• Everyday AI
  • Posts
  • Ep 840: The Next 12 months of AI: 19 Predictions Every Business Leader Needs to Hear

Ep 840: The Next 12 months of AI: 19 Predictions Every Business Leader Needs to Hear

OpenAI CRO leaves, Google releases Gemini 3.7 Flash, OpenAI and Cerebras team up for speedy GPT-5.6 and more

Author

Everyday AI
August 13, 2026

 

👉 Subscribe Here | đź—Ł Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI

Sup y’all 👋

Starting next week, the Everyday AI newsletter and livestream is taking a lil break. 

With the school year approaching, we’re going to go back to school to kick off our popular Start Here Series from the beginning. 

This is a great opportunity to follow along on the 30-part series if you’re looking to go from AI Newbie to AI Champion.

So if you’re still looking for your daily dose of Everyday AI, make sure to follow the Start Here Series to completion on Spotify or Apple Podcasts. 

AI University is back in session and starts Monday. Don’t be tardy. 

✌️

Jordan 

Outsmart The Future

Today in Everyday AI
7 minute read

🎙 Daily Podcast Episode: AI is changing fast. Here are 19 predictions for the next 12 months. Give today’s show a watch/read/listen.

🕵️‍♂️ Fresh Finds: OpenAI says frontier firms are getting 8X more from AI, DeepSeek open-sourced its agent framework, and Google AI Studio added an Agents tab. And more. Read on for Fresh Finds.

đź—ž Byte Sized Daily AI News: OpenAI CRO leaves, Google releases Gemini 3.7 Flash, OpenAI and Cerebras team up for speedy GPT-5.6 and more. Read on for Byte Sized News.

đź’Ş Leverage AI: AI is changing too fast for a traditional roadmap. Here are the priorities that should shape your company’s next nine months. Keep reading for that!

↩️ Don’t miss out: Miss our last newsletter? We covered: Grok 4.6 drops with impressive benchmarks, Google expands Gemini connectors, Grok Bot brings competition to Codex and more. Check it here!

Ep 840: The Next 12 months of AI: 19 Predictions Every Business Leader Needs to Hear

The next 12 months of AI leaked.

Kinda.

For the past 90ish days, we've been quietly collecting evidence of what's next.

1,030 saved posts. 90 Podcasts. Countless conversations. Every model drop, every leak, every quiet product update the big labs hoped you'd scroll past.

Then we connected the dots.

What came out the other side: 19 calls on where AI goes over the next 12 months. And some of them are uncomfortable.

Bring your team's AI roadmap. You'll want to edit it. 👇

Also on the pod today:

• Reactive chat is dying 📉 
• Voice becomes default AI control 🎤 
• Proactive agents work overnight 🤖 

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – Dograh is an Open Source workflow builder for Voice Agents, Assembly Studio is The platform firms run on and build on, Unsloth Desktop is The first desktop app to run and train AI models.

OpenAI Reports — Frontier firms are using AI way more deeply, generating over eight times the output per user than typical companies.

DeepSeek Open Source Framework — DeepSeek just dropped its open-source agent framework and AI model, both under MIT license.

AI Studios Agents Tab — Google's AI Studio is rolling out a new Agents tab to streamline everything from agent management to artifact editing in one spot.

Alibaba New Model — Alibaba dropped Qwen3.8-2.4T-A95B, a 2.4 trillion parameter model with near-frontier abilities.

Twitch AI — Twitch now lets streamers opt out of having their content used to train Amazon's generative AI.

Gemini Notebook Upgrade — Google’s Gemini Notebooks now let you copy any notebook, sources and all.

DeepSeek V4 — DeepSeek just dropped its 1.6T-parameter V4 Pro model with open weights and a 1M-token context window.

Google Health Guardian — Google Health Guardian now tracks subtle health changes like insulin resistance right on your Pixel or Fitbit.

Google Pixel 11 — Google’s Pixel 11 lineup puts its smartest AI yet front and center, aiming to change how you use your phone.

Anthropic Eyeing $2 Trillion — Anthropic could smash IPO records with a $2 trillion+ debut, investors say.

Microsoft Data Center — Microsoft just got its first 50MW AI data center from IREN in Texas, part of a $9.7B deal.

Anthropic and Decart — Anthropic is eyeing a $6 billion buyout of Decart AI to boost its chip efficiency and handle soaring demand.

Google Pixel Watch 5 — Google's new Pixel Watch 5 drops with health tracking upgrades, tougher display, and fresh sizing but still faces battery life criticism.

1. OpenAI CRO Denise Dresser exits with IPO around the corner đź‘€

OpenAI said Thursday that Chief Revenue Officer Denise Dresser is leaving less than a year after joining, handing enterprise sales leadership to former Wiz president and COO Dali Rajic. Her exit follows longtime executive Brad Lightcap’s departure announcement days earlier, adding to a recent run of senior leadership changes.

The timing is especially notable because OpenAI is preparing for a potential IPO and needs its business-sales operation to keep pace with rival Anthropic. Dresser will remain briefly to support customers and transition responsibilities.

2. IBM and OpenAI Partner to Bring GPT-5.6 Into Enterprise Consulting 🤝

Announced August 13, IBM says it will embed OpenAI tools, including GPT-5.6, Codex, and ChatGPT Work, into its Consulting Advantage platform as companies rush to put generative AI into everyday operations.

