AI NewsAI AgentsWeekly RoundupNews

The Week in Agents

Menlo’s $177B, the agentic web, and 29 more finds — sorted into what an agent is, what it is not, and where the money actually goes.

A night-time network of connected dashboards and drones, representing the week's agent news

The week’s agent news, sorted. Menlo’s numbers are the headline: deployment, tokens and growth all point the same way — spend follows verifiable work.

Under that: the definitions, the assistant-versus-agent arguments, the enterprise money, the critics and the builders. Thirty links, one rule.

The short version

  1. The definitions converged.Perceive, reason, plan, use tools, act toward a goal. The argument is about thresholds, not words.
  2. The market separated.Assistants await, automation follows rules, agents take initiative — and spend only follows where results verify.
  3. The web grew a second audience.Agents read, click, and buy. Design for both, and decide where humans stay.

Definitions

6 finds
02Definition

IBM — What Are AI Agents?

IBM defines agents as systems that autonomously perform tasks for a user or system.

Why it mattersThe enterprise baseline: autonomy on someone's behalf, not just answering.

ibm.com →

Friction

7 finds
08Assistant vs agent

IBM — Agents vs Assistants

Assistants await you, agents take initiative on specific tasks.

Why it mattersThe week's core confusion: conversation is not delegation.

ibm.com →
12Assistant vs agent

GWI — Difference

Agents framed as fully autonomous with little intervention.

Why it mattersThe maximalist claim to check against evals.

gwi.com →
Featured · Menlo Ventures, Nov 3

Menlo Ventures 2026: State of Enterprise AI (Nov 3)

Menlo finds deployment hit $177B by September (was $33B, 5.4x) with $1.70 back per $1 building and 55% net-new budgets. Agents now drive 80% of OpenRouter tokens and 37x growth since February, so spend scales with work — and coding ($34B) plus chatbots ($25B) take 76% of apps because results verify.

$177BEnterprise agent deployment by September, from $33B a year earlier
80%Of tokens on a major model router now come from agents
37xAgent token growth since February
$1.70Returned for every $1 spent building models

Why it mattersIt replaces definition debates with a verification rule for the whole week.

menlovc.com — $177B deployment, agents 80% of tokens →

Enterprise + analysts

6 finds

Critics + community

6 finds

Builders + agentic web

4 finds

The rule of the week

Follow the verification.

Autonomy is a claim; a checkable result is the evidence. The feature turns the 30 finds into a test you can run on any demo.

Read the feature →

All 30 items link to their original source. Featured data: Menlo Ventures, 2026 State of Enterprise AI (Nov 3). Roundup assembled from the Nov 2 content week; definitions, critics and builder finds are quoted and linked, not paraphrased as fact.

Get the next field note

New articles, show notes, and practical lessons from the Strattegys AI crew.

Keep reading
More field notes from the stack.
Privateering in the AI Caribbean: Riding the Waves of the Fifth FrontierFeatured

Privateering in the AI Caribbean: Riding the Waves of the Fifth Frontier

The fifth frontier: PC, internet, cloud, SaaS, AI all consolidated the same way. Finding the privateer’s viable lane between the great powers.

See How AI Changes the Calculus for GTM StrategyInsights

See How AI Changes the Calculus for GTM Strategy

What we learned from a Mirror Teams strategy session about how AI lowers the barriers, increases the speed, and changes the output of go-to-market analysis.

Should AI do this? A five-question checklist across four lanes: build, buy, delegate, keepLearn

Should AI Do This or Should I?

Should AI do this, or should you? A five-question checklist, a build / buy / delegate / keep scorecard, and a five-task exercise.

SORINT.us logo

Platform Services

Design, build, staff, and run the platforms your teams depend on.

Visit site →

Your brand here

Get your founding sponsor spot now.