On this week's "Good Enough Isn't" podcast episode, I told Artem Koren, the co-founder of Sembly AI, that when we pitch a new website now, we don't hand over a deck. We just build the site and hand that over instead.
I've used that line before, but it landed with Koren because he spent a decade in management consulting before he built Sembly, an AI-powered meeting intelligence platform that turns conversations into structured, client-ready outputs. He knows exactly why a McKinsey or Goldman deck is always flawless. The deck used to be the proof of competency. If a deal was worth a billion dollars, the deck had to look like an organization that could handle a billion dollars. Perfect slides signaled a perfect process.
AI tools like Sembly have broken that game. Now you can feed AI a client, a goal, a brand, and it builds the deck: sourced, styled, and on-brand in minutes. As I say in the episode: a $5,000 project can look like a $30,000 presentation now.
So the "deck quality equals high-caliber organization" signal is gone.
A flawless deck doesn't prove anything about the people who made it anymore. Truthfully, it never really did, but it was the closest stand-in we had.
What's left to judge is what's always mattered: the judgment and taste supporting the work. Koren said AI will get better at that too, eventually, but today is not that day.
That's why we stopped pitching sites and started shipping them. It's an invitation to judge the work, which is really the only remaining quality signal.
Now for this week's updates…
Patrick Patterson
CEO at Level Agency
What’s interesting this week
Level made Adweek's 2026 Fastest Growing Agencies list This one feels good. Adweek ranked Level No. 28 of 72 agencies worldwide and No. 4 among agencies our size, on a list built from certified financials. It's a big moment for our team, and a bigger one for the clients who grew with us and the new partners who chose us this year. Read the full announcement, and if you're one of those clients, thank you.
Key takeaway: Growth like this is an outcome of client results, and the numbers behind it are all audited.
Reddit opens AI-powered Max campaigns to every advertiser Reddit's Max campaigns, in limited release since January, now handle targeting, creative selection, placements, and budget allocation for any advertiser. Beta advertisers saw 17% lower CPAs and 27% more conversions on average, per Reddit. New asset groups let you separate product lines inside one campaign and report on each.
Key takeaway: If Reddit sat on your test list because of the learning curve, that reason just got thinner.
LinkedIn is the second most cited source in AI answers, and individual profiles do most of the work Semrush and Meltwater data compiled by HubSpot show LinkedIn trails only YouTube as a cited source across ChatGPT, Perplexity, and Google AI Mode. Google AI Mode cites it in 14.3% of responses. The detail that matters for your leadership team: 75% of LinkedIn citations come from individual profiles, and 51% come from accounts under 10,000 followers.
Key takeaway: Your SMEs posting on LinkedIn now counts as AI visibility work.
New this issue: a running list of the events and sessions where you can catch the Level team, virtually and in person.
Mastering the New Era of Search: How Brands Win Customers in an AI-Led Buyer Journey Thursday, Sept. 17, 1 PM ET · Virtual, hosted by Search Engine Land
Brad Stephenson, Stacey Heubeck, and Ernie DeCoite join Search Engine Land's Danny Goodwin to lay out one framework for organic, paid, and AI visibility. The starting point: 51% of B2B buyers now begin research inside an AI chatbot.
AI-powered media buying, put to the test (Virtual Insider Session with PubMatic) Tuesday, Sept. 22, 1 PM ET · Virtual, 30 minutes plus a 15-minute live Q&A
Ernie DeCoite and Anjlee Majmudar walk through Level's head-to-head test for American Career College. PubMatic's agentic supply-side platform delivered over double the qualified reach per dollar and cut campaign setup from hours to minutes. PubMatic's Harry Tong joins them for the live Q&A. Attend live for a chance to win Apple AirPods Pro 3!
MAICON 2026 Oct. 13 to 15 · Huntington Convention Center, Cleveland, OH
Level is presenting and sponsoring at the Marketing AI Institute's flagship conference. The speaker list includes Andrew Yang, Karen Hao, Kevin Roose, and Paul Roetzer. If you're in Cleveland, find us on the floor. Use code LEVEL200 for $200 off conference registration.
Digital Marketing for Financial Services New York Summit Oct. 28 to 29 · New York Hilton Midtown, New York, NY
Level is presenting at the 13th annual DMFS, the only North American event series built for financial services marketers. Speakers include marketing leads from MetLife, JPMorgan Chase, and US Mortgage Corporation.