The deal also creates a dedicated IBM OpenAI Practice staffed by thousands of trained consultants and engineers, aimed at sectors where security and old systems remain major hurdles.

3. Google Rolls Out Gemini 3.7 Flash Weeks After Its Last Update ⚡

Google is releasing Gemini 3.7 Flash today, replacing the three-week-old 3.6 Flash with a cheaper introductory price and stronger coding and workflow automation scores.

The update improves several benchmarks, particularly software engineering and complex-document processing, but it is an incremental step rather than the flagship Gemini 3.5 Pro launch Google had promised for June. The quick refresh keeps Google’s AI lineup moving while questions remain about whether the company’s biggest model upgrade has slipped further behind schedule.

4. Apple Seeks Publisher Deals to Feed Siri AI With Live News 🎤

Apple is in talks with publishers on multiyear content deals that could give its upgraded Siri access to current news and information, according to the Wall Street Journal.

The proposed agreements may pay publishers when Siri uses their material, with Apple reportedly discussing a nine-figure budget.

5. Claude in Chrome Now Syncs Cowork Sessions Across Devices 📲

Anthropic has turned its Chrome side panel into a full Claude Cowork session, letting users carry browser-based tasks and conversation context across its desktop, web, and mobile apps.

The update is available now for Max and Team subscribers, with Pro access rolling out in coming weeks, while Enterprise admins can keep it disabled or restrict it to approved sites.

6. OpenAI Previews GPT-5.6 Sol “Ultrafast” Tier at Up to 750 Tokens Per Second ⚡

OpenAI is previewing an API-only Ultrafast service tier that it says can run its GPT-5.6 Sol model up to 14 times faster than standard processing, according to 9to5Mac.

Powered by Cerebras, the service is aimed at tasks where responses need to arrive almost immediately, including voice systems, support tools, coding agents, finance, and security work. The company says the speed has already helped its teams turn overnight research and incident-analysis jobs into several workday iterations.

Annual AI predictions are pretty much useless now. Models ship two to three times faster than they did nine months ago, so a December roadmap is kinda stale by February.

So what now?

With the school year about to kickoff, we skipped the syllabus and went straight to the answer sheet: 19 things your company should focus on for the entire year ahead.

This breakdown might be the closest thing to a cheat sheet for your company's next nine months. We don't say that lightly.

(You know, we kinda do this EVERY DANG DAY.)

On today's Everyday AI, we hand over all 19 answers for the year ahead, from proactive agents and multiplayer workspaces to the return of token maxing.

Let’s dive in shorties.

1. Stop Prompting. Start Reviewing Overnight Work 🔥

The new normal is waking up and reviewing what your AI already shipped overnight, because Codex and Claude Cowork pretty much made the around the clock agent real.

What does this actually look like?

You hand an agent tomorrow's deliverable at 5 PM, and a manager thread, think chief of staff, coordinates the other threads while you sleep, so context stops living in one lonely chat.

The gain is leverage. Your team stops babysitting conversations and starts directing a staff that never clocks out.

Teams still typing one prompt at a time are gonna get lapped.

Try This

Tonight, hand one repeatable deliverable to an agent instead of a chat window: a first draft report, a weekly summary, a data cleanup, whatever eats your Tuesdays.

Tomorrow morning, review and redirect instead of starting from scratch, and feel that flip from operator to editor.

Make review-first mornings your Monday default, then watch how fast the habit spreads across your whole team.

2. Turn Repeat Wins Into Transferable Skills ⚡

Most teams stuff their best AI workflows into projects and call it progress. That's vendor lock-in wearing a trophy costume.

Skills are different. A saved skill is a repeatable process any agent can invoke, share, and rerun, which makes it real AI IP your company owns.

Why does that matter?

Skill reuse is about to become a legit company metric, the way SOPs once measured operational maturity. Lines of code and token leaderboards told us absolutely nothing.

Reuse tells you who actually turned AI into an owned asset, because every prompt marathon that ends in a great deliverable and never becomes a skill is compounding value left sitting on the table.

Try This

Pick one process your team repeats weekly, like a status report or a client brief, then save the winning workflow as a skill and share it with two teammates.

Then track one number: how many times it gets reused this month.

When reuse climbs, you're building an asset your company keeps. When it stays flat, you're just renting software with extra steps.

3. Quit Renting the Biggest Brain Available 🚀

Here's a spicy one: most of your work does not need the top model. Prolly 90% of knowledge work runs fine on 10% of a frontier model's intelligence.

So what's the move?

Model routing. Let the big model orchestrate the first and last 10% of a project while cheaper models grind through the middle.

Nobody needs a frontier model to rewrite an email, yet whole departments burn premium tokens on exactly that.

If your provider keeps prices high while everyone else races down, y'all should be shopping.

Try This

Audit one week of your team's AI usage and sort every task into heavy lifting versus routine work. Most of it lands in routine.

Route the routine 80% to a cheaper or faster model and watch the quality hold.

Then set a monthly scoreboard: usage should climb while cost per task falls. Get that trend moving before someone upstairs asks about the AI bill.

Reply

or to participate.