What I learned from the Google AI Professional Certificate, and where it fell short

What I learned from the Google AI Professional Certificate, and where it fell short

I’ve just completed the Google AI Professional Certificate. The sections on research and data analysis were the strongest parts for me. They offered useful ways to think about bringing AI into everyday work. The app development and launch sections were less convincing: very basic, with too much emphasis on using Gemini tools and not enough depth on what happens when you need something dependable.

That makes this a mixed review, but a positive one in the areas I’m most likely to apply to my workflows.

Google describes the programme as practical training in using AI at work, covering areas including planning, research, communication, content creation, data analysis and app building. That workplace focus is where I think it earns its place. My expectations were higher when it moved towards building and launching applications.

Research was one of the most useful parts

Research is an obvious place to use AI, but it is also an easy place to mistake a convincing answer for a reliable one.

The research section was valuable because it connected AI to a recognisable piece of work: gathering information, making sense of it and using it to inform a decision. Google’s course description includes using Deep Research and a source-based research workspace to summarise material and pressure-test ideas.

For my own workflows, the useful application is a more structured research process. Before asking for a report, I can define the decision I’m trying to make, the evidence I need and the questions that remain open. AI can help organise the material and compare different perspectives. I still need to check whether the sources support the conclusions.

For example, if I’m assessing an AI tool for a Valdris client, a useful research brief would compare its capabilities, pricing, integrations and data-handling terms against that client’s requirements. It should flag anything it cannot establish. A polished list of features would not be enough.

I want to carry that approach into researching business opportunities and preparing consulting recommendations. It gives the research a clear purpose and makes the result easier to review.

Data analysis felt directly applicable

The data analysis section was the other standout for me. Google’s description covers cleaning data, analysing it and producing visualisations and formulas with Gemini in Google Sheets. These are practical tasks that businesses already need to do.

The appeal is being able to explore a question in plain language, then inspect the calculations and assumptions behind the answer. That could make it easier to work through a spreadsheet, investigate an unexpected result or decide which comparison would be useful.

One application I can see for Valdris is reviewing enquiries. Given an appropriately handled dataset, AI could help group enquiries by service, explore where they came from and suggest ways to visualise the results. I would then check the categories, totals and missing information before making decisions about where to focus.

The checks matter. An enquiry is not a sale. An empty field is not necessarily a zero. A chart can look perfectly reasonable while answering the wrong question.

My takeaway is to use AI to help investigate the data while keeping the underlying records and calculations available for review. I also need to check what information is appropriate to upload and which account settings and terms apply, especially when client data is involved.

App development and launch needed more depth

This was where the course fell short for me.

I found the app development and launch material too basic. It leaned heavily on using Gemini tools, and I wanted more depth on the work involved in turning an initial build into something people can rely on.

There is value in helping someone build their first small tool. Seeing an idea become interactive can make app development feel more approachable. For someone starting from scratch, that may be enough to make the section worthwhile.

But my expectations extend beyond getting an app to run. I want to understand how it handles real data, who can access it, what happens when an integration fails and how changes are tested before users receive them. Deployment also brings questions about costs, monitoring, maintenance and support.

I would have liked more attention to those concerns. My criticism is about the level of depth I needed, rather than a claim that an introductory course should turn someone into a software engineer.

For anyone planning to build a client-facing application, I would treat this part of the certificate as an introduction and plan for further learning and technical review before launch.

How I’ll apply it to my workflows

The most useful next step is to pick a repeatable task and make the AI-assisted version clear enough to evaluate.

For a research workflow, that means starting with a defined question, collecting relevant sources, asking AI for a structured comparison and reviewing the evidence before turning it into a recommendation.

For an analysis workflow, it means defining the question, checking the data, using AI to help explore it and validating the calculations before sharing the findings.

In both cases, I want to keep reusable briefs and review criteria rather than start with a blank prompt each time. I also want to assess the whole task: how much checking it requires, whether the output is useful and whether it reduces rework. Generating something quickly only helps if the result is worth using.

The course’s Google focus makes sense for a Google certificate. Still, I want the working methods to transfer beyond Gemini. The tool should fit the task, the data and the systems already in use.

I’m glad I completed the certificate. I’d recommend considering it for structured, practical exposure to AI at work, particularly research and data analysis. Experienced builders should set their expectations lower for the app development and launch material.

My next step is to apply the research and analysis approaches to a small set of repeatable workflows at Valdris, keep the review steps visible and judge the results on the quality of the work they help produce.

---

Course reference: Google AI Professional Certificate. The assessments above reflect my experience of completing the course; the workflow examples describe how I intend to apply it, rather than measured results already achieved.