SENS ADVISOR

Applied AI Notes · What Businesses Are Asking AI to Do · November 2026

What Moved Since October

Selected measures from the October edition, set against November.

Each month I review requirements that businesses publish online for AI systems that act through software tools. In this edition, about 200 projects published over roughly two weeks to mid-November 2026 are set against about 220 in the October edition, using identical sources, window and codebook. On most measures, November looks much like October. Where shares moved further, they are set out below, although two months of data collected in one place cannot establish a trend.

85%ask for an “agent” (October 87%)
41%of those describe a fixed workflow (October 49%)
46%state no success check (October 39%)
13%are flagged as possible agent cases (October 13%)

Several business areas changed share

E-commerce and retail operations now account for 8% of projects, compared to 2% a month earlier, while AI products that businesses are building to sell account for 12%, against 5%. Customer support dropped to 4% (down from 10%), matching a smaller share of chat and voice assistants (20%, against 27%), although that second difference does not reach unadjusted 5% significance. Trading and crypto projects, at 4%, remain too few to interpret with any confidence.

0%10%20%Sales and lead generation: October 16%, November 12%Sales and lead generation16% → 12%Internal operations: October 12%, November 10%Internal operations12% → 10%Marketing and content: October 11%, November 9%Marketing and content11% → 9%AI products built to sell: October 5%, November 12%AI products built to sell *5% → 12%Customer support: October 10%, November 4%Customer support *10% → 4%Real estate: October 5%, November 8%Real estate5% → 8%E-commerce and retail: October 2%, November 8%E-commerce and retail *2% → 8%Healthcare: October 4%, November 4%Healthcare4% → 4%Data and analytics: October 4%, November 3%Data and analytics4% → 3%Software and developer tools: October 4%, November 3%Software and developer tools4% → 3%Personal productivity: October 3%, November 3%Personal productivity3% → 3%Trading and crypto: October <1%, November 4%Trading and crypto *<1% → 4%
OctoberNovember* passes an unadjusted test at 5%
Share of projects, by business area; areas under 3% in both months are left out.

Autonomous operations accounted for a larger share

Among the tasks requested, autonomous operations, meaning systems intended to run part of a business with little direction, were the only task category whose difference reached unadjusted 5% significance, reaching 7% of projects compared to 2% the month before. Business workflow automation, multi-agent platforms and tool integrations through the Model Context Protocol (MCP) each stayed within a point or two of their October shares.

0%10%20%30%Chat or voice assistant: October 27%, November 20%Chat or voice assistant27% → 20%Business workflow automation: October 15%, November 16%Business workflow automation15% → 16%Multi-agent platform: October 10%, November 11%Multi-agent platform10% → 11%Tool integration (MCP): October 8%, November 9%Tool integration (MCP)8% → 9%Content pipeline: October 8%, November 7%Content pipeline8% → 7%Research or monitoring: October 4%, November 7%Research or monitoring4% → 7%Autonomous operations: October 2%, November 7%Autonomous operations *2% → 7%Outreach sequencing: October 4%, November 2%Outreach sequencing4% → 2%Knowledge Q&A: October 4%, November 3%Knowledge Q&A4% → 3%Document extraction: October 2%, November 4%Document extraction2% → 4%Testing, reliability, safety: October 4%, November 1%Testing, reliability, safety4% → 1%Coding agent or dev tool: October 2%, November 3%Coding agent or dev tool2% → 3%Other: October 9%, November 10%Other9% → 10%
OctoberNovember* passes an unadjusted test at 5%
Share of projects, by the task the AI system is to perform.

The calls for agents are no more agentic

Of the projects that ask for an “agent”, 18% describe a model choosing its next step, also 18% after rounding in October. Fixed workflows fell to 41% of these projects (previously 49%), while requirements too vague to tell rose to 36% (27%). A larger share was classified as too vague to tell, although that difference does not reach unadjusted 5% significance.

0%10%20%30%40%50%Fixed workflow: October 49%, November 41%Fixed workflow49% → 41%Model chooses next step: October 18%, November 18%Model chooses next step18% → 18%Plain code or one model call: October 6%, November 5%Plain code or one model call6% → 5%Too vague to tell: October 27%, November 36%Too vague to tell27% → 36%
OctoberNovember* passes an unadjusted test at 5%
Share of projects that ask for an “agent”, by the design their requirements describe.

Checks and delivery are close to October

Projects that state no success check rose to 46% (up from 39%), while those requesting testing or evaluation dropped to 46% (53%). Neither difference reaches unadjusted 5% significance, and more editions will be needed to assess whether either pattern persists. Human review, approval or handoff, mentioned in 44% of projects, is essentially unchanged.

0%10%20%30%40%50%Reference check: October 49%, November 45%Reference check49% → 45%Metric or model-based assessment: October 6%, November 4%Metric or model-based assessment6% → 4%Someone's judgment: October 6%, November 5%Someone's judgment6% → 5%No check stated: October 39%, November 46%No check stated39% → 46%
OctoberNovember* passes an unadjusted test at 5%
Share of projects, by the strongest success check stated.
0%20%40%60%80%Names the systems to connect: October 70%, November 75%Names the systems to connect70% → 75%Human review, approval or handoff: October 46%, November 44%Human review, approval or handoff46% → 44%Deployment, monitoring, maintenance: October 48%, November 54%Deployment, monitoring, maintenance48% → 54%Testing or evaluation: October 53%, November 46%Testing or evaluation53% → 46%Guardrails, privacy, compliance: October 46%, November 51%Guardrails, privacy, compliance46% → 51%A step that needs interpretation: October 38%, November 32%A step that needs interpretation38% → 32%Cost or speed limits: October 22%, November 16%Cost or speed limits22% → 16%A simpler option: October 13%, November 13%A simpler option13% → 13%
OctoberNovember* passes an unadjusted test at 5%
Share of projects that state each requirement.
0%20%40%60%Events or schedules: October 53%, November 50%Events or schedules53% → 50%A person hands over a job: October 23%, November 22%A person hands over a job23% → 22%Unclear: October 24%, November 28%Unclear24% → 28%
OctoberNovember* passes an unadjusted test at 5%
Share of projects, by what starts the work.

