Aztek Marketing News Roundup (04/20 - 04/24)

Searchlight is Aztek's marketing news roundup that brings together the week’s most relevant developments in marketing, search, AI, and digital strategy, all in one place. We update this article throughout the week with news we think is worth your time, along with context to help you understand what changed, why it matters, and what it could mean for your business.

This week's topics:

• LinkedIn Feed Algorithm 2026: What the LLM Overhaul Means for Your Reach
• AI Design Wars: How ChatGPT Images 2.0 and Claude Design Are Reshaping Visual Prototyping
• Prove It or Lose It: Authenticity Signals That Make E‑E‑A‑T Matter in 2026

4/21: LinkedIn Feed Algorithm 2026: What the LLM Overhaul Means for Your Reach

LinkedIn hit the reset button on its feed ranking system in March 2026. The platform swapped a patchwork of heuristic rules for one large language model that reads every post for meaning, gauges conversation quality, and shows each member a blend of personal-network and topic-based content.

Many brands woke up to shrinking impressions even though their publishing schedule never changed. Let's get into what happened, why it matters, and how to adapt without chasing gimmicks.

What Changed in March 2026

LinkedIn’s new model uses a single retrieval and ranking pipeline that scores posts on three pillars:

• semantic relevance
• a depth score
• member preferences

Semantic relevance measures how closely a post matches the interests a member shows in previous activity. Depth score aggregates dwell time, saves, meaningful comments, and private shares. Member preferences include signals like language, connections, and mute settings. Poll spam, meme reposts, and click-bait that once gamed reaction counts now sink because they lack depth.

Inside the Depth Score

Depth score blends four actions: time spent reading, saves, original comments with at least a full sentence, and private shares through direct messages. The platform treats each metric as proof that the content delivered value worth passing on. A high depth score gives a post a second life outside the author’s network, which is why some niche experts are gaining reach while company pages fall back.

Why LinkedIn Rebuilt the Feed

Developers saw that the old stack had five different ranking stages that were hard to keep consistent. A modern language model can understand nuance in text and run the whole process in one pass, which cuts serving costs and improves content diversity. LinkedIn also wanted to spotlight expertise beyond the user’s first-degree network. That decision aligns with Microsoft’s wider plan to strengthen professional discovery features across products.

Winners and Losers Under the New Signals

Winners include subject-matter experts who share clear opinions, back them with insight, and stay on topic. Posts that spark thoughtful replies and follow-up questions rise quickly once the model sees members spending time in the thread.

Losers include batch-scheduled link drops that send readers off-site without context, generic motivational posters, and shallow polls. Those formats draw quick clicks yet rarely earn saves or detailed comments, so their depth score stays low.

Company Pages vs Personal Profiles

Company pages saw reach drop by as much as sixty percent after the change, while personal profiles often held steady. The model favors human voices that show expertise, nuance, and willingness to engage in discussion. Corporate broadcast posts tend to feel impersonal and rarely draw long comments, so their depth score underperforms.

Brands that rely only on page updates risk losing visibility unless they empower employees to share insights in their own words.

How to Optimise for the LinkedIn Feed Algorithm 2026

1. Publish point-of-view driven narratives tied to one main topic. Stick to your lane so the model can classify your expertise.
2. Invite discussion in the post itself and dive into the comments early. The first hour often sets the trajectory.
3. Replace “link in comments” tactics with short summaries that give value even if the reader never clicks through.
4. Coach employees to share the post with an original take rather than a simple repost.
5. Audit creative formats. Carousels and short native videos keep users in the feed, which helps dwell time.

Key Questions Before You Publish

• Does the opening line state a clear angle within 140 characters?
• Does the post add original thinking or real-world proof?
• Will it spark genuine conversation among peers, not just one-click reactions?
• Have you set time aside to reply to early comments quickly?
• Would you feel comfortable bookmarking this post for later reference?

If you answer yes to these prompts, the piece likely meets the depth score threshold.

