Google Cuts Agent Token Costs, BMS Buys Nvidia Pods, US Tests Health AI

Sponsored by

AI HUSTLE | July 23, 2026

Welcome to this week's edition of AI Hustle. As business operators, we all know that the ultimate leverage isn't just about working harder—it’s about building automated systems that run, think, and scale in the background while you focus on high-level strategy. Today, we are breaking down a brilliant growth workflow that turns the tedious process of manual influencer outreach into a fully automated customer-acquisition machine. Plus, we’ll dive into how the latest hardware and model upgrades from Google and Nvidia are drastically lowering the cost of running these automated systems. Let's get into it.

Leading AI expert shares one email per day on what’s new & what’s next

AI moves faster than any single newsletter can cover. To actually stay current you'd need to read every major AI newsletter, listen to the AI podcasts that move thinking, watch what every AI lab ships, and scan public-company earnings calls, investor letters, and GitHub. Almost no one has time for that. 

MavSource fixes this in 5 minutes a day. We aggregate all of those sources, summarize what matters, analyze the trends shaping AI — across both public and private markets, apps and infrastructure — and send it as one daily email with founder commentary on what the day's signal actually means. 

No checking dozens of sources. No hype. One brief, every morning. Built for the people who can't afford to fall behind. Free.

Get The Free Brief →

The Hustle: The Micro-Army Machine: How to Automate Niche Influencer Recruitment at Scale

The Goal: Find and recruit hundreds of niche, highly engaged creators to promote your product without spending a single second on manual vetting, emailing, or order setup.

The Tools:

* Apify / Phantombuster: For automated social platform scraping.

* Make.com: To orchestrate the data flow and connect your apps.

* OpenAI (GPT-4o) / Anthropic (Claude 3.5 Sonnet): To analyze creator profiles, calculate authentic engagement, and write hyper-personalized pitches.

* Hunter.io / Lemlist: For automated email outreach and sequence tracking.

* Shopify / ShipStation: To automate product fulfillment once a creator joins.

Step 1: Pinpoint the Targets (The Input)

Set up a daily recurring scraper in Apify or Phantombuster. Configure it to search Instagram or TikTok for accounts with a follower count between 2,000 and 20,000 (the sweet spot for highly engaged micro-influencers). Have the scraper target specific keywords, hashtags, or competitor mentions relevant to your industry. Each day, the tool automatically extracts the creators' bio, follower count, email address (if public), and links to their latest 5 posts, and pushes this raw data into a Google Sheet.

Step 2: Vibe Check & Verification (The Trigger)

Once a new row is added to your Google Sheet, it triggers a Make.com scenario. Before doing any outreach, the system must verify the creator's value. Make.com calculates their true engagement rate by pulling the likes and comments from their latest posts. It screens out accounts with low engagement or signs of bot activity (such as hundreds of generic, single-emoji comments).

Step 3: The Custom Pitch Formulation (The AI/Logic)

Next, Make.com sends the filtered creator's bio, industry niche, and recent captions to GPT-4o or Claude 3.5 Sonnet. The AI performs two critical tasks:

1. Brand Alignment Match: It scores the creator from 1 to 10 on how well their content matches your brand's aesthetic and voice.

2. Dynamic Copywriting: If the score is an 8 or higher, the AI drafts a highly personalized outreach email. It references a specific topic from one of their recent posts to prove it's not a generic blast, and outlines a tiered affiliate structure (e.g., a free product to review plus 15% commission on any sales they generate).

Step 4: Secure the Partnership (The Output)

The personalized pitch is automatically sent via your outreach tool (like Hunter.io or Lemlist). When a creator replies and agrees to the deal, an AI-powered email classifier detects their positive intent and replies with a link to a simple form where they can input their shipping details. Once submitted, Make.com automatically generates their unique affiliate discount code and pushes a free product order directly into your Shopify or ShipStation queue for shipping.

Why This Hustle Works:

* Unmatched Authenticity: Micro-influencers boast up to 60% higher engagement rates than mega-influencers, giving you a highly trusted, hyper-targeted sales force.