Intent Is the New Code: Why the Next 100 Years Belong to Communicators, Not Engineers Ft. Artem Koren, Co-Founder and Chief Product Officer of Sembly AI
As I mentioned above, Artem Koren joined Myles and me on the podcast this week, and we covered a lot more ground than I fit in the log. His bigger claim: the last century of progress belonged to STEM, and the next belongs to whoever can state intent most sharply. He calls it intent technology.
Artem earned that view the long way, from computer science at Columbia to a decade in management consulting to a 2019 transformer on a single $800 GPU in his uncle's spare room. What you'll take from the episode:
Why a team of strong classical engineers failed to adopt large language models in their first year, and the mindset shift that fixed it
How I've started directing AI like a manager instead of a developer: acceptance criteria and what done looks like, no tech stack
Why the career pyramid is thinning in the middle and starting to look like the Eiffel Tower, with a training path closer to medicine
Where taste, judgment, and real-world endpoints still sit outside what AI can absorb, at least for now
The sharper the intent you hand the machine, the sharper the work you get back.
What primate research says about reach, faces, and outrage
Nicholas Spiro, Chief Commercial Officer at Viral Nation, went looking for social media lessons in primate research. His premise is one every social team should sit with. The brain your ad lands in evolved millions of years before the feed, and it still runs on touch, faces, and score-keeping. Your dashboard measures none of those. Spiro builds the argument on four studies.
Robin Dunbar found primates spend more of the day grooming each other as group size grows. Connection costs attention, and a million impressions buy none of it. Doris Tsao's team at Berkeley found six small regions in macaque brains that respond far more to faces than to any other image. That's why a creator with a face often outperforms a brand with a logo, regardless of production value. Sarah Brosnan and Frans de Waal watched capuchins throw back a cucumber after a neighbor got a grape. Comparison and outrage are the cheapest engagement to engineer, and they build the relationship on resentment.
The hopeful study: de Waal also documented uninvolved chimps consoling the loser after a fight. Spiro argues the campaigns that build community all carry some version of that consolation, some deliberate vulnerability.
For a brand that reports on impressions, the useful question is which of these four instincts your creative rewards.
The "be on Reddit" advice is aimed at the wrong stage of the funnel
Every GEO channel on Slack says the same thing: be on Reddit, be on YouTube, meet buyers where the community already trusts. Ten Speed tested that advice at the moment it matters most. That's when a B2B buyer stops asking what a category is and starts comparing named vendors inside ChatGPT, Perplexity, Claude, or Gemini.
At that stage, the models overwhelmingly cite pages a marketing team writes and owns. Product pages led everything. Comparison pages earned more citations than their share of buyer questions. Reddit and YouTube barely registered. Close to nine of every ten citations pointed to content the brand controls.
Greg Jarboe at Search Engine Journal pressed Ten Speed with six questions before publishing, and the answers are as useful as the finding. The headline split is descriptive, with no statistical test behind it. The sample is one agency's B2B client base at a single moment. Ten Speed said so plainly.
Two moves for this quarter. Rewrite product pages and your homepage as if a model is reading them cold. Say what it does, who it's for, and what it integrates with, in plain language. Treat your G2 and Capterra profiles as content, since directory copy feeds the shortlist. And before any AI visibility stat lands in a client deck, ask what the denominator is.
You've read about Sembly twice already in this issue, so here's the short version of what it does. You give it a stated goal ("pitch this product to Acme Co"), and it pulls context from your meetings, documents, CRM, and the client's website. Then it builds a branded, sourced deck or Word document, with the client's colors, fonts, logo, and tone derived from a URL. Every output exports as an editable .pptx or .docx.
The difference from Gamma or a general-purpose chat model is that it never starts from scratch. Client context and branding carry over from one deliverable to the next, and the free trial comes with credits to test it.
Your AI visibility, scored.
When did you last audit what AI thinks of your brand?
Don't rely on an unreviewed algorithm to manage your brand’s reputation. Our scorecard pulls back the curtain on how AI platforms frame your brand sentiment and which sources they use to vet your authority.
Inside your custom report:
Visibility score and share of voice: The percentage of AI prompts your brand appears in, benchmarked against your top competitors across all major models.
Sentiment and positioning: Whether AI describes your brand favorably, neutrally, or negatively, and whether you're showing up as the primary recommendation or an afterthought.
Optimization roadmap: The specific content structure and authority signals you need to improve your score.
