Digital Transformation with Google Cloud practice questions

In cloud computing, infrastructure means the underlying compute, storage and networking resources, physical or virtual, that applications and data run on.

When a company stops running its own infrastructure, a provider now operates the servers, storage and networks. The company's applications and data still exist and still belong to it; they simply run on the provider's infrastructure. Contracts and policies are governance, not infrastructure.

1. Picture the layers: infrastructure at the bottom, applications and data on top.
2. Match the word to the bottom layer: compute, storage, networking.
3. Eliminate applications, data and governance documents, which sit on or around that layer.

Exam tip: Infrastructure is the compute, storage and networking layer everything else runs on.

Agentic AI refers to AI systems that work toward a goal by reasoning, planning a series of steps and using tools or other systems to act, with limited human direction.

The bank's system must decide what to do next, call several systems (transaction lookup, case management, email) and finish the job. That combination of planning and acting is agentic AI. Predictive models score or forecast, generative models produce content, and rule-based bots follow a fixed script.

1. Note what the system must do: decide steps and act in other systems.
2. Ask which kind of AI plans and takes actions toward a goal.
3. Eliminate options that only predict, only generate content, or only follow fixed rules.

Exam tip: Agentic AI plans and acts toward a goal; generative AI produces content when prompted.

Agentic AI builds on generative models but adds goal pursuit: the system plans steps, uses tools and systems, and adapts as it goes.

A generative model responds to a prompt with content. An agent uses a model to reason about a goal, decides which steps to take, calls tools such as APIs or databases, and continues until the task is done. A longer conversation history still leaves the system answering rather than acting; forecasting describes predictive ML; fixed rules describe classic automation.

1. Identify what a prompt-answering model does: it replies with content.
2. Find what only agents add: planning steps and taking actions with tools.
3. Reject descriptions of a better chat model, predictive ML and rule-based automation.

Exam tip: Generative AI creates; agentic AI plans and acts.

Generative AI produces content; agentic AI uses models plus tools to complete multi-step tasks across systems.

Two facts decide this. First, the gap is action, not wording: better drafts or forecasts leave staff making every change. Second, the requests vary widely. A fixed rules engine does act in the booking systems, so it is the tempting answer, but it only handles the change it was programmed for. An agent can plan the steps for each different request and carry them out across the three systems through tools.

1. Locate the gap: replies are fine, but nobody automatically changes the bookings.
2. Rule out options that improve content or forecasts without acting.
3. Compare the two options that act: a fixed rule versus an agent that plans each request.
4. Apply the constraint: many kinds of change means the steps must be worked out per request.

Exam tip: Fixed rules automate one known path; agentic AI plans and acts on varied goals.

Digital transformation is using modern digital technology to create or change business processes, culture and customer experiences to meet changing business and market needs.

The insurer changed where its systems run but not what the business does. Transformation is measured by changes to operations and customer experience, for example faster claims or new digital products. Digitizing records is a narrower step, cost-driven rehosting can leave everything else unchanged, and nothing requires replacing every system before transformation begins.

1. Note what changed: only the location of the servers.
2. Recall what digital transformation changes: processes, culture, customer experiences.
3. Separate it from digitization and from simple rehosting.

Exam tip: Moving servers is migration; changing how the business works is transformation.

Open source software is software whose code is made available under a license that lets anyone use, inspect, modify and share it.

The deciding needs are reading, changing and sharing the code. Only open source licensing grants all three. Freeware is about price, not code access; proprietary software keeps code private; and software built to an open standard may itself be closed source.

1. List the needs: read the code, change it, share it.
2. Ask which licensing model grants all three.
3. Separate price (freeware) and specifications (open standards) from code access.

Exam tip: Open source is about access to the code; free of charge is not the same thing.