top of page

AI Agent Strategy: The Week Agent Supply Outran Agent Demand

Writer: Ling Zhang
Ling Zhang
4 days ago
8 min read
Everyone Shipped AI Agents This Week. Almost Nobody Used Them.

November 9 – November 15, 2026: When Agent Supply Outruns Agent Demand – AiFullest Weekly


Guiding Question: What changed this week, and what should Data & AI leaders do about it?


This week told two stories about AI agents at the same time. The supply of agents, agent platforms and agent tooling grew faster than ever. Meanwhile, some of the best-known consumer companies in the world told investors that almost nobody is using AI agents to transact with them. If you are setting your AI agent strategy for 2027, the gap between those two stories is the most useful thing you can study right now.


AI Agent Strategy: The Week Agent Supply Outran Agent Demand


Lead Signal: Agent Supply Is Booming. Agent Usage Is Not.

On November 10, Fortune reported what Uber, DoorDash and Airbnb said on their recent earnings calls about outside AI agents transacting on their platforms. According to Fortune, Uber said AI agents booked four rides and 500 Uber Eats deliveries out of 320 million combined rides and deliveries during the week of Halloween. DoorDash CEO Tony Xu described agent-driven orders as “maybe several hundred orders in total” since launch. Airbnb CEO Brian Chesky said a “very non-material number” of its nearly half a million daily bookings came through AI agents.


The same morning, IBM and Microsoft announced Imagine Enterprise, which they describe as “an agentic platform designed to help enterprises build, run and govern AI agents.” IBM Consulting brings more than 6,000 prebuilt AI agents and assets to it, the industry frameworks start with financial services, insurance and retail, and forward-deployed engineering support is offered for qualifying engagements.


And the week kept going. On November 13, LangChain announced a $300 million Series C at a $3 billion valuation, saying more than 7,000 organizations rely on its LangSmith platform. On November 12, Kindo announced a $30 million raise to scale AI agent management, and Nebuly announced a $21 million Series A to help enterprises measure the value of their AI. Google released Gemini 4 Hydrogen, which it calls its “daily driver for agentic workflows,” and Databricks laid out three ways enterprises can put OpenAI models to work on governed Databricks data.


Notes: the Uber, DoorDash and Airbnb figures are Fortune’s reporting of earnings-call remarks, and I have not read the call transcripts myself. The Uber number covers a single week, not a quarter.


Why It Matters

Look at where the money and the launches went this week: platforms to build agents, tools to manage agents, and tools to measure whether agents create any value. That last category is the tell. When investors fund measurement, it is usually because buyers cannot yet prove the value on their own.


The consumer numbers are not a verdict on agentic AI. As Fortune reports it, Chesky drew a useful line: agents are good at simple, routine tasks, but planning a trip is different. Ordering an Uber already takes under a minute. When the human path is already fast, or even enjoyable, an agent has little to add. The same logic applies inside your company.



The AI Harness Layer: Why Scaling AI Was Never the Hard Part — Proving AI ROI Is

The workforce side is lagging too. In CIO Dive’s coverage of Gartner’s 2027 CIO agenda survey of 2,002 CIOs and technology executives, just 4% of CIOs reported plans in place to revamp roles with a human-agent workforce in mind. Agent supply is accelerating. Role design is not.


Read more: if the hard part of AI was never scaling but proving value, this is the piece I would revisit before your next agent investment.



My Perspective: Your AI Agent Strategy Should Start With Demand

I do not read this week as a warning against agents. I read it as a reminder that availability is not adoption. Most agent conversations I have with leaders start with a platform, a vendor or a count of use cases. Very few start with the question that matters most: which work, done by which people, is slow, painful or error-prone enough that they will actually hand it over?


A sound AI agent strategy starts with demand, names an owner for every agent, and measures usage before it measures anything else. An agent nobody uses is not a pilot. It is a cost with a dashboard.


AI Agent Accountability: Who Owns the Agent When It Makes a Mistake?

Read more: an agent with no owner is a risk whether it is used or not. This is how I think about accountability when an agent gets it wrong.




What Mid-Level Leaders Should Do
  1. Count real usage. For every agent your team runs, find out how many people used it last week and for what. If you cannot get that number, that is your first project.


  2. Pick work people want to give away. Choose tasks your team finds slow or tedious, not tasks that are already quick. Ask them rather than assume.


  3. Name one owner per agent. One person who answers for its usage, its errors and the date it gets reviewed or retired.



What Executive Leaders Should Do
  1. Ask for adoption before expansion. Before approving the next agent platform or bundle, ask for usage and outcome data on the agents you already run.


  2. Fund measurement inside the program. Budget for usage tracking and value measurement from day one instead of adding it after the money is spent.


  3. Put role redesign on the same timeline. If Gartner’s 4% describes your organization, your agent roadmap is ahead of your workforce plan. Close that gap on purpose.



Practical Exercise: The Agent Demand Test

Pick one agent your organization runs or plans to buy, and answer three questions honestly:


  • Who asked for it? Name the people whose work it changes, and whether they requested it.


  • What does it replace? Describe the current human path. How long does it take today, and how do people feel about doing it?


  • What would make you retire it? Write down the usage level below which you would switch it off.



If you cannot answer the first question with names, you have a supply-driven agent. That is not a failure, but it is a signal to go back to the people before you go back to the vendor.


If AI Development Itself Is Accelerating, What Should You Be Learning Right Now? A Case for Durable AI Judgment Over Tool Fluency

Read more: the skill that keeps its value as tools multiply is judgment about where AI belongs. I made that case earlier this week.





