In partnership with
Your growth team woke up to a briefing they didn't ask for.
Monday 7am. Three messages in #growth.
Stripe revenue by channel, Meta and Google spend reconciled against GA4, Klaviyo flow performance, Shopify AOV by source. Posted by Viktor at 6am.
The campaign brief he wrote sits in #campaigns. Brand monitoring scrape runs every six hours. Competitor pricing update lands every Friday.
Your media buyer, content lead, and CMO open Slack to the same prepared room. 3,000+ integrations including every ad platform, CDP, and CMS you run.
"Viktor is like the most capable all-round colleague you can imagine." Sam, CEO, Givr.
Hey,
Big week. Brussels forced open the most valuable real estate in AI, Google's flagship slipped again, and two open models are about to land that could reset what you pay for high-volume work. Then: 13 new tools in the directory, and the guides we published this month.
This week in AI
1. Brussels pried open Android for rival AI assistants
July 16 · European Commission · binding
The European Commission issued two binding Digital Markets Act decisions against Google. The first forces Google to open 11 Android features that only Gemini can fully use today. In practice: you will be able to summon a rival assistant by voice, the way "Hey Google" works now, and let it act inside your apps, from booking a taxi to replying to messages. The second requires Google to share anonymized search signals (ranking, query, click, and view data) with competing search engines and AI chatbots on fair, reasonable terms.
The dates: search data sharing begins January 2027, and the Android changes reach users with the next Android release in 2027. No fines came attached, though non-compliance can cost up to 10% of global annual turnover. Google's Kent Walker objected that the decisions risk undermining privacy and security guardrails for millions of Europeans.
Why it matters: which AI assistant lives on your phone has been a distribution decision, not a quality one. This is the first ruling that turns it into an actual choice, and every assistant that is not Gemini just got a legal path onto two billion devices.
2. Gemini 3.5 Pro slips a third time
Reported · unconfirmed by Google
Google's flagship reportedly missed a July 17 target, the third slip, and the company is said to be weighing a stopgap Gemini 3.6 Flash release to put something in market. Google has published no official model card, pricing, or benchmarks, so treat every circulating spec as unconfirmed. Alphabet shares fell about 4% on the delay reports.
Why it matters: if you have been holding a decision waiting for Gemini's next flagship, stop waiting. Buy on what ships today, not on a roadmap.
3. The open-weight countdown: July 24 and July 27
Two dates worth marking
DeepSeek's V4 stable release lands July 24, ending the preview churn that kept cautious teams off it. Three days later, on July 27, Moonshot AI releases the open weights for Kimi K3, a 2.8-trillion-parameter model that just took the top spot on a major coding leaderboard. The honest caveat: K3 sits around ninth on general chat, so it is a coding and agent specialist rather than an all-round replacement.
Why it matters: a free specialist that beats paid generalists at the one task you actually run all day changes the budget math. If you are spending heavily on API calls for high-volume coding, the last week of July is the moment to benchmark.
4. AI detectors have a blind spot, and it is measurable
Epoch AI study
Epoch AI tested Pangram, GPTZero, and Originality.ai on 495 human passages from 99 authors, all written before ChatGPT existed, plus AI text in two flavors. On plainly prompted AI writing, all three were near flawless: false negatives topped out at 0.7%. Then the researchers fed models five samples of a specific author's work and asked them to imitate the style. About 13% of those AI passages slipped through on average, with individual detectors missing between 10% and 18%. Scientific writing was the weak spot, at roughly 26% missed.
On genuine human text the picture flipped: Pangram and GPTZero wrongly flagged nothing, while Originality.ai flagged 19 of 495 human passages, about 3.8%.
Why it matters: detectors are strong against lazy AI writing and weak against anyone who bothers to prompt for a style. If a detector result is being used to make a decision about you or your work, these are the numbers to bring.
Also worth knowing
Oracle is cutting up to 30,000 jobs, roughly 18% of its workforce, to free an estimated $8 to $10 billion a year for its Stargate data-center buildout with OpenAI and SoftBank. The clearest accounting yet of what the AI infrastructure boom actually costs.
SAP completed its Prior Labs acquisition and committed over 1 billion euros across four years to tabular foundation models: AI built for spreadsheets, ledgers, and databases rather than prose. An unglamorous category that may quietly outperform another chatbot.
Microsoft is preparing Project Perception, an AI security platform that routes cheap models for triage and calls frontier models only for hard reasoning. The cost engineering is the point: it makes always-on vulnerability scanning affordable instead of theoretical.
New in the directory
Heads up: the tool links in this section are affiliate links. If you sign up through them, we may earn a commission at no extra cost to you. As always, it never affects our scores or what we recommend.
AI Video
InVideo · Type a prompt, get a finished video with scenes, voiceover, and music. It routes to models like Kling, Veo, and Runway, so it doubles as a one-stop workflow while the model leaderboard keeps reshuffling.
PixVerse · Fast text- and image-to-video. Stands out on stylized and anime looks and effect-driven shorts rather than photoreal footage.
Hailuo AI · One of the stronger free tiers in text-to-video, with motion quality that punches above its price.
AI Image & Photo
Skylum (Luminar Neo) · A desktop AI photo editor built for photographers: one-click sky swaps, relighting, and precise AI masking, without a subscription treadmill.
YouCam · AI photo and selfie studio: professional headshots, retouching, makeup try-on, and avatars, mostly mobile-first.
OpenArt AI · Many image models under one login, plus consistent characters and custom style training for people who need the same face or look across a series.
Audio & Notes
HappyScribe · Transcription and subtitles in 100+ languages, with both AI-fast and human-accurate options on the same platform. Useful when "close enough" is not enough.
HyNote AI · An all-in-one AI note-taker. Record a meeting or drop in PDFs, videos, and links, then get summaries and action items you can chat with across everything you have saved.
Slides & AI Office
Present Intelligence (Pi) · Turns a prompt or an existing document into a structured, interactive deck rather than static slides, in about two minutes.
Skywork AI · AI office agents that produce source-cited docs, slides, spreadsheets, even podcasts and web pages from a single brief. The citations are the differentiator.
Flowith · An agentic AI canvas. Instead of one chat thread, agents research, write, and build in parallel across an infinite workspace.
Build & Grow
Appy Pie · No-code builder for apps, websites, chatbots, and automations, now AI-assisted end to end.
Klaviyo · Email and SMS marketing built for ecommerce, with the segmentation and automation flows DTC brands actually rely on.
Published this month
New guides, all pricing verified
How to Track Your Brand's AI CitationsWhat to measure across ChatGPT, Google AI Overviews, Gemini, and Perplexity, the free manual method, and how to do it at scale.
Best AI Marketing Tools in 2026: 12 Picks Worth Paying ForHubSpot AI, Semrush, ActiveCampaign and nine more, with verified pricing and the cons vendors leave out.
Best AI Accounting Software in 2026Six tools that genuinely automate the books: Digits for bookkeeping, Ramp for spend, DataSnipper for audit.
And the free-tier series. We went through the free plans line by line and ranked them by what you can actually do without paying:
Chatbots Writing Coding Note-takers Image-to-video Voice Music Logos Resumes Detectors Presentations
10x the context. Half the time.
Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.
That is the issue. One careful email a week. No hype, no ads.
ToolChase is reader-supported and ad-free. Some links in this email are affiliate links, and they never influence our scores or rankings.