Claude is named more often

Claude is now named in the requirements of 39% of projects, compared to 26% a month earlier, as is Python (39%, against 26%). OpenAI held at 30%, while low-code automation platforms declined to 30% (previously 36%), a difference that does not reach unadjusted 5% significance. Because these figures count tools named in the requirements, they may not capture every tool in use.

0%10%20%30%40%n8n, Make or Zapier: October 36%, November 30%n8n, Make or Zapier36% → 30%OpenAI: October 30%, November 30%OpenAI30% → 30%Claude: October 26%, November 39%Claude *26% → 39%Python: October 26%, November 39%Python *26% → 39%MCP: October 18%, November 17%MCP18% → 17%Agent frameworks: October 12%, November 13%Agent frameworks12% → 13%
OctoberNovember* passes an unadjusted test at 5%
Share of projects whose requirements name each tool. Agent frameworks: LangChain, LangGraph, CrewAI, AutoGen, OpenAI Agents SDK or LlamaIndex.
0%10%20%30%Fixing or extending an existing system: October 20%, November 21%Fixing or extending an existing system20% → 21%Brief specification: October 26%, November 27%Brief specification26% → 27%Flagged as a possible agent case: October 13%, November 13%Flagged as a possible agent case13% → 13%
OctoberNovember* passes an unadjusted test at 5%
Share of projects.

What held steady

Most of what the October edition described still holds, with the figures below moving by less than three points.

For a full analysis behind each measure, the October edition remains the reference: Called an agent, specified as a workflow, How will anyone know it works?, Before you commission an agent.

All figures as a table
What the system is to doOctoberNovember
Chat or voice assistant27%20%
Business workflow automation15%16%
Multi-agent platform10%11%
Tool integration (MCP)8%9%
Content pipeline8%7%
Research or monitoring4%7%
Autonomous operations *2%7%
Outreach sequencing4%2%
Knowledge Q&A4%3%
Document extraction2%4%
Testing, reliability, safety4%1%
Coding agent or dev tool2%3%
Other9%10%
Business areaOctoberNovember
Sales and lead generation16%12%
Internal operations12%10%
Marketing and content11%9%
AI products built to sell *5%12%
Customer support *10%4%
Real estate5%8%
E-commerce and retail *2%8%
Healthcare4%4%
Data and analytics4%3%
Software and developer tools4%3%
Personal productivity3%3%
Trading and crypto *<1%4%
Design described, among projects asking for an “agent”OctoberNovember
Fixed workflow49%41%
Model chooses next step18%18%
Plain code or one model call6%5%
Too vague to tell27%36%
How success is to be checkedOctoberNovember
Reference check49%45%
Metric or model-based assessment6%4%
Someone's judgment6%5%
No check stated39%46%
What else the requirements specifyOctoberNovember
Names the systems to connect70%75%
Human review, approval or handoff46%44%
Deployment, monitoring, maintenance48%54%
Testing or evaluation53%46%
Guardrails, privacy, compliance46%51%
A step that needs interpretation38%32%
Cost or speed limits22%16%
A simpler option13%13%
What starts the workOctoberNovember
Events or schedules53%50%
A person hands over a job23%22%
Unclear24%28%
Tools namedOctoberNovember
n8n, Make or Zapier36%30%
OpenAI30%30%
Claude *26%39%
Python *26%39%
MCP18%17%
Agent frameworks12%13%
Kind of work and detailOctoberNovember
Fixing or extending an existing system20%21%
Brief specification26%27%
Flagged as a possible agent case13%13%
Method and limits

For this edition I collected requirements that businesses published online for AI projects over roughly two weeks to mid-November 2026. From these, I keep projects that ask for a system to act through software tools and that aim to build new systems or to fix or extend existing ones. A language model then classifies each set of requirements against a fixed codebook, built on the design process and task-screening worksheet I use in practice. The October figures are recomputed here with the same filters, from that edition's retained labels.

One limit is scope: because the requirements were published in one place over about two weeks, they describe these projects rather than the whole market. All shares refer to projects rather than businesses, since a single business can publish several projects. In a separate review of the success-check label in each edition, a second model labelled every project twice, and where its labels disagreed with the original, a third model settled the label under written rules, although agreement by itself cannot establish correctness. Unanimous labels were also checked against the same rules, with any proposed changes settled by the third model. Other labels were never independently re-coded, with no manual audit of any label. The classification read the full text of each set of requirements, so the result reflects what businesses state rather than the system eventually built. Changes describe the mix of requirements, not market growth.

Use of AI tools

A language model classified the requirements against my codebook and helped to draft the analysis code and this text. Other AI models re-checked the success-check labels and reviewed the method and claims. Having designed the framework and the codebook and checked the figures, I am responsible for the conclusions.