{{first_name | Hi}}, your strategic AI update is here.
Wall Street is questioning June's 57,000-job numbers, Microsoft bet $2.5B on embedding its own engineers inside customer operations, and Zuckerberg told staff Meta's AI agents have not delivered. Buying the tools is the easy part. The 73% are not behind on budget; they are behind on operating model design, and that is what the 10% solved first.
"If you want to build a ship, don't drum up people to gather wood, divide the work, and give orders. Instead, teach them to yearn for the vast and endless sea."
— Antoine de Saint-Exupéry, French writer, poet, and pioneering aviator
In today's lineup:
Robots
73% use AI. Only 10% run on it.
Real ROI from Gen AI
When AI advises vs. when AI decides
People
Jon Wang turns patient access into a 24/7 AI operation
Love
How to embed AI without a change management war
The Coming Wave by Mustafa Suleyman
Reading time: 3 min.
ROBOTS 🤖
How are robotics and AI changing industries? We break down the latest news, tools, and innovations for you.
Top Insights
Wall Street analysts question the US government's June jobs numbers as payrolls add only 57,000 against a 115,000 forecast and tech layoffs reach 142,000 year-to-date.
Microsoft Frontier embeds 6,000 engineers inside enterprise operations at Unilever and Novo Nordisk, backing $2.5B on implementation over models to close the AI ROI divide.
Zuckerberg tells staff AI agents have underdelivered despite $145B in planned infrastructure and 7,000 employees redirected to Meta's agent transformation unit.
Tesla caps employee AI spending at $200 per week as token-based billing turns prompt habits into capital budget problems for operations teams.
OpenAI proposes giving the US government a 5% equity stake as voluntary AI model standards approach the August 1 deadline.
73% use AI. Only 10% run on it.
A new Publicis Sapient Global Enterprise AI Report of 1,550 AI decision-makers across six countries finds that 73% of organizations use AI regularly. Only 10% say AI is core to how their business operates.
Nigel Vaz, CEO of Publicis Sapient, names the root cause: "The enterprise was not designed for the speed, scale and autonomy that AI makes possible."
That 63-point spread comes down to organizational design.
Three moves the 10% make differently:
Redesign who owns decisions and how output is measured before deploying the next tool
Question whether the process should exist before automating it
Fund governance and workflow redesign on the same budget line as deployment
The 73% bought access. The 10% rebuilt the architecture.
What stops AI from becoming core to your operations?
Real ROI from Gen AI
How human-centric AI delivers business results
BMW ran humanoid robots through 30,000+ X3 builds at Plant Spartanburg over 11 months and is now deploying next-gen Figure 03 into logistics sequencing as part of a Physical AI production strategy.
Mastercard launched Agent Pay for Machines across 30+ partner platforms, enabling machine-to-machine transactions at network speed with no human approval loop.
Fluence improved engineering efficiency 30–50% deploying Google Gemini Enterprise across its teams, cutting technical document creation time across energy storage projects.
Tool Spotlight
Your weekly briefing on tools that create competitive leverage
When AI advises vs. when AI decides
Taktile raised $110M in Series C funding led by Goldman Sachs, expanding its agentic decision platform for financial institutions into underwriting, fraud, onboarding, and compliance.
Best for: Finance, insurance, and fintech teams that need AI to close decisions at speed and scale without routing every edge case to a human.
Choose if you...
Run lending, underwriting, or fraud operations and need AI to execute the decision, not just flag a risk score.
Have risk or compliance teams that spend more time reviewing AI recommendations than acting on them.
Are moving from pilot-stage decisioning experiments to production workflows handling thousands of cases daily.
Best for: Large enterprises with established risk governance frameworks that need real-time decisioning layered on top of legacy systems, bureau data, and existing models.
Choose if you...
Operate in a heavily regulated environment where every decision requires an auditable logic trail.
Have existing credit or risk models in production and need an orchestration layer, not a replacement.
Manage a diverse global portfolio where different markets require different decision logic from one platform.
3 other tools to explore
Alloy suits banks and fintechs who need identity, KYC, and onboarding orchestration upstream of the credit decision.
Zest AI suits US credit unions and banks who need an ML-powered credit scoring layer that extends into automated approval.
IBM Operational Decision Manager suits large enterprises with deep rule-based decision architectures that want to add AI on top of existing governance frameworks.
PEOPLE 👥
Meet the innovators turning bold ideas into real-world impact.
Transformation Champion
Jon Wang turns patient access into a 24/7 AI operation
Jeffery Liu and Jon Wang, Co-Founders & Co-CEOs, Assort Health
Jon Wang trained as a physician at UCSF before moving into machine learning, watching call centers bleed care time to administrative backlog. He co-founded Assort Health to fix the patient intake layer with AI: autonomous call routing, staff copilots, and a shared Patient Journey Memory. The company raised $120M Series C this past week at a $1.2B valuation; customers report 97% fewer hold times and 115% more labor capacity.
LOVE ❤️
Practical wisdom, growth tactics, and a must-read book that will challenge the way you think.
How to embed AI without a change management war
Boeing dropped a planned manufacturing system upgrade and ran AI inside the tools their teams already used. The result: $92M saved in training costs, quality inspection times cut 75%, with no one retrained on anything.
That is the Invisible Upgrade: the system becomes more intelligent without becoming more complicated for the teams.
Three ways to apply it:
Map every screen your team uses today. AI should improve what happens inside those screens, not add new ones.
Measure success by what your team stopped doing manually, not by what system they learned.
If the rollout needs a training program, redesign the rollout.
What would it take to make your next AI deployment invisible to the teams it helps most?
The full story is in chapter 4 of 100x.
Transformative Reads
One book, handpicked from my conversations with friends, industry leaders, and tech innovators:
In The Coming Wave, Mustafa Suleyman (co-founder of DeepMind, CEO of Microsoft AI) argues that organizations treating AI as a gradual shift are miscalibrated. Cost curves, capability curves, and adoption curves are compressing together. The 73% are planning for a transition the 10% already completed.
Perfect for: CEOs and boards deciding how much urgency their AI strategy actually reflects.
In Culture
73% of organizations are this meme. The 10% skipped the chant and redesigned the org.
THANK YOU!
Lighthouse takes us many hours to create, and we don’t run ads or charge anything. We are dedicated to making it the best AI transformation newsletter possible.
We just ask for one thing: tell one friend about Lighthouse today!
Want daily AI & Innovation insights? Connect with me on LinkedIn or X.
Got suggestions? Just reply to this email!
Much Love,
Matt and the Future Works team
