AI HAS A HYPE PROBLEM. WE DON'T.

The Executive Briefing · Twice weekly

The Executive Briefing — Tuesday, November 17, 2026

Living proof · Powered by AIXF.ai

Run by Axiom Intelligence. 24/7/365. For more than two years.

Every publication and podcast in the DX Today ecosystem is produced by AIXF.ai’s Axiom Intelligence, around the clock, every day of the year. This is not a demo of what AI could do. It is the live, two year proof of what Axiom Intelligence already does.

Key takeaways

  1. Menlo Ventures Puts Healthcare AI Spending at $3.7 Billion, Up 2.5x, as Incumbents Take 68% of New Scribe Dollars and Life Sciences Spend Jumps 23x. Treat the scribe market as a lesson in platform risk. When an incumbent bundles a capability, a standalone vendor can lose share even while its revenue grows, so negotiate contracts with exit terms and data portability. Put the next wave of budget where Menlo says growth is fastest, in clinical copilots, patient engagement and drug discovery, and prepare for payers to answer provider AI coding with AI of their own.
  2. Sierra Releases Poppy 1.0, a Protocol for Personal AI Agents, With Shopify Switching It On for Every Merchant and Meta's Muse Checking for It First. Plan now for customers who arrive through their own agents. Ask your web and API teams what it would take to publish a Poppy file, decide which actions you would allow an agent to take without a human approval, and write those rules down before an agent asks. Companies that make themselves easy for trusted agents to use will be easier to buy from.
  3. The Commerce Department's BIS Issues an Interim Final Rule on Remote Access to Advanced Chips, With a New License Requirement for Five Countries and an End to the AI Diffusion Rule. Inventory every contract where your AI workloads run on, or rent capacity from, data centers in Indonesia, Malaysia, Singapore, Thailand or Brazil, and identify any customer headquartered in Macau or Country Group D:5. Confirm whether the U.S. headquartered end user exclusion and its reporting steps apply to you, and consider filing comments within the 60 day window, since the rule is interim and its terms can still change.
  4. Fortune Reports Capital Economics Says the AI Stock Boom Is in Its Final Stages, as Growth Runs at 3.4% While Hiring Slows and Labor's Share Falls to Its Lowest Since 1947. Use the window Goldman describes to redesign entry level pathways and to retrain workers in the occupations the Apollo study flags as most exposed. Report productivity gains alongside workforce plans, because a rising output per worker and a falling labor share will draw scrutiny from employees, investors and regulators. Stress test budgets for a market in which public enthusiasm for AI valuations cools.

Welcome to the DX Today Executive Briefing

This edition follows money, protocols, rules and paychecks. Healthcare has become the largest vertical market for AI spending, a new protocol is trying to give personal AI agents a common way to deal with companies, Washington has written a new rule aimed at remote access to advanced chips, and the data on growth and jobs is raising hard questions about who captures the gains from AI.

Each story below is tied to a named source and a specific date, so you can see who said what. Where a figure comes from a report rather than from the company or agency it concerns, we attribute it to the outlet that reported it, and we leave out claims we could not trace.

01

Menlo Ventures Puts Healthcare AI Spending at $3.7 Billion, Up 2.5x, as Incumbents Take 68% of New Scribe Dollars and Life Sciences Spend Jumps 23x

Menlo Ventures published its report 2026: The State of AI in Healthcare, its second annual edition, on November 16, 2026, and the headline number is large. Menlo reported that healthcare AI spending reached $3.7 billion, up 2.5x in a year. The firm said healthcare is AI's largest vertical market, with spending more than double the next largest vertical, legal, at $1.8 billion. For executives who still think of healthcare as a cautious late adopter of software, the report points the other way: this is now where the largest pool of AI spending sits.

Where the money goes matters as much as how much there is. Menlo said 50% of healthcare AI spending goes to healthcare specific applications, about twice the enterprise average, and that providers account for 77% of spending, or $2.9 billion. Payers' external AI spending was $150 million, or 4% of the healthcare AI market. Providers are the main buyers, while payers remain a small external customer for AI vendors.

