KIMP’s Picks July 17th: Weekly Highlights In Design, Marketing, AI, & More

What did you miss this week? Maybe it was a new AI tool making headlines, a marketing trend gaining momentum, or a platform update that could change how brands connect with audiences. With so much happening every day, it’s easy for important updates to slip through the cracks.

That’s where KIMP’s Picks comes in. We’ve gathered the stories, ideas, and industry insights worth your attention and wrapped them into one easy-to-scan recap.

Welcome to the KIMP’s Picks’ July 17th edition.

In the Marketing Realm 

Fine-tune attribution windows in Google Analytics 

Google Analytics shared a post highlighting how it now allows marketers to customize conversion lookback windows to help them assign attribution credit more accurately based on how long customers take to convert. Since many users complete a purchase or key action days or even weeks after first interacting with an ad, relying on a standard attribution window may not always reflect the full customer journey.

To adjust these settings, users can go to the Admin section in Google Analytics, open Key Events, select a conversion event, and choose a lookback window that best fits their business needs. Changes apply only to future data and are reflected across all attribution models and reports within the property.

This update gives brands greater control over attribution accuracy, helping them better understand the true impact of their marketing efforts.

Gap and Hailey Bieber tap into nostalgia for new denim collection 

Gap has partnered with Hailey Bieber on a denim capsule collection inspired by 1990s fashion, a decade that heavily influences Bieber’s personal style. The collection centers on the Hailey Jean, based on her favorite vintage Gap denim. 

Custom details include Bieber’s signature inside the pocket lining and 1996 branding, referencing both her birth year and a notable era in Gap’s denim history.

The collaboration is part of Gap’s broader brand revival strategy. Since 2023, the retailer has focused on culturally driven marketing, celebrity partnerships, and updated creative direction to reconnect with consumers. Recent collaborations and campaigns have helped fuel momentum, with Gap reporting a 10% increase in first-quarter comparable sales.

This partnership combines celebrity influence, nostalgia marketing, and heritage branding to strengthen Gap’s relevance with modern consumers.

Liquid Death taps into humor for its latest campaign 

Liquid Death’s latest campaign promotes its Sparkling Energy drink by poking fun at one of the category’s biggest clichés – the idea that energy drinks help people perform extreme feats. Instead of showing athletes pushing limits, the ad features a humorous rooftop fall, reinforcing that the drink contains a moderate amount of caffeine and zero sugar rather than promising superhuman results.

The campaign supports a consumer promotion offering DoorDash gift cards, positioning the reward around buying chicken wings and other food rather than focusing solely on product benefits. By leaning into self-aware humour, Liquid Death differentiates itself from traditional energy drink advertising and stays true to its unconventional brand personality.

The key marketing lesson from this campaign is the power of category disruption. Rather than repeating the same performance-driven messaging as competitors, Liquid Death creates memorability by challenging industry norms and embracing entertainment. The campaign also shows how brands can build stronger audience connections through honesty and humour, especially when consumers are increasingly sceptical of exaggerated marketing claims.

In the Content Marketing Realm 

How marketers can use AI to build smarter content strategies 

A post from Search Engine Land talks about a new workflow that shows how combining Semrush, Google Search Console, Google Analytics, and Claude can help marketers identify and prioritize content opportunities based on business impact, not just search volume.

A content gap analysis helps brands discover topics competitors rank for but they do not. While SEO tools can uncover thousands of keyword gaps, the bigger challenge is deciding which opportunities are worth investing in.

The process combines competitor insights from Semrush, first-party search data from Google Search Console, and engagement metrics from Google Analytics. Claude can then analyze the combined data, group related topics, identify patterns, and recommend which content areas deserve attention.

The approach focuses on finding quick wins, such as improving existing pages already gaining search visibility, along with larger opportunities that require new content or broader topic coverage. Marketers can score opportunities based on factors like business relevance, search demand, ranking difficulty, existing authority, and required effort.

The workflow also emphasizes human review, as AI can organize and analyze data but cannot replace business knowledge and strategic judgment.

In the Design Realm 

Google Fonts you need to try 

Google shared its recent font recommendations highlighting typefaces that can add personality and flexibility to websites, apps, and creative projects.

The post spotlights three fonts with different design purposes. Sekuya is an all-caps display serif font that combines classic curves with a bold, modern style, making it suitable for attention-grabbing headlines and hero sections. 

MoMo Family is a typography system developed for Vietnam’s e-wallet app MoMo. It includes Trust Sans and Display for a clean, rounded look, along with Signature, a handwritten style designed to create a warmer and more human brand voice.

