by Tan Aik Keong (AK)
If you've been hearing the term "Agentic AI" lately, along with the technology behind it — the Model Context Protocol, or MCP — don't let it feel unfamiliar. This technology, which sounds fairly technical, is steadily working its way into daily life, and it's going to change how you and I live.
What is MCP? From Anthropic to OpenAI
To explain MCP properly, it helps to start with two companies that have risen fast in the past few years.
First, Anthropic — an American AI research company founded in 2021 by former OpenAI employees. They built MCP with the goal of letting AI interact with humans more safely and effectively. In simple terms, MCP is a technology that lets AI remember and understand the context of a conversation — effectively giving it a memory module, so it responds more intelligently and more like a person actually following along.
OpenAI later recognised MCP's potential too, gradually adopting and supporting the technology's development, making AI models "smarter" and more logically consistent, more context-aware, and more natural to use.
The magic of MCP: AI that actually understands you
You've probably run into this before: chatting with a chatbot or customer-service bot that seems to remember only your last sentence, then forgets everything by the next one. It's a bit like talking to a forgetful friend — genuinely annoying.
MCP is built to fix exactly that. It gives AI a form of short-term memory, letting it hold onto conversational context, what you've said before, and what outcome you're actually after. That means a future AI assistant could act more like a real personal assistant or secretary — proactively managing your schedule, remembering your favourite restaurants or film genres, even automatically reminding you about bill payments or restocking everyday items.
Agentic AI: can AI actually make decisions on its own?
That brings us to the other term: Agentic AI. This kind of AI has the ability to make its own decisions and proactively complete tasks.
Say you used to tell an AI: "book me a restaurant table for 8pm tomorrow." A traditional AI could search or recommend options, but you'd still have to make the actual booking yourself.
With Agentic AI and MCP working together, AI doesn't just find a suitable restaurant — it can call the restaurant directly, or use its booking system, and complete the reservation itself. If the restaurant is fully booked, it can automatically find alternatives and check back with you, without you having to lift a finger.
What this could look like in daily life
MCP combined with Agentic AI is opening up a wave of "smart living" scenarios that make life genuinely easier:
Personal finance: AI proactively manages your bank accounts, automatically flags unusual charges, and warns you of potential fraud.
Healthcare: An AI assistant monitors your health in real time, and if something looks off — an abnormal heart rate or blood sugar reading — prompts you to see a doctor or adjust your habits.
Driving and mobility: An AI-equipped car automatically schedules servicing and repairs, predicting faults before they happen to keep your trips safe.
Logistics and delivery: AI automatically optimises your parcel delivery routes, and can adjust delivery timing in real time, so you always know exactly when a package will arrive.
Closing: AI's next stop is your everyday life
The MCP technology Anthropic pioneered, now actively pushed forward by OpenAI too, is moving AI from a state of "only follows instructions" toward one of proactive service and decision-making. Before long, we may all be used to living alongside an AI assistant that genuinely understands us.
This particular shift probably won't turn the world upside down overnight — but it will make daily life more comfortable, and a little more human.
Next time you hear "Agentic AI" or "MCP," remember: these aren't distant science fiction. They're steadily becoming part of your everyday life.
Part of the AK AI Corner column. Originally published in Oriental Daily (东方日报) on Apr 5, 2025.