Next Steps

Plan a ninety-day test. Pick two or three employees with strong subject knowledge and help them craft weekly posts that follow the playbook. Compare their depth scores to your company page reach, adjust topics, and double down where you see rising saves and quality comments.

LinkedIn’s feed now rewards substance over volume. Brands that lean on genuine expertise and invite dialogue will earn compounding reach while broadcast-only tactics fade.

04/22: AI Design Wars: How ChatGPT Images 2.0 and Claude Design Are Reshaping Visual Prototyping

Speed used to be the one thing generative image models offered that traditional design workflows couldn't match. They could spit out endless concepts in seconds, but most of those concepts still required heavy lifting before they were safe for client eyes. That gap narrowed this week when two releases landed almost back-to-back. OpenAI introduced ChatGPT Images 2.0, an upgrade that adds higher resolution, multilingual text, multi‑image storyboards, and a quick web fact‑check before the model starts drawing.

Anthropic followed with Claude Design, a prompt‑to‑prototype tool that builds interactive layouts you can tweak inside a sidebar editor and then export into your usual design stack. Both moves shift generative visuals from novelty to a practical starting point.

Why Marketers and Designers Should Pay Attention

The gains start with speed‑to‑concept. A cross‑functional team can move from vague idea to presentable draft in a single working session, not a week of bouncing tasks between copy, design, and development. The entry bar also drops. Product managers or content strategists who know what they want can sketch a first pass without touching design software. Designers regain bandwidth to focus on higher‑level visual decisions because the blank canvas is already filled with something coherent.

There is also competitive pressure. Companies experimenting now will iterate faster and pressure rivals to match that pace. We already saw a signal when Figma’s share price dipped after the Claude announcement, an investor‑level acknowledgment that prompt‑to‑prototype is a needle‑moving capability.

ChatGPT Images 2.0 vs. Claude Design: Strengths, Gaps, and Best‑Fit Scenarios

Guardrails, Brand Risks, and Governance Steps

Speed introduces risk unless your process evolves alongside it. Put these guardrails in place before AI drafts start flowing across Slack channels:

• Style‑guide overlay. Maintain a one‑page checklist that covers colors, typography, logo placement, and tone. Compare every AI‑generated asset against it before approving.
• Rights and attribution. Auto‑generated icons or stock‑looking photos may resemble protected assets. Treat them like any external resource that requires a license check.
• Version control. Store AI outputs in a dedicated folder with date stamps. Make sure half‑finished comps cannot be mistaken for final files.
• Accessibility audit. Contrast ratios and alt text still matter. AI will not catch every issue, so keep your usual compliance steps.

Why Human Oversight Still Matters

Generative design tools can shave hours off an early draft, but they are not replacements for experienced creatives. If anything, faster first passes make human oversight more important.

• AI produces options; people pick the winners. ChatGPT Images 2.0 and Claude Design can flood you with viable layouts, but a designer still decides which concept fits the brief, feels on‑brand, and tells a coherent story.
• Nuance is where brands live. Color balance, typography hierarchy, accessibility compliance, those details build trust. AI guesses; designers know.
• Context beats convenience. A headline that reads fine to a model may contradict the campaign’s larger message. Strategic alignment happens when humans connect visuals to business and audience insight.
• Risk lives in the margins. Licensing, inclusivity, and legal disclaimers remain human responsibilities. AI does not check every box, and a missed box can cost real money.

Navigating AI‑Driven Design

ChatGPT Images 2.0 and Claude Design arrived within days of each other, and together they paint a clear picture of where AI‑assisted design is headed. Static visuals are getting smarter, prototypes are getting faster, and cross‑functional teams are gaining the ability to test ideas in near real time.

The opportunity is obvious: accelerate exploration without sacrificing brand credibility. The threat is equally clear: move too slowly or skip the guardrails, and you risk a flood of off‑brand, legally questionable assets that erode trust instead of building it.