The scribe market is the clearest case study in how quickly the competitive map can change. Menlo said incumbents captured 68% of new scribe dollars and now hold 46% of the $950 million scribe market, up from 33% a year ago. Menlo said every major electronic health record vendor now offers its own scribe. JP Sanday, a Partner at Menlo Ventures, said Microsoft and Epic now capture roughly two of every three new scribe dollars. Menlo said AI native scribes grew 28% but lost market share. AI native vendors are learning how fast a platform owner can bundle a feature that once justified a standalone company.

Growth is also moving away from paperwork and toward care itself. Menlo said clinical copilots grew 5.5x to $550 million, and that patient engagement AI grew 4.5x to $450 million. The biggest jump was in life sciences, where Menlo said AI spending rose 23x, from $30 million to $700 million, the fastest growth of any healthcare AI segment. Menlo also said 51% of pharma companies now run AI at scale, up from 15% a year earlier.

The research and development numbers show optimism about direction and caution about speed. Menlo said 84% of R&D leaders expect AI to cut development timelines by at least a year over the next three to five years. Menlo said 93% of R&D leaders expect the first approval of an AI discovered drug within five years, but only 28% expect it within two years. Johnny Hu, a Principal at Menlo Ventures, said a new drug takes more than a decade and billions of dollars, with a 90% clinical failure rate. That context explains why even a modest improvement in discovery odds attracts so much capital.

The report also surfaces a tension between buyers on opposite sides of the same claim. Menlo said 43% of payers, and 57% of large payers, say provider AI coding raised their 2026 medical cost trend, up from 23% in 2025. Menlo said prior authorization is payers' top AI priority at 82%, up from 59%. In plain terms, providers are using AI to document and code more completely while payers are using AI to scrutinize and approve less automatically, and each side's spending is a response to the other's.

Menlo surveyed 550 U.S. healthcare and life sciences executives at director level and above who took part in procuring AI, so the findings reflect the views of people who actually buy these tools. Leaders should read the numbers as a map of where budgets are moving and where competitors are likely to be standing a year from now.

Strategic Takeaway

CIOs, Chief Medical Information Officers and Healthcare Strategy Leaders

Treat the scribe market as a lesson in platform risk. When an incumbent bundles a capability, a standalone vendor can lose share even while its revenue grows, so negotiate contracts with exit terms and data portability. Put the next wave of budget where Menlo says growth is fastest, in clinical copilots, patient engagement and drug discovery, and prepare for payers to answer provider AI coding with AI of their own.

02

Sierra Releases Poppy 1.0, a Protocol for Personal AI Agents, With Shopify Switching It On for Every Merchant and Meta's Muse Checking for It First

Sierra released Personal Agent Protocol 1.0, called Poppy, on November 13, 2026, in a post by Bret Taylor and Clay Bavor. Sierra described Poppy as a protocol that defines how personal agents connect with companies that implement it. The release matters because it tries to answer a question every consumer facing business will soon face: when a customer's own AI agent arrives at your website or API, how does it identify itself, what is it allowed to do, and who stays in control?

Sierra said Poppy is largely built on existing standards, namely MCP, OpenAPI and the Universal Commerce Protocol (UCP), and uses OAuth for authentication. Building on familiar pieces lowers the cost of adoption, because engineering teams that already run these standards are not starting from zero. Sierra said version 1.0 reflects about five weeks of design partner testing against real websites, APIs and agents. Sierra said any company can implement Poppy on its own.

The protocol is designed to be extended rather than rewritten. Sierra said Poppy extensions let industries build on the core protocol without changing it, and gave a commerce extension coauthored by Shopify as the example. That structure matters for governance, because banks, airlines and retailers have different risks and will want different rules layered on a shared base.

The partner roster shows how quickly the effort has grown. Sierra said the eight initial design partners were Meta, Genesys, Instinct, NiCE, Rocket, Shopify, Stripe and Walmart. Sierra said 60 new companies joined as design partners, bringing the total to more than 100. Sierra's list of new design partners includes Delta Air Lines, Intuit, Salesforce, Zoom, Cisco, Hyatt and Wayfair.

Distribution is what turns a specification into a standard, and Sierra reported movement at both ends. Sierra said Poppy is on for everyone who uses Meta's AI agent Muse, and that Muse automatically looks for Poppy first on every site it visits. Sierra said Poppy is enabled for every merchant on the Shopify platform, because every Shopify store supports UCP by default and now automatically publishes a Poppy file. Sierra said Shopify powers businesses in more than 175 countries. An agent that looks for the protocol and a platform that publishes it from the start give Poppy a real chance to be used rather than merely admired.

