AI toolsTrendinghigh impact

OpenAI Signals Data Maps ChatGPT Use Across 165 Work Activities in 100 Countries

By InnovaAI Research2 min readLatent

OpenAI released its Signals dataset on October 26, 2026, covering 165 work activities across nearly 100 countries to show how people use ChatGPT for professional tasks. At the same time, Supabase co-founders outlined infrastructure changes needed to support AI-built applications at scale, signaling a broader shift in how agencies should think about AI tooling.

Key Facts

01OpenAI released its Signals dataset on October 26, 2026, documenting ChatGPT usage across 165 work activities in nearly 100 countries.
02Marketing content development and business data analysis are explicitly listed among the tracked work activities.
03Supabase co-founders outlined the need for agent-friendly backends capable of handling long-lived workloads as AI agents build more software.
04Agencies can download the full OpenAI Signals dataset to map AI use against their own service lines.
05The convergence of AI labor data and AI infrastructure design signals a structural shift, not just a tooling upgrade.

Why does this matter for agencies?

▶Having 165 work activities mapped across nearly 100 countries gives agencies concrete data to build client-facing AI disclosure policies instead of relying on internal estimates.
▶The explicit inclusion of marketing content development in OpenAI's dataset means clients in that space will increasingly expect transparency about how AI is used in deliverables.
▶Supabase's infrastructure discussion highlights that automation stacks not designed for AI agents may fail to scale as client operations grow more AI-dependent.
▶Agencies that audit their tooling now against agent-readiness criteria will avoid costly migrations later when client demands accelerate.

What should agencies do?

Download the OpenAI Signals dataset and map the 165 documented work activities to your agency's current service lines to identify where AI is already embedded and where disclosure policies are needed.

medium effort

Draft or update an agency AI-use policy that references the OpenAI Signals research as supporting evidence, specifying which tasks are AI-assisted and what human review steps apply.

medium effort

Review your current workflow automation stack for agent-readiness by testing whether tools like Make or n8n can handle parallel, long-lived, or agent-triggered workloads without manual intervention.

high effort