* 100% Hands-Off Scaling: It completely eliminates the hours spent manual-scrolling social media, hunting for emails, typing custom pitches, and manually creating trial orders.

* Zero-Risk Growth: You only send products and pay commissions to creators who are highly aligned and actively agree to promote your brand.

What is an EOR—and why are companies using it?

Opening entities in every country can be slow, expensive, and hard to scale.

That's why more companies are using EOR to hire globally faster.

See how Oyster helps teams hire, pay, and support talent in 180+ countries while staying compliant along the way.

Get all the answers

🚀 The AI Pulse: 3 Signals to Watch This Week

Google's Gemini 3.6 Flash & 3.5 Flash-Lite Target Enterprise Agent Token Costs

Google has released Gemini 3.6 Flash and 3.5 Flash-Lite, specifically targeting the high token costs and latency associated with running autonomous enterprise agents. Because running background agents through endless reasoning loops can quickly rack up massive API bills, Google focused heavily on efficiency. The new 3.6 Flash model boasts a 17% reduction in output token generation compared to previous versions (and up to 65% in developer tests), while costing only $1.50 per million input tokens. Meanwhile, the ultra-fast 3.5 Flash-Lite operates at 350 output tokens per second, making it the fastest in its class for high-volume, low-latency background processing. Platforms like Figma, Harvey, and Hebbia have already integrated these models for complex design prototyping and legal document parsing.

The Hustle Take: The economics of AI are shifting rapidly in favor of the operator. Up until now, building highly active, multi-step autonomous workflows (like our Micro-Army Machine above) carried the risk of unpredictable, runaway API costs. With Google aggressively slashing token costs and introducing tools like "native computer-use" directly into the API, you can now build deeper, more complex agentic workflows that run 24/7 for a fraction of what it cost just six months ago.

Bristol Myers Squibb Buys Nvidia AI System for Drug Discovery

Pharmaceutical giant Bristol Myers Squibb (BMS) has purchased Nvidia’s newly introduced DGX SuperPOD system, built on the advanced Vera Rubin architecture. BMS plans to use this massive computing cluster to scale its "Predict First" workflow. Instead of manually synthesizing and testing thousands of molecules in a physical lab, researchers use AI-generated predictions to exclude unviable molecules before they are ever physically created. This approach has already accelerated their early-stage drug development timeline by 20% to 30%, with expectations to hit a 50% reduction in the near future.

The Hustle Take: Do not look at this as just a "science story." The "Predict First" framework is a masterclass in risk mitigation for any business. Whether you are launching a physical product, a SaaS tool, or a marketing campaign, the goal should always be to use AI to simulate, predict, and stress-test your ideas before spending a single dollar on physical execution, manufacturing, or ad spend.

US Public Health Departments to Test OpenAI and Anthropic Models

In a massive step forward for public sector AI adoption, 10 US public health departments are set to pilot generative AI tools under a new program called PULSE (Public Health Use Case and Learning Scaling Engine). Supported by the Coalition for Health AI (CHAI), Accenture, OpenAI, and Anthropic, the initiative will provide up to 2,000 public health practitioners with enterprise-grade access to top-tier AI models. The program will test five critical use cases, including automated clinical-data retrieval, biosurveillance, and multilingual public communications, with the ultimate goal of publishing concrete, compliant AI implementation playbooks for public agencies by 2027.

The Hustle Take: When highly regulated, deeply risk-averse institutions like public health departments start actively deploying generative AI, it signals a massive shift in market trust. For business operators and B2B developers, this is your cue that the market for highly secure, compliance-first AI integrations is wide open. If you can build AI tools that address strict data privacy and governance, you will find a highly lucrative customer base waiting in industries that were previously too afraid to adopt the technology.

Go from AI overwhelmed to AI savvy professional

AI will eliminate 300 million jobs in the next 5 years.

Yours doesn't have to be one of them.

Here's how to future-proof your career:

• Join the Superhuman AI newsletter - read by 1M+ professionals

• Learn AI skills in 3 mins a day

• Become the AI expert on your team

Start learning AI now