SUSE Mono is a hybrid monospace font focused on balancing geometric structure with readability. Its flexible design makes it suitable for both digital interfaces and print applications.

The series aims to help designers discover open-source fonts that can support different creative needs, from bold brand expressions to functional digital experiences.

So if you are looking for fresh fonts to try in your next design project, this is a good list to begin with. 

Tinder refreshes its brand identity 

Tinder has unveiled its first major rebrand in nearly a decade to create a more expressive identity aimed at connecting with Gen Z while maintaining the app’s recognizable personality.

The refreshed identity includes an updated flame icon, an all-caps wordmark, a new serif typeface, and an expanded color palette that moves beyond Tinder’s traditional red and orange tones. The visual system combines bold typography with mixed-media imagery, including anime references, fine art, photography, and meme-inspired visuals to reflect how younger audiences experience relationships and culture.

The rebrand also places a stronger focus on Tinder’s voice, built around a fictional dating columnist named “T.” The personality blends humor, honesty, and empathy while moving away from traditional dating expectations like “finding the one” and highlighting different forms of connection.

Tinder’s recent rebranding shows how brands can evolve for new audiences while preserving the recognizable elements that made them successful.

In the AI Realm 

Codex Micro – OpenAI’s new AI agent controller 

OpenAI has partnered with Work Louder to launch the Codex Micro. This is a custom hardware controller designed to make working with AI coding agents more efficient. The device gives users physical controls for managing Codex tasks, monitoring agent activity, and accessing frequently used workflows.

The Codex Micro features RGB-lit agent keys that display live status updates, allowing users to see whether an agent is thinking, running, waiting, or complete without switching between chats. It also includes a joystick for triggering common Codex actions like code reviews, debugging, and refactoring.

The device adds dedicated command keys for actions such as accepting or rejecting changes, starting new chats, and push-to-talk controls. A rotary dial allows users to adjust reasoning levels based on the complexity of their tasks.

Built with mechanical switches, custom keycaps, and aluminum components, the device works with Mac and Windows through Bluetooth or USB-C connectivity. It includes a custom Codex icon keyset with additional keycaps designed around AI workflows.

Spotify tests AI-powered conversations for Premium users to discover content 

Spotify is introducing a new conversational AI experience that allows eligible Premium users to interact with the app through voice or text. The feature helps listeners discover music, understand content, and explore their listening history without leaving Spotify.

Available through the Home and Now Playing views on mobile, the experience lets users ask Spotify to create playlists, adjust recommendations, save songs, add tracks to queues, and refine the mood or style of what they are hearing. Users can also ask questions about songs, albums, artists, podcasts, and audiobooks to learn more about the content.

The feature uses a listener’s playlists, favorite artists, repeat plays, and listening history to answer personalized questions, such as when they first heard a song or which genres they have recently explored.

The conversational experience is rolling out in beta to eligible Premium users ages 18 and older in the U.S., Ireland, and Sweden on iOS and Android in English. 

OpenAI unveils GPT-Red to automate AI safety and fortify models 

After introducing GPT-Live last week, OpenAI has recently introduced GPT-Red, an automated red-teaming model designed to help identify AI vulnerabilities at a larger scale. The system aims to address a growing challenge in AI safety: traditional human-led red-teaming is valuable but difficult to scale as models become more capable.

GPT-Red works by creating adversarial attacks against AI systems, including prompt injection attempts, then using those findings to improve model defenses. OpenAI used GPT-Red during the training process of GPT-5.6 Sol, helping the model become more resistant to prompt injection attacks while maintaining its general capabilities.

The model was trained through self-play, where GPT-Red generates attacks and other AI systems learn to defend against them. OpenAI reports that GPT-Red successfully identified vulnerabilities across various testing environments, including attacks involving browsing, tools, and AI agents.

Google Images celebrates 25 years with new AI-powered features 

Google is marking 25 years of Google Images by introducing new ways for users to explore and create visual content through Search. The updates build on Google’s shift from traditional text-based search toward more interactive, image-focused experiences.

A new browsable Google Images home experience will feature a personalized gallery of images from across the web, updated in real time based on user interests. Users will also be able to save ideas into collections and easily return to topics they want to explore further.

Google is also bringing image generation into AI Overviews in Search. Powered by its Nano Banana model, the feature allows users to create custom images directly from text prompts, helping turn specific ideas into visuals when existing images are not available.