Supporting Signals

NVIDIA Announces a CEO Transition

On November 12, NVIDIA announced that Jensen Huang has resigned as President, Chief Executive Officer and a member of the board, effective immediately. CFO Colette Kress is serving as Interim CEO while the board searches for a permanent successor, and Harvey C. Jones becomes the company’s first-ever Chair of the Board. The release gives no reason for the resignation and says the board expects operations, customer commitments and the product roadmap to continue without interruption. NVIDIA reports third-quarter results on November 17 as previously scheduled.


For Data & AI leaders, the lesson is not about one person. It is about concentration. If your AI roadmap depends heavily on a single infrastructure supplier, this is a good week to ask what your continuity plan looks like, the same way you would for any critical vendor.


Advocacy Groups Ask a Court to Freeze Next-Generation AI Models

Al Jazeera reported that on November 12, four nonprofits (Irreplaceable Education Fund, Public Citizen, Heat Initiative and Alliance for Secure AI) filed a civil complaint in San Francisco Superior Court against Anthropic, Google, Meta, OpenAI and SpaceX. They argue the companies violated California’s unfair competition law and ask for an emergency injunction on next-generation models until the technology can be shown to be under human control. I have not read the complaint itself, so please treat the details as reported.


Whatever the outcome, governance pressure is now arriving through the courts as well as through legislatures. It is one more reason to write down how your own organization decides which AI it deploys, and why.


AI Tools & Releases Worth Knowing

What it is: New versions of Anthropic’s models. Anthropic says they generate outputs up to 30% faster than the 5.1 versions, at $8 per million input tokens and $40 per million output tokens, with an optional fast mode at $16 and $80.


Who it is for: Developers and enterprise teams building on Claude. Access to Mythos 5.2 is more restricted, so check Anthropic’s page for eligibility.


Why leaders care: Faster, cheaper output changes the running cost of every agent you operate. Re-run your cost model before your 2027 budget locks.



What it is: Google’s fast model for agentic work, with four thinking levels, priced flat across its 1 million-token context window at $1.50 per million input tokens and $7.50 per million output tokens.


Who it is for: Developers through the Gemini API and AI Studio, enterprises through Gemini for business and Antigravity, and consumers in the Gemini app and Search.


Why leaders care: Cheap, fast models make it easy to build many agents. That makes demand discipline more important, not less.



What it is: Google’s strongest model, opened to paying developers and enterprises after an initial release to cyber defenders. Introductory pricing is $2 and $10 per million input and output tokens until November 30, then $4 and $20.


Who it is for: Developers and enterprises through the API and Gemini for business, and Google AI Ultra subscribers.


Why leaders care: If your team is piloting it, budget for the post-introductory price now, so the pilot’s business case survives December.



What it is: Claude joins a Google Meet call as a visible participant, takes notes, answers questions and builds documents, whiteboards and slides during the call.


Who it is for: Claude Team and Enterprise plans.


Why leaders care: An agent in the meeting raises consent, recording and record-keeping questions. Settle your policy before usage spreads on its own.



Signals to Watch
  • NVIDIA’s November 17 results. The first quarterly report under an interim CEO. Why it matters: listen for any change in supply commentary that touches your AI infrastructure plans.


  • Anthropic’s IPO pricing. Anthropic launched its roadshow on November 10, and RTTNews reported a range of $70 to $78 per share for a Nasdaq listing under “ANT.” Reported pricing dates are unconfirmed. Why it matters: frontier-lab revenue and risk will soon be priced by public markets every quarter.


  • Gemini 4 Carbon’s introductory pricing ends November 30. Why it matters: a pilot’s cost can double overnight if no one is watching the meter.


  • Agent management and measurement as a category. Kindo, Nebuly and LangChain all raised money this week to manage or measure agents. Why it matters: expect your vendors to start offering usage and value dashboards, so ask for them now.


  • EU AI Act enforcement procedures. The Commission’s feedback period on draft enforcement procedures closed on November 9, per regulations.ai; the drafts are not yet adopted. Why it matters: procedural rules decide how an investigation would actually run.




This week’s lesson is simple. The world is not short of agents. It is short of agents people actually want to use. The leaders who do well in the year ahead will be the ones who start their AI agent strategy with the people doing the work, not with the catalog of what is possible.


I often remind the leaders I coach that adoption is earned, not installed. The same is true for every agent you deploy.


Closing Reflection

If you switched off every AI agent in your organization tomorrow, which ones would people actually ask to have back?



Ready to Grow Into Your Next Level?

If your organization is investing in AI but struggling to turn activity into measurable business value, my AI & Data Strategy Consulting Framework can help you build a clear, governed, value-driven roadmap.


If you are a Data & AI leader who wants to grow from technical contribution to strategic influence, my Data & AI Leadership Winning Blueprint can help you strengthen your executive presence, communication, and transformation leadership.


If you are navigating career growth, personal growth, or leadership reinvention in the AI era, my coaching programs can help you clarify your next chapter and grow with confidence.


And if you want to approach this season with greater financial clarity, my financial education and holistic check-in conversation can help you review protection, risk, growth, and tax through the lens of the life you are building.


You do not have to navigate this season alone. Book a complimentary strategy conversation and take your next step toward leading, growing, and building with clarity, confidence, and purpose.


May you grow to your fullest in your data science & AI!

May you grow to your fullest in your data science & AI!

Subscribe Grow to Your Fullest and

Comments


bottom of page