Two live examples show what customers can do today and where approvals remain. Sierra said Mercury is live on Poppy, where customers can ask a personal agent to check balances and review transactions, and a payment or transfer the agent starts goes through the account's approval policy first. Sierra said Turkish Airlines is live on Poppy, where agents can check fares, flight status and stopovers, Miles&Smiles members can have agents find business class upgrades that customers approve before paying, and completing a purchase inside the conversation is still in progress.

Notice the pattern in both examples: the agent can look, compare and prepare, but money moves only after the customer or the account's own policy says yes. That design choice is likely to decide whether risk and compliance teams accept agent initiated activity. It also shows that the hardest part of agent commerce is not the technology of calling an API but the discipline of deciding which actions an agent may take on its own.

Strategic Takeaway

Chief Digital Officers, Heads of Ecommerce and Customer Experience Leaders

Plan now for customers who arrive through their own agents. Ask your web and API teams what it would take to publish a Poppy file, decide which actions you would allow an agent to take without a human approval, and write those rules down before an agent asks. Companies that make themselves easy for trusted agents to use will be easier to buy from.

03

The Commerce Department's BIS Issues an Interim Final Rule on Remote Access to Advanced Chips, With a New License Requirement for Five Countries and an End to the AI Diffusion Rule

The Commerce Department's Bureau of Industry and Security (BIS) issued an interim final rule titled Addressing Remote Access to Advanced Semiconductors in November 2026. BIS said the rule was written by and at the direction of the Department of the Treasury and the National Security Council. BIS said the rule is meant to ensure existing restrictions on sensitive American technology cannot be circumvented through remote computing in data centers in third countries.

The rule works in two main ways. First, it creates a new license requirement tied to Indonesia, Malaysia, Singapore, Thailand and Brazil, which the rule describes as having a higher than normal risk of diversion. Second, according to the rule text, it requires a license for exports for data center use that would give companies headquartered in Macau or Country Group D:5 access to more than 12,000,000 aggregate TPP. BIS also gave existing data centers serving companies in Macau or Country Group D:5 90 days from the rule's effective date to obtain a license.

The rule contains exclusions that matter for ordinary business. It excludes transactions with a U.S. headquartered ultimate consignee or end user from the new license requirement for Indonesia, Malaysia, Singapore, Thailand and Brazil if the new reporting requirements are followed. It also excludes computing power used exclusively for a commercial recommender system. Companies that operate or rent capacity in the named countries will need to map their customers carefully to see which side of those lines they fall on.

The legal footing is unusual and worth noting. The rule cites Executive Order 14434, Inaugurating the Era of Super Intelligence, and section 1762(a) of the Export Control Reform Act of 2018, which allows the changes without prior notice and comment. BIS nonetheless asks for public comments no later than 60 days after publication in the Federal Register. Because the rule is interim, its details can still change, and affected companies have a chance to shape the final version.

The rule also closes a chapter. It formally rescinds the AI Diffusion Rule, published January 15, 2025, which BIS had announced it would rescind in May 2025. Reuters reported that the Biden era rule went unenforced from early 2025, which it said let billions of dollars' worth of restricted Nvidia chips flow to data centers used by Chinese AI companies.

The enforcement backdrop is political as well as technical. Reuters reported that Megaspeed, a Singapore based cloud company and Nvidia partner, has faced scrutiny over possible illegal diversion of chips to China. Reuters reported that Megaspeed said on November 15 that it complies with applicable laws, including U.S. export control requirements. Reuters reported that Treasury Secretary Scott Bessent said on Friday that four government agencies favored putting Megaspeed on a trade blacklist but the Commerce Department, which was in charge, did not.

For executives, the practical message is that where compute sits, and who rents it, is now a compliance question rather than only a cost question. A cloud contract with a provider in one of the named countries, or a model training arrangement that serves a company headquartered in a restricted group, can carry licensing consequences.

Strategic Takeaway

General Counsel, Chief Compliance Officers and Cloud Infrastructure Leaders

Inventory every contract where your AI workloads run on, or rent capacity from, data centers in Indonesia, Malaysia, Singapore, Thailand or Brazil, and identify any customer headquartered in Macau or Country Group D:5. Confirm whether the U.S. headquartered end user exclusion and its reporting steps apply to you, and consider filing comments within the 60 day window, since the rule is interim and its terms can still change.