Additionally, Google highlighted major milestones in visual search, including Similar Images, Search by Image, Google Lens, Multisearch, Circle to Search, and AI-powered visual search tools that allow users to explore objects, scenes, and products through images.

Thinking Machines Lab releases a new open AI model, Inkling 

Thinking Machines Lab, is the AI startup founded by former OpenAI CTO Mira Murati. The company has officially launched its first in-house AI model, Inkling. Breaking away from the traditional closed-system approach of major rivals, Inkling is an open-weights model. This means developers and businesses can download the model weights from Hugging Face. And they can host and customize the AI completely on their own terms. 

Architecturally, it is a massive 975-billion-parameter Mixture-of-Experts model. However, it only activates 41 billion parameters per query to keep processing costs low.  Inkling stands out with native multimodal capabilities. It can seamlessly process and reason across text, images, and audio right out of the box. Moreover, it features a customizable “thinking effort” toggle. So it lets developers dial reasoning up or down to strike the perfect balance between speed and accuracy. 

To make customization easy, Thinking Machines Lab has paired the model launch with Tinker, a dedicated platform for fine-tuning the model to specific tasks. 

Anthropic announces Claude for Teachers 

Anthropic has launched Claude for Teachers, offering verified K-12 educators in the US free access to premium Claude features, teaching resources, and curriculum support tools.

The program connects Claude with Learning Commons, giving teachers access to academic standards across all 50 states and curriculum resources from organizations including OpenSciEd and Illustrative Mathematics. The goal is to help educators create standards-aligned lessons, adapt materials for different learning levels, and reduce planning workload.

Claude for Teachers also integrates with education platforms such as Canva Education, MagicSchool, Diffit, Brisk Teaching, and others to help teachers create classroom materials, analyze student progress, and design personalized learning activities.

Social Media Feature Updates and Algorithm Changes

You can now add “Flair” to your posts on Threads 

Threads is expanding its community flair feature, allowing users to add icons and emojis to posts within certain communities to show specific interests or discussion topics.

The flair appears below a user’s name on community posts and helps members quickly understand the focus of a post. For example, users in BookThreads communities can use icons to indicate interests such as romance, writing, or other book-related topics.

Once selected, a user’s flair will automatically appear on future posts within that community, though it can be changed anytime. Threads first tested community flair in December and is now making the feature available more broadly.

The update builds on Threads’ growing focus on topic-based communities, which launched for all users in October. The platform has continued expanding these groups as discussions around specific interests grow, including communities that are developing into smaller subtopics.

View on Threads

X adds the “Mentions Boost” feature 

X has launched Mentions Boost, a new feature designed to help Premium Business subscribers promote organic conversations and customer mentions around their brands.

The tool is part of X’s updated Ads Manager experience and builds on the platform’s existing Boost options. It allows businesses to turn real user discussions and customer advocacy into promoted content while keeping the original voices behind those conversations visible.

X says the feature is designed to help brands capture attention from existing conversations, strengthen engagement, and connect organic discovery with advertising campaigns.

Mentions Boost is now available to all Premium Business subscribers through the X app.

X tweaks the algorithm to show more posts from accounts you follow back 

X is updating its recommendation algorithm to increase the visibility of posts from accounts users follow, especially mutual connections where both users follow each other.

Nikita Bier, Head of Product at X, said the platform found that information about mutual follows was missing from the algorithm, causing replies and conversations to feature more unfamiliar accounts. The change aims to make discussions feel more connected and help interest-based communities form more naturally.

The update addresses a wider challenge across social platforms, where AI-driven recommendation systems often prioritize engagement signals over users’ explicit choices, such as the accounts they intentionally follow.

TikTok expands AI transparency tools on the platform 

TikTok is introducing new measures to help users better understand, identify, and interact with AI-generated content on the platform.

The company is expanding AI literacy efforts through new educational resources, including a guide created with media literacy organizations NAMLE and Henry Ajder. TikTok also plans to launch an in-app hub that will help users learn how to recognize AI-generated content.

To reduce AI-generated spam, TikTok is testing improved detection systems aimed at identifying accounts that use AI to mass-produce content in areas such as politics, financial advice, and medical topics. The platform said it removed more than 86 million fake accounts in the first three months of 2026.

A Final Word…

That’s a wrap for this week’s KIMP’s Picks. The worlds of marketing, design, and AI continue to move quickly, with new tools, trends, and ideas emerging every day. Whether it’s a fresh creative approach, a smarter workflow, or a platform change worth knowing about, we’ll keep bringing you the highlights that matter most. We’ll see you in next week’s edition of KIMP’s Picks.