If you have spent any time with a screen on this week, you have likely felt the strange, temporal distortion of the Olympics.
Time moves at an irregular pace when there is a constant stream of figure skating and curling in the background. But compared to the current pace of AI and digital marketing, the Olympics actually feel slow.
In the time it takes these athletes to reach the 2030 Games, we’ll have gone from leveraging platform-specific features like AI Max to having AI agents negotiate media buys with other AI agents.
Because our industry is currently moving faster than a downhill ski run, I have come up with a few suggestions for the Olympic committee. If they want to see true endurance, they should forget the triple axel and watch a growth marketer handle a Monday morning:
AI Prompt Biathlon: Ski five miles, then write a prompt that gets a usable CSV on the first try.
Approval Slalom: Dodging a series of gates as the creative baton is passed from Copy to Design to Legal to Design back to Legal, then to the CMO, then the CMO’s first cousin once removed, who finally says, "Let's go in a different direction."
Screen-Share Figure Skating: The art of elegantly hiding 47 open browser tabs in the 1.5 seconds before your presentation begins.
Image Optimization Curling: Precision-sliding a 5MB hero image down to 50KB without it looking like a Minecraft screenshot.
But unlike the Games, there is no four-year wait for the next iteration. In our industry, the closing ceremony for one platform usually starts five minutes before the opening ceremony for the next.
So, before you head back to the "Approval Slalom," let’s dive into the signals that actually matter for 2026.
Patrick Patterson
CEO at Level Agency
Summit highlight
A quick break from the headlines.
We pulled together a short highlight reel from our recent Summit, and I finally had a chance to show it at our monthly company meeting. It’s equal parts fun and useful, because you can feel how quickly teams align when they share context in the same room.
If you want a glimpse of the energy, here’s the video:
Proud of the work. Even more excited about what it unlocks next.
And if you’re curious how we’re structuring AI across performance, CRM, and brand systems, we shared more context around the Level + BAM Strategy integration and the Level Intelligence Suite here.
Podcast spotlight
I recently sat down with our new SVP of AI, Krish Sailam, to talk about what happens when the AI honeymoon phase ends and the actual work begins.
Krish has spent two decades driving innovation for giants like Amazon and BMW, so he has a blunt, refreshing take on why the market-maker mindset matters far more than the tools themselves.
We spent time digging into the looming trust crisis where deepfakes and AI noise make it harder to maintain authentic brand connections. Krish also explained why AI-generated applications have effectively killed the traditional resume and why human empathy is actually your most critical technical tool for designing breakthrough services.
If you’re tired of the cinematic AI hype and want to know what actually changes in the messy, Monday-morning work, this is the episode you need.
ChatGPT ads spotted and they are quite aggressive OpenAI is testing search ads that look like sponsored answers. These placements hog significant space and bury organic results. This shift means winning in AI will eventually require paying for the pole position in the chat window. Read more
Reddit is testing a new shopping product experience in search Reddit is leaning into its status as a massive product review engine by testing a shopping tab in search. This feature combines product data with actual subreddit conversations. E-commerce brands should treat comment sentiment as a vital part of the conversion funnel. Read more
Why do budgets overspend even with a Target ROAS in place? Google’s smart bidding often prioritizes high-intent conversion clusters over strict daily limits. This leads to spending spikes even when you have tROAS settings in place. Remember that tROAS is a directional north star for the algorithm rather than a hard budget brake. Read more
Google’s new Gemini Pro model has record benchmark scores (again) Gemini 3.1 Pro is crushing records in reasoning and coding benchmarks. It is especially strong at long-context understanding compared to its peers. Growth teams can use this to feed huge datasets into a single prompt to find patterns a human would miss. Read more
The hidden failure of AI PDF parsing AI still struggles with the visual logic of PDFs. Tables and multi-column layouts often get mangled during the extraction process. You should verify any automated reporting that relies on document parsing before trusting the output. Read more
UGC content moderation is the new brand safety frontier Brand safety now involves avoiding subtle misinformation and off-brand "vibes" in creator content. As UGC grows, the burden of moderating these assets is exploding. A strong moderation stack is essential if you want to protect your reputation at scale. Read more
Deep dive: AI
Better context > better prompts
Early AI adoption felt like clever prompting. Write the right instruction and the model does something impressive.
Tessl’s article frames the next step as context craft. Not more tricks in the prompt box, but a disciplined way to build, refresh, and serve the information an agent needs to work safely inside real codebases. That includes architecture decisions, naming conventions, dependency relationships, and the why behind past tradeoffs, not just the current files.
The useful part is the lifecycle mindset. Context gets created, curated, versioned, and maintained as the system changes. When teams treat it like a living asset with ownership, agents stay grounded and results get more repeatable. When context is ad hoc, outputs swing from brilliant to risky depending on what the agent happened to see.
This matters outside engineering too. Growth teams are now running AI across analytics, forecasting, reporting, and creative workflows. The same rule applies: define the terms, encode the constraints, and keep the context current. That’s how you get speed without drift.
Related signal: Agent Skills are emerging as a more domain-specific evolution beyond MCP-style integrations, with major systems like Claude beginning to support them. If you want to go deeper into the technical layer shaping this shift, this new resource is tracking the space.
Deep dive: Growth
Marketing and finance are arguing about different scorecards
Last week’s newsletter hit on the trust gap that shows up when attribution gets treated like truth instead of a model. This Level piece on financial services marketing measurement does something most attribution debates avoid. It names the core tension directly: marketing and finance are often working from different scorecards.
Marketing optimizes for leading indicators and platform-reported performance. Finance optimizes for recognized revenue, timing, and risk-adjusted outcomes. Both can be internally consistent and still fundamentally misaligned.
The piece frames it as the “two scorecards problem,” and then walks through a pragmatic fix: map finance’s conversion definitions and timing windows, expose where tracking diverges, standardize reporting in finance’s language, and then adjust acquisition strategy toward the conversion points finance actually recognizes.
It’s written for financial services, but if you’re in any high-consideration category with long cycles, compliance, or offline conversions, it’s a masterclass in internal alignment strategy.
Codex is OpenAI’s agentic coding workspace built for real engineering tasks: feature work, refactors, migrations, and multi-step changes that usually stall in backlog limbo.
Even if you’re not writing code, you live downstream from it. Tracking fixes, landing page experiments, CRM integrations, analytics QA. Codex reduces translation overhead and speeds up execution loops.
In a year where velocity matters, shrinking the gap between idea and implementation is a real advantage.
What does ChatGPT actually say about you?
When a potential buyer asks an AI agent for a Top 5 list in your category, do you make the cut? More importantly: Why or why not?
Don't leave your brand’s reputation to an algorithm you haven't audited. 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:
Sentiment analysis: Is the AI's summary of your brand accurate or outdated?
Citation audit: Discover which third-party sites are feeding the AI your info.
AEO roadmap: Immediate steps to improve your "AI discoverability."