04

Fortune Reports Capital Economics Says the AI Stock Boom Is in Its Final Stages, as Growth Runs at 3.4% While Hiring Slows and Labor's Share Falls to Its Lowest Since 1947

Fortune reported on November 16 that Capital Economics wrote on Monday that the AI stock boom is in its final stages. Fortune reported that Anthropic goes public on Thursday in a deal estimated to raise about $64.35 billion at the top of its range. Citing Renaissance Capital, Fortune reported that the raise is nearly twice what all U.S. IPOs raised together last quarter. The offering gives markets a live test of how much appetite remains for AI valuations, and it arrives at an unusual moment for the real economy.

That economy is growing without adding many workers. Fortune reported that the U.S. economy grew at a 3.4% annual rate in the third quarter of 2026, citing the Bureau of Economic Analysis. Fortune reported that employers added an average of 48,000 jobs a month over the 12 months through September, and that October added 165,000 jobs. Fortune reported that Goldman Sachs estimates AI spending drove more than half of business investment growth in the first three quarters of 2026.

Forecasters are trying to measure how long this pattern can last. Fortune reported that Morgan Stanley's 2027 outlook, summarizing chief U.S. economist Michael Gapen, puts the jobs breakeven at about 50,000 a month. Fortune reported that Gapen sees hyperscalers spending $1.6 trillion on capital expenditure in 2027, adding 0.9 points to growth. Fortune reported that Goldman estimates AI investment adds 0.6 percentage points to true GDP growth in 2027 and 0.2 points to counted GDP. Fortune reported that J.P. Morgan Asset Management's David Kelly sees real GDP growth slowing to 1.6% over the next four quarters.

Output per worker is rising while labor's slice of the pie is shrinking. Fortune reported that nonfarm business productivity rose at a 3.6% annual rate in the third quarter, per the Bureau of Labor Statistics. Fortune reported that labor's share of output fell to 53.0%, the lowest since the series began in 1947. Those two numbers together are the core of the jobless growth argument: more output, and a smaller portion of it flowing to wages.

Younger workers appear to feel it first. Fortune reported that the unemployment rate for 20 to 24 year olds was 7.5% in October, against 4.1% overall. Fortune reported that an August Stanford Digital Economy Lab study found employment of 22 to 25 year olds in the most AI exposed jobs was about 19% below where it would have been had it kept pace with peers in less exposed work. Fortune reported that an Apollo study by economists Sania Edlich and Torsten Slok found that after 2023 real wage growth in the most AI exposed occupations ran about 6.7 percentage points behind less exposed jobs, with no statistically significant effect on employment.

The Apollo findings add detail on who is affected. Fortune reported that the Apollo study found the lowest wage quartile of exposed jobs took the biggest hit, about 10.7 points, and that about 5.8 million workers are in the most exposed occupations. Studies like these are best read as signals rather than settled proof, since labor markets are still working through the aftermath of the pandemic.

Looking ahead, Fortune reported that Goldman expects the low hire, low fire job market to last about another year before AI displacement begins to have a larger impact. That timeline suggests a window for planning. The pressure on entry level roles and on lower paid workers in exposed occupations is already visible in the data, and the larger effect that Goldman anticipates has not yet arrived.

Strategic Takeaway

CHROs, CFOs and Workforce Strategy Leaders

Use the window Goldman describes to redesign entry level pathways and to retrain workers in the occupations the Apollo study flags as most exposed. Report productivity gains alongside workforce plans, because a rising output per worker and a falling labor share will draw scrutiny from employees, investors and regulators. Stress test budgets for a market in which public enthusiasm for AI valuations cools.


The Analysis

The Bottom Line

Taken together, these stories describe an AI economy that is maturing in three directions at once. Spending is concentrating in sectors where the work is high value and heavily regulated, standards are forming so that agents can act for customers without losing their guardrails, and governments are drawing lines around where advanced computing can run.

The common thread is accountability. Healthcare buyers are weighing platform risk, Poppy builds approvals into agent actions, the Commerce rule ties compute to licenses and reporting, and the jobs data invites hard questions about who benefits. Leaders who can show their work on all four fronts will be better placed than those who simply adopt.

Your next step is practical. Pick the one story that touches your business most directly, assign an owner, and set a date to report back with a specific decision.