Agentic AI should be easier to use.

Today, using agentic AI often means learning several tools, connecting accounts and building an elaborate setup before the real work can begin.

A creative professional overwhelmed by several devices and a cluttered work desk

The same creative professional calmly working with paper sketches, with one phone on the desk

Core Team3 agents working on 5 tasks

Meeting scheduled at 5:00 PM

Set up the technologyYou may need to choose tools, connect accounts and work out which settings to turn on before you can ask for help.Core Team starts with an agent that already knows the job.
Repeat the same contextEvery new chat can mean explaining the same people, plans and limits again.Bills stay with Dara and home jobs stay with Home HQ.
Track work across chatsResearch, choices and follow-ups can end up in different places.Work, handoffs and the next decision stay together.

People ask different people for different jobs.

One person choosing between money, repair, language and travel help — Use AI the way you already work with people: choose the right help, explain the outcome and stay involved where it matters.

What this could help with.

Core Team is an idea for everyday agentic AI. You start with the job, not the tool or the prompt.

Mei preparing catered food in her kitchen

Local work

A team that works the way Mei works.

Mei says, “Next week is busy—help me get ahead.” One agent reviews margins, one manages bookings, one protects preparation time and one prepares marketing. The team gives Mei one plan without asking her to change how she runs the business.

Fatima organising project notes in an office

Work that improves

Better work, with less to explain.

With permission, agents could learn a person’s useful formats, regular deadlines and recurring constraints. Privacy-protected patterns from similar tasks could improve the quality of the work—not just the tone of a reply.

Mehul checking stock in his grocery shop

Local knowledge

Agents that understand the local market.

Commercial agents could use a shared, permissioned knowledge base built from similar businesses and local signals—such as supplier reliability and seasonal demand—while each business’s customers, money and private work remain separate.

Hi, I’m Chinmay Phatak.

Core Team is currently a design concept in the problem-validation stage. I created this demo to test whether people want a simpler way to work with agentic AI—and whether agents organised around familiar jobs make sense.

The product does not exist in its final form yet. This demo is a way to question the problem, test the interface and learn what should be built next.

My intention is to keep developing the idea towards a real product.

See the demo Contact me