04/24: Prove It or Lose It: Authenticity Signals That Make E‑E‑A‑T Matter in 2026

AI has made it very inexpensive to crank out articles, images, and even full videos; trust is another story. Google now bakes invisible SynthID watermarks into Gemini‑generated media, the EU AI Act attaches fines to undisclosed AI content, and Instagram applies visible labels to every Reel it detects as machine‑made. Together, these shifts turn “content provenance” from a theory into a ranking and compliance requirement.

Google’s December 2025 core update gutted low‑effort AI pages across every niche, not just health or finance. Experience, Expertise, Authoritativeness, and Trustworthiness (E‑E‑A‑T) still define quality, but the era of taking a brand’s word for it has ended. Authenticity signals, like verifiable machine‑readable proof points, are now the baseline requirement.

The New Proof Layer

Modern content integrity now hinges on a technical "proof layer" that travels with every asset and confirms its origin and edits. Here is what that layer looks like in practice:

Watermarks and invisible tags
SynthID hides imperceptible signatures inside each pixel, letting Google confirm whether an image or video came from an AI model. If you publish machine‑generated visuals without the mark, you may be mislabeling origin and risking a visibility hit when detection tightens.

Cryptographic signatures (C2PA)
Major publishers are adopting the C2PA standard, which bakes a tamper‑proof manifest into every asset. Anyone can inspect the chain of custody: who created it, when, and what edits were made. It is nutrition labeling for media.

Mandatory platform labels
Instagram’s new policy demands an on‑screen disclosure for any AI‑touched Reel. Other networks will follow because regulators want easy‑to‑audit rules. If your campaign leans on AI visuals, plan creative so the label does not undercut the message.

These proof tags don't show up in the on‑page text; they live in the background data that search engines read and trust. Leave them out, and Google will probably treat the content with suspicion.

Algorithm Reality Check

Google has spent the past year proving that it now judges authenticity by action, not intention. Two recent updates show exactly how:

December 2025 Core Update
Google expanded its quality filters beyond sensitive topics. Sites that leaned on quick AI rewrites, thin listicles, or syndicated guest posts lost as much as 80% of their organic traffic within days. The common thread was a lack of human perspective, original evidence, or technical proof tags.

Perspectives Refresh 2026
The new Perspectives filter favors content that feels like it came from a real person. It looks for first‑person language, unique images, ProfilePage schema, and outbound citations to primary sources. Pages with those signals gain extra carousel slots while textbook‑style explainers drop out of view.

Building Your Authenticity Signals Stack

Content provenance checklist

• Enable SynthID whenever Gemini or another model assists an image or video.
• Export graphics through tools that support C2PA or Adobe Content Credentials.
• Store a visible changelog on evergreen pages so Google can see how content evolves.

Schema and metadata upgrades

• Add ProfilePage, Author, and Review schema.
• Link author bios to verified social handles or professional directories.
• Include DOI or source links for proprietary data so citations can be crawled.

First‑person evidence

• Embed screenshots, photographs, or data exports that only a practitioner could capture.
• Use short video or audio snippets from subject‑matter experts and transcribe them for accessibility.

Combined, these authenticity signals travel with the asset wherever it is embedded or scraped. Even if someone copies your article, they lose the watermark and the cryptographic proof, handing the originality win back to you.

Next Steps to Win With Authenticity Signals

1. Audit your live content library for missing bylines, duplicate author bios, or AI images without SynthID.
2. Implement watermarking or C2PA on every new multimedia file and update your style guide to enforce it.
3. Update schema sitewide and add an “Updated on” note with a mini‑changelog for major pages.
4. Train writers and designers on when and how to disclose AI assistance.
5. Monitor Search Console impressions and the Google Perspectives filter for lift tied to your new authenticity signals.

Brands that show their work, sign their content, and back claims with first‑hand evidence will keep the rankings, reach, and trust that generic AI clones cannot touch. Prove your expertise, or prepare to watch it disappear.