apply (for example retargeting Google Traffic; or YouTube rts, etc) 4. If applicable; send out a direct mailer. This has worked well for anyone with a retail store. Works well for service or CPG biz. Execute a series of 3-6 over a period of 2-4 months. Another tangible way to get in front of customers + can be customized (i.e. send to top 15 accounts). 5. Execute initial campaigns: Launch cold layers via digital > retarget traffic; cold emails to test winning offers; capture early hand raisers (opened email; high degree of ad engagement) & put into an outbound SDR cadence. 7-10 dials per contact to reach & have conversation; establish an acceptable "booked meet threshold" to be profitable & work against it. Test creative, test audiences, test hooks, etc. 6. Iterate creative based on winning offers; increase content exposure around these buckets (think Ebooks, webinars, silly videos, landing pages, etc). 7. Rinse & repeat until you work through your account list. We typically run initial tests on <30% of our total TAM to validate angles.

8

One of the most common mistakes I see restaurant brands make: treating every marketing dollar as an acquisition dollar. The data tells a different story. Repeat guests spend 67% more per visit than first-timers. Retaining a customer costs a fraction of what it takes to acquire a new one. And roughly 70% of first-time guests never return—not because of a bad meal, but because no one gave them a reason to come back. The second visit is the most important conversion in the customer journey. Yet most restaurants invest nothing in making it happen. At Evok, we've built our restaurant marketing practice around what we call the "Trial to Ambassador" framework—a systematic approach to moving guests up the value ladder instead of letting them drift away after a single transaction. Our team just published a guide breaking down the retention strategies that drive measurable results: the 7-day follow-up window, loyalty program design, behavioral segmentation, email and SMS tactics, and the metrics that actually matter. If you're a restaurant marketer tired of pouring budget into acquisition with diminishing returns, this one's worth your time. https://lnkd.in/dA6MNPEd

16

1 Comment

Renting your advantage is a short-sighted trap if you are not building your own AI-powered apps. When you build in-house you tap into a truly proprietary edge that no off-the-shelf tool can match. You leverage your unique catalog customer behavior and creator data to craft listings that reflect your brand voice and drive higher click through rates and conversions on platforms like TikTok Shop and Amazon. You win on speed and margin by automating work that drains hours so you launch faster with fewer stockouts while safeguarding your contribution margin without bloating headcount. You gain control and resilience when marketplace rules shift faster than a vendor roadmap so you protect your intellectual property and comply with live updates every day. Start with one painful workflow define one success metric and launch a small agent with human oversight then let the results drive your next build. Comment build if you want the blueprint.

2

1 Comment

Most GTM teams are using AI wrong. They treat it like a personalization tool. "Let's use Clay to customize 1,000 emails." "Let's scale outreach with Instantly." Then they wonder why response rates tank. Here's the problem: Personalization without value is just noise with your name on it. The teams actually getting replies? They're not optimizing for personalization. They're optimizing for permissionless value. What that looks like: → Use AI to identify specific problems prospects are facing right now → Lead with insights they can act on immediately → Deliver value before asking for anything Your outreach shouldn't need permission to be useful. It's not about sending more. It's not even about personalizing better. It's about leading with value so clear they'd be stupid to ignore it. The best systems I'm building: → Process signals to find real problems (not just "congrats on the funding") → Route prospects based on pain, not just title → Deliver actionable insights in the first touch Most people are still asking "how do we personalize at scale?" The ones winning are asking "how do we deliver value at scale?" Big difference. Which question are you asking?

7

3 Comments

If you’re still relying on Apollo exports or buying lists from vendors, you’re burning cash and nuking your domain reputation. Here’s the full, step-by-step workflow my team uses to: Build custom, intent-driven lead lists (zero recycled data) Enrich every contact with relevant data points like job changes, tech stack, and recent news/signals Hit 5%+ reply rates and 20%+ positive reply rates 1) Data Sourcing Start with Crunchbase/Apollo/Clutch/LISN filters (ICP, region, company size, tech used) Pull company + contact data with Clay (not just emails – grab LinkedIn, company news, funding, etc.) Use Apollo for the bulk of your prospect email contact info 2) Enrichment and Validation Run every contact through Clay for: Tech stack enrichment (BuiltWith integration) Recent company news (auto-scraped for personalisation) Validate emails with DeBounce before adding to your campaigns 3) List Cleaning and Segmentation Deduplicate by LinkedIn profile or email using Clay's native feature Segment by persona (for example: Topmen, Marketing Decision Makers) Remove any catch-alls or role-based emails 4) Campaign Launch Import to Instantly, Smartlead, or EmailBison (each support high-volume, multi-inbox sending) Personalise first/PS lines with AI (Claude prompt: “Write a custom opener/PS line for {FirstName} at {Company}, referencing their latest funding/news”). Try to make this as relevant to the email/value proposition as possible Set up a 2 or 3-step sequence with varied CTAs 5) Deliverability Infrastructure Use warmed domains (2+ weeks minimum) Warm up inboxes with EmailBison, Instantly, or Smartlead's in-built system Monitor bounce and spam rates daily to know when to rotate new inboxes in and old ones out

2

1 Comment