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

AI moves fast. Industries are adapting to these changes faster. So, you cannot afford to miss the AI news that everyone’s talking about. AI updates that will probably change the way you work, the way you communicate with your team and the way you automate your workflows in the coming weeks. Social media has become a go-to destination for catching up on these updates but keeping up with the constantly refreshing feeds can be overwhelming. 

Let’s simplify that, shall we? Welcome to the KIMP’s Picks April 17th Edition – we’ve handpicked the latest social media buzz so you do not have to spend all weekend scrolling endlessly to find out what you missed. AI news and announcements from AI startups, marketing news and insightful marketing discussions, changes coming to social media algorithms and so much more. Ready for a quick recap? Let’s get started. 

In the Marketing Realm 

Brands capitalizing on creator marketing 

YouTube recently shared a post highlighting the power of creator marketing and how brands around the world are leveraging the platform for creator marketing. 

Reportedly, about 87% of YouTube U.S. viewers consider the platform a leading source of product discovery thus making it the top platform influencing purchase decisions. 

Popular cookie brand Crumbl built on creator trust to demonstrate the use cases of its app. Supergoop, on the other hand, launched a high-frequency Video View campaign to boost awareness and increase ROI. Canva with its use of Demand Gen and Creator partnerships on YouTube, demonstrated how the platform makes creator-led campaigns easily scalable. 

This trend shows that creator-led content brings strong growth potential, meaning brands that invest early in authentic creator partnerships are more likely to influence purchase decisions at scale. 

Google updates Dynamic Search Ads to AI Max 

Google is upgrading Dynamic Search Ads (DSA) to a new system called AI Max. This brings a shift toward more automated and AI-driven Search campaigns. The rollout moves out of beta and will gradually replace legacy DSA setups starting in September 2026.

AI Max is designed to improve performance by combining website content, advertiser inputs, and real-time intent signals. Google reports an average lift of around 7% more conversions or conversion value when using the full feature set compared to traditional search term matching.

Existing campaigns will be automatically transitioned, with their settings preserved to maintain performance stability. Accordingly, Google is encouraging advertisers to switch early to test features and avoid disruption.

This change reflects a clear move toward AI-first campaign management, where automation and intent-based targeting take priority over manual keyword control.

How AI is reshaping market research 

A recent post from MarTech explores how market research is entering a new phase where AI is no longer just a helper for isolated tasks, but a core part of how insights are created, stored, and validated. 

Today, many teams already use AI to summarise interviews, generate survey questions, or clean data. But these are still small, task-level improvements. But the structure of the market research process itself is changing. Tools like Anthropic’s Projects in Claude give AI long-term context and allow it to analyze patterns across data and act like a living memory for insights. 

Secondly, tools like Google’s Gemma allow organizations to run AI inside their existing systems ensuring the security of sensitive customer data. In other words, these tools help teams analyze private datasets securely. 

Finally, the growing popularity of multi-AI systems means that businesses are deploying different models to handle different tasks like summarization, fact-checking and sentiment analysis. All these results are then compared in order to enhance accuracy and reduce bias. 

This matters because it turns AI into the backbone of modern insight generation, reshaping how organisations understand customers and make decisions

How AI is shaping buying decisions 

Semrush shared some insightful data from a recent research to understand the influence of AI on consumer shopping behavior and purchase decisions. 

According to Semrush, 85% of the surveyed consumers use AI every week with about 55% of them using it for researching products in particular. But yes, traditional search still holds a significant place, with 77% of consumers using both Google and AI for information about the same purchase. 

On the whole, data shows that being mentioned “first” is not the most crucial goal at the moment but being mentioned by AI is the real deal. So, the goals for marketing are changing. 

Traditional search isn’t going anywhere so brands need to focus on that but accommodating strategies to enhance AI visibility is more important now than ever. 

In the Content Marketing Realm 

Understanding Gen AI traffic 

We just spoke about the need for brands to focus on AI search visibility. But then comes the question, which AI tools matter? 

Traffic to popular generative AI sites can be a good indicator to understand this because when you are working on content that strengthens your brand’s visibility, you need to know where your customers are. 

According to Similarweb, ChatGPT seems to be consistently leading in terms of web traffic with Gemini and Claude picking up quickly. So these are platforms that most brands need to be considering in their AI content optimization strategies. 

In the Design Realm 

Canva x TikTok simplifying designs for ads on TikTok 

TikTok and Canva have expanded their partnership to make it easier for businesses to create and launch ads directly on TikTok. The integration connects Canva’s design tools with TikTok’s advertising platform, cutting out extra steps between creative production and campaign launch.

Many brands already build their marketing assets in Canva, from brand kits to social media visuals. Now those same designs can be turned into TikTok-ready videos and published straight to Ads Manager without leaving the platform.

