The real pattern in all the noise
Scan the headlines and you see three things on repeat:
- Google keeps rewriting the rules of search (core updates, AI overviews, I/O announcements).
- Everyone is shipping “AI for X” (content, product marketing, analytics, prospecting).
- Marketers are scrambling to patch performance (signal decay, cannibalization, title rewrites, hashtag tools, Shorts hooks).
Underneath all of that is one problem that actually matters:
Your growth engine is sitting on decaying signals and rented platforms, while your team is distracted by tool-of-the-week AI hacks.
If you’re a CMO, performance lead, or media buyer, the question is no longer “How do we use AI?” or “What’s the next channel?” It’s:
How do we build an AI-resilient, signal-aware growth system that survives search overhauls, privacy shifts, and platform volatility?
What’s actually changed (and why your old playbook is rotting)
1. Search is no longer a neutral listing service
Between Google’s “biggest search update in 25 years,” AI overviews, and rolling core updates, search is drifting from:
“We send you traffic for good answers” → “We keep users and summarize your answers.”
That means:
- More zero-click experiences and fewer organic visits per impression.
- Brand visibility increasingly mediated by AI summaries, not just blue links.
- SEO wins becoming more brittle; one update can erase quarters of work.
If your growth model assumes “search will keep sending us compounding free traffic,” you’re running on a shrinking asset.
2. Signal decay is eating your funnel from the top down
Privacy changes, shorter attribution windows, and cross-device chaos mean your performance data is:
- Less complete (missing impressions, missing conversions).
- Less durable (data you used to rely on for 90 days is now useful for 7-14).
- Less portable (platforms hoard their own signals; your view is stitched and laggy).
Media platforms are compensating with more automation:
- “Smart” bidding and budget routing.
- Broad match and Advantage+ style campaigns.
- Black-box optimization that you can’t fully audit.
The result: your top-of-funnel performance looks worse, not always because the creative or audience is broken, but because the underlying signals are thin and decaying faster.
3. AI is flooding the pipes with content and tools, not strategy
You can now:
- Automate content and product marketing with agents.
- Crank out SEO pages, title tags, and Shorts scripts at scale.
- Generate AI visibility reports, AI analytics, and AI prospecting lists.
The bottleneck is no longer “Can we produce enough?” It’s:
- “Can we produce anything distinct enough to matter?”
- “Can we measure impact in a world of decaying signals?”
- “Can we keep control of our message instead of outsourcing it to generic models?”
Tooling has exploded. Strategy has not.
The job now: design an AI-resilient, signal-aware growth system
The operators who win the next three years won’t be the ones with the most AI tools. They’ll be the ones who:
- Treat platforms as rented distribution, not foundations.
- Design for signal decay instead of pretending it’s 2018.
- Use AI to compress time-to-insight, not to write more generic content.
1. Rebuild your growth stack around durable signals
Start by sorting your signals into three buckets:
- Rented signals: platform pixels, view-through conversions, in-platform ROAS, engagement rates.
- Owned signals: first-party events, CRM data, email behavior, product usage, LTV.
- Inferred signals: modeled conversions, media mix models, incrementality tests.
Then make two moves:
-
Prioritize owned signals in your decision-making.
- Shift from “How did this ad set perform?” to “What did this cohort do over 30-90 days?”
- Connect ad platforms to a central warehouse or CDP; stop letting each channel grade its own homework.
- Standardize event schemas across web, app, and offline so your signals are actually comparable.
-
Use rented signals tactically, not as the source of truth.
- Optimize in-platform for speed; validate in your own data for direction.
- When platform and internal data disagree, treat it as a hypothesis, not a crisis.
2. Design for signal decay instead of fighting it
You can’t reverse privacy or force perfect tracking. You can adapt your operating system:
- Shorten your feedback loops.
- Move from quarterly to biweekly or weekly experimentation cycles.
- Use leading indicators (engagement, CTR, scroll depth, reply rate) as early signals for creative and offer quality.
- Build always-on incrementality testing.
- Geo holdouts, PSA tests, or ghost ads where possible.
- Even simple on/off or bid-down tests beat blind trust in platform reporting.
- Accept fuzzier attribution and upgrade your questions.
- Stop asking “What’s the exact ROAS of this ad set?”
- Start asking “What’s the directional impact of this channel on new high-value customers?”
3. Make AI your analyst and orchestrator, not your copywriter-in-chief
Most teams are using AI where it’s least strategic (writing more generic copy) and underusing it where it’s most valuable (sense-making and orchestration).
Reassign AI’s job description:
- AI as analyst
- Automate weekly “growth health” briefs: channel shifts, cohort behavior, anomalies.
- Have agents scan search console, analytics, and CRM for patterns: cannibalization, content gaps, decaying pages.