Two new tools power this update. The first is TikTok in Canva Apps Marketplace, which lets users design videos in Canva and send them directly to TikTok’s asset library or Ads Manager. The second is TikTok in Canva Grow, an AI-powered campaign builder that generates ad creatives from business inputs and publishes them in a few clicks.

Together, these features streamline the full workflow from idea to live campaign. Businesses can create, optimize, and launch TikTok ads faster, while keeping creative control inside Canva.

This update reflects a broader push toward integrated marketing tools, helping brands move from content creation to campaign execution with far less friction and faster speed to market.

Adobe introduces Firefly AI Assistant – a new era of agentic creativity 

Adobe is transforming design workflows with the introduction of Firefly AI Assistant. Moving beyond simple text-to-image prompts, this “creative agent” acts as a conversational partner capable of orchestrating complex, multi-step workflows across the entire Creative Cloud.

Instead of manually navigating menus in Photoshop, Premiere, or Illustrator, users can simply describe their vision, and the assistant executes the technical heavy lifting behind the scenes.

The assistant features Creative Skills, allowing creators to trigger task sequences like resizing images for multiple social platforms, animating stills, or optimizing file formats with a single command. It maintains context as you move between different Adobe apps and learns your personal aesthetic preferences over time. 

Importantly, all outputs remain fully editable in native Adobe formats, ensuring professionals keep pixel-level control while the AI handles the repetitive, busy work of production and feedback loops.

In the AI Realm 

Automating workflows in Chrome with “Skills”

Google is rolling out a new feature called Skills in the Chrome browser. This feature is designed to make AI-powered browsing faster and more efficient. Instead of rewriting the same prompts across different websites, users can now save frequently used prompts and reuse them instantly.

Here’s the idea. If you create a useful prompt, like comparing product specs or summarising long documents, you can save it as a Skill. The next time you need it, just trigger it with a simple command. It will run across the page you’re viewing and even multiple tabs.

According to Google, early adopters are building practical workflows with some analyzing nutrition in recipes and others using the feature to compare products when shopping. 

Additionally, Chrome is also launching a ready-made Skills library. Users can pick pre-built prompts for tasks like gift selection or product breakdowns, then tweak them as needed.

Anthropic has redesigned the Claude desktop app 

Anthropic has announced a redesign of the Claude desktop app. The redesigned version focuses on how developers actually work today. Instead of handling one task at a time, the app now supports running multiple coding sessions in parallel, making it easier to manage complex workflows across projects.

A new sidebar keeps all active and recent sessions organised in one place. Developers can switch between tasks, filter by project or status, and quickly resume work. Sessions automatically archive once completed, keeping the workspace clean.

Also, the app now includes built-in tools like an integrated terminal, file editor and others. This means developers can review, edit, and ship code without leaving the app. A flexible drag-and-drop layout also lets users customize their workspace based on how they prefer to work.

In addition to the updates in design, Anthropic has also added some performance updates. 

OpenAI makes progress in cybersecurity with GPT-5.4-Cyber

Last week, Anthropic announced Project Glasswing, reinforcing AI’s strength in cybersecurity. This week, there’s news from OpenAI. The company is scaling its cybersecurity efforts with an expanded Trusted Access for Cyber (TAC) program and a new model variant, GPT-5.4-Cyber. The goal is to give defenders stronger AI tools while keeping safeguards in place.

The TAC program is now opening access to thousands of verified individuals and organisations. Users who meet identity and trust criteria can unlock more advanced capabilities designed specifically for cybersecurity tasks.

GPT-5.4-Cyber is built to support defensive use cases like vulnerability detection and even analysing compiled software without source code. It has fewer restrictions for legitimate security work, but access is tightly controlled and rolled out gradually.

These moves from major players in the AI segment signal a future where AI becomes central to cybersecurity workflows, meaning organizations must invest in smarter, AI-driven defence systems to stay ahead.

Personal intelligence in Gemini is now expanding to more countries globally 

In January, Google introduced Personal Intelligence in Gemini a few months ago to Google AI Pro and AI Ultra subscribers in the U.S. Now Personal Intelligence is expanding to more users across the globe.

Personal Intelligence gives Gemini access to data across its ecosystem to deliver highly personalised responses. The feature connects with apps like Gmail, Drive, Calendar, Photos, Maps, and YouTube to understand user preferences without needing repeated prompts.

The goal here with the introduction of Personal Intelligence is to make Gemini more context-aware and proactive. With the update, Gemini can now recommend products based on past purchases, troubleshoot devices using receipt data, and suggest travel plans tailored to your schedule, location, and interests.

But it’s not just that. Users can also discover new hobbies or explore personalised travel ideas based on past behaviour and subtle preferences. Google is handing users full control over which apps Gemini can access and the option to turn off Personal Intelligence when they like. 

Automate workflows in Claude Code with “Routines” 

Anthropic has launched Routines in Claude Code, a new feature that lets developers automate repeatable tasks across their workflow. Currently in research preview, routines can run on a schedule, via API calls, or in response to events like GitHub activity.

Here’s how it works. Developers set up a routine once by defining a prompt, connecting repositories, and choosing triggers. From there, Claude handles tasks automatically in the background, without needing a local machine running.

Scheduled routines can fix bugs overnight or manage backlogs. API-triggered routines can respond to alerts, verify deployments, or triage issues in real time. GitHub-based routines can review pull requests, flag changes, and even respond to updates as they happen.

Teams with early access are automating documentation updates, running code reviews, and syncing changes across multiple codebases. Each routine runs as its own session, adapting as new inputs or events come in.

Google announces a native Gemini Mac app 

Google has introduced the Gemini app for Mac, offering a native AI experience built for everyday desktop workflows. This brings AI closer to where users already work without needing to switch tabs or juggle multiple tools to get answers and execute tasks. 

Gemini Mac app allows users to write, brainstorm, and code directly from their desktop. A key feature is the global shortcut Option + Space, which instantly opens Gemini from anywhere on the system. This makes it easy to get quick help without interrupting focus.

Another standout feature is contextual assistance. Users can share their active window with Gemini, giving the AI real-time context. This means it can summarise text, debug code, or analyse complex visuals based on what’s on screen.

All these features are available for access when the user grants relevant permissions on their Mac like screen recording and accessibility. Google is working on making Gemini a more proactive desktop agent. 

Google introduces more expressive AI speech with Gemini 3.1 Flash TTS 

Google has announced its next big update in the text-to-speech realm with the launch of Gemini 3.1 Flash TTS. This new model is designed to deliver more natural, expressive, and controllable AI-generated voice. It’s rolling out in preview across developer tools, enterprise platforms, and Google Workspace.

According to Google, the model enhances speech quality and ranks highly on industry benchmarks. It supports over 70 languages, making it suitable for global applications. It also enables multi-speaker conversations, allowing more dynamic and realistic dialogue.

In addition to all these, the model also introduces audio tags. These let users control tone, pacing, and delivery using simple instructions within the text. Developers can guide how a voice sounds, shift emotions mid-sentence, and create consistent character voices across projects.

To address safety concerns, all generated audio is watermarked using SynthID, helping identify AI-generated content.

Anthropic introduces Claude Opus 4.7

Anthropic has released Claude Opus 4.7 to bring major improvements in advanced coding, task reliability, and multimodal capabilities. The model is now generally available across Claude products, APIs, and major cloud platforms.

The biggest upgrade is in software engineering. According to Anthropic, Opus 4.7 can handle complex, long-running coding tasks with minimal supervision. It follows instructions more precisely and even checks its own outputs before responding thus improving trust in results.

The model’s vision capabilities have also improved significantly. Hence, the model can process higher-resolution images, enabling more detailed analysis of diagrams, screenshots, and visual data. This opens up new use cases in areas like data extraction and UI design.

Social Media Feature Updates and Algorithm Changes  

X launches Cashtags 

X has introduced a new Cashtags feature designed to make financial discussions more interactive and data-driven directly within the platform.

The update allows users to search or post cashtags and get automatic suggestions for matching stocks or crypto tokens. This helps ensure users are referencing the correct asset while posting or browsing.

When a cashtag is tapped, users can instantly view related posts along with a live price chart. This removes the need to switch between apps, keeping both market data and community insights in one place. The feature is currently available globally for iOS. 

Snap x Qualcomm announce next-gen AR glasses 

Snap’s subsidiary, Specs Inc. has announced a multi-year partnership with Qualcomm to power its upcoming AR glasses, Specs. The collaboration will use Qualcomm’s Snapdragon chips to deliver high-performance, low-power computing for immersive digital experiences.

Specs are designed as standalone, see-through smart glasses that blend digital content with the real world. Users will be able to see, hear, and interact with augmented experiences in real time, without needing a separate device.

The partnership builds on years of collaboration between Snap and Qualcomm, which previously worked together on earlier Spectacles devices. 

Snap also plans to support a growing ecosystem of creators and partners building for Specs, making it easier to scale new AR applications over time.

A Final Word…

We know that’s a lot to take in but that’s exactly why you have KIMP’s Picks. Whether you are a marketer trying to keep up with the social media updates or an entrepreneur wondering what’s happening in the world of AI, KIMP’s Picks is designed to give you the highlights without feeling overwhelmed. We’ll be back next week with another round of updates, until then, keep experimenting and stay curious!