- AI as workflow router
- Trigger creative refresh tasks when performance drops below a threshold.
- Auto-generate test matrices (angles, hooks, offers) from winning patterns.
- AI as draft assistant, not final voice
- Use it for first passes on briefs, outlines, and variations.
- Keep a human gate for anything customer-facing, especially in a SaaS recession where trust is already thin.
The goal is not “AI did this.” The goal is “We moved from idea to validated test in days, not weeks.”
4. Build portable workflows, not platform dependencies
With AI agents, APIs, and automation everywhere, it’s tempting to wire everything tightly into one ecosystem. That’s comfortable-and dangerous.
To stay resilient as platforms and policies change:
- Standardize your primitives.
- Define common objects: campaigns, audiences, offers, creative concepts, experiments.
- Map each platform’s quirks into that shared model instead of building one-off workflows per channel.
- Keep your logic outside the walled gardens.
- Budget allocation rules, bidding strategies, and experimentation frameworks should live in your own systems (or at least your own docs), not as one-off tweaks in ad managers.
- Use AI to make migrations cheaper.
- When you move from one ESP, CRM, or ad platform to another, use AI to translate naming conventions, audiences, and workflows.
- Document these patterns once; reuse them across tools.
From “channels and hacks” to “offers and memory”
In a world of decaying signals and AI summaries, the things that compound are boringly old-school:
- Positioning that sits clearly in the customer’s head.
- Offers that are sharp, specific, and hard to copy.
- Brand memory built across touchpoints, not just last-click wins.
1. Treat SEO and content as brand distribution, not a traffic faucet
With AI overviews and generative search, a lot of your “how to” content will be read by machines before it’s read by humans. That changes the job:
- Shift from keyword-chasing to idea ownership.
- Pick a few core problems you want to be known for solving.
- Build deep, opinionated content around those, not 500 thin variants.
- Optimize for quotability, not just rankings.
- Clear definitions, sharp frameworks, and data points that AI systems are likely to cite.
- Structured content (FAQs, summaries, schemas) that’s easy for models to ingest.
- Measure brand search and direct demand, not just pageviews.
- Track branded queries, category+brand searches, and assisted conversions.
- Use these as your north star for content, not raw organic sessions.
2. Make creative and offers your primary optimization levers
As targeting becomes more automated and opaque, your biggest controllable levers are:
- What you say (message, positioning, promise).
- What you offer (pricing, bundles, guarantees, trials).
- How you present it (format, hook, story, proof).
Build a simple, ruthless system:
- Maintain a live “offer backlog” with hypotheses: new bundles, guarantees, bonuses, payment options.
- Run structured creative sprints: 1-2 key narratives per quarter, tested across channels.
- Use AI to generate variations within a strong concept, not to invent your positioning from scratch.
3. Invest in brand memory as a performance asset
The Burger King CEO’s quote-“Invest in brand or you’ll surely die”-is not sentimental. It’s math.
When signals are weak and attribution is fuzzy, strong brand memory does three commercially useful things:
- Improves click and conversion rates because people have seen you before.
- Reduces your effective CAC over time as more conversions are “assisted” by prior exposure.
- Makes you more resilient to platform changes because people search for you, not just the category.
Practically, that means:
- Protecting a portion of spend for reach and frequency, even when short-term ROAS looks worse.
- Aligning social, search, email, and site experience around a consistent story, not channel-specific gimmicks.
- Using AI to monitor brand mentions and sentiment, then feeding that back into creative and product decisions.
What to actually do in the next 90 days
If you strip away the noise, a practical 90-day plan for an AI-resilient, signal-aware growth engine looks like this:
- Audit your signal stack.
- List your top channels and map which signals are rented vs owned.
- Identify one place where you’re over-trusting platform numbers and under-using your own data.
- Stand up one simple incrementality test.
- Pick a major channel and run a geo holdout or on/off test.
- Use AI to help design, monitor, and summarize the test.
- Redefine AI’s role on your team.
- Document where AI is allowed to create, where it’s allowed to assist, and where it’s banned.
- Assign one owner to build a “growth analyst agent” that generates a weekly report from your core data sources.
- Pick one core narrative and one core offer to pressure-test.
- Run it across paid social, email, and site. Use AI to generate variations, but keep the spine human.
- Measure impact on high-intent behaviors and brand search, not just immediate ROAS.
- Document your portable workflows.
- Write down how you plan, launch, and evaluate campaigns-independent of any one platform UI.
- Where possible, move that logic into your own tools or at least your own templates.
The environment will keep changing: more AI in search, more automation in ad platforms, more privacy constraints, more tools promising shortcuts. Your job isn’t to keep up with every headline. It’s to build a growth engine that still works when the next update lands and the next signal decays.