Garage Eight: How AI Is Changing Analysts' Work and Why Junior Roles Are Disappearing

Garage Eight analyst Vladimir Syropyatov has outlined six AI trends that are already changing day-to-day work with data. His main thesis is blunt: neural networks do not eliminate the profession, but they quickly devalue routine work and raise the entry bar.

Whom AI Will Replace

According to the author, junior analysts feel the pressure of AI the most. What not long ago counted as a normal starting job - ad hoc queries, simple dashboards, draft reports, basic SQL and Python debugging - is now increasingly done in minutes with a model. As a result, companies keep hiring analysts, but look less and less for people to execute routine tasks and more and more for specialists who can give a model the right context, check the result and tie it to a business decision. The author's formula is simple: one mid-level analyst with AI replaces several juniors at once.

• typical junior positions built around routine are disappearing;
• managers answer some analytical questions themselves through AI;
• vague wording gives way to clear prompts and specifications;
• training shifts from syntax to logic, validation and working with LLMs;
• demand is growing for hybrid roles at the intersection of analytics, domain knowledge and AI.

The author sees a separate effect in the redistribution of value. Cheap AI tools give small and mid-sized businesses access to capabilities that used to belong only to large teams with dedicated analytics resources. In this setup, the analyst is needed not as someone who simply "pulls the numbers" but as a solution architect: they explain the limits of the data, gather context, verify the model's output and make sure a fast machine answer does not turn into an expensive business mistake.

Garage Eight in Practice

To show this is not just theory, Garage Eight describes several internal cases. To work with sensitive data, the team is testing local models, including DeepSeek-R1 14B, so that managers and analysts can get SQL hints and investigate the causes of falling conversion without the risk of a leak. According to the author, in this scenario the speed of getting insights rose 32-fold, and employees came to understand the origin of the metrics better, because they began working with them directly rather than only through a queue of requests to analysts.

Three more uses look just as pragmatic: AI summarizes meetings, checks code before release and helps assemble drafts of reports and presentations. The article claims that AI validation speeds up bug detection 3-5 times and cuts the number of production bugs by 70%, while report preparation drops from roughly three hours to one. Inside the company this has already settled into a set of good-practice rules: run code through a local LLM and review presentations with AI before showing them to an audience.

"AI will not replace analysts.

But analysts who use AI will replace those who do not."

How to Prepare Now

The author draws two practical conclusions from these observations. For managers the advice is simple: do not wait for the perfect moment, break processes into parts and introduce AI step by step. At the same time, teach the team, work out agreements with security and compliance, and create an environment where a hypothesis can be tested quickly instead of spending weeks approving a pilot. Garage Eight is already looking at automatic task prioritization, personal AI assistants, AI mentors for newcomers and regular searches for non-obvious hypotheses in the data.

The plan for analysts is also clear. They need to master the tools now, but not at the level of "press a button - get an answer"; rather at the level of critical verification, product thinking and communication. The more routine the model takes on, the more valuable the skills become of explaining complex things in simple words, seeing the process behind the numbers and knowing where AI can be trusted and where its output must be rigorously rechecked. Otherwise a specialist risks competing not with a colleague but with a cheap and very fast automated layer.

What This Means

The Garage Eight article captures a shift already visible in many data teams: AI does not abolish analytics, it moves it up the value chain. The winners will be specialists who can combine the model's speed with their own understanding of the business, the data and the risks. The losers will be those who keep selling the market only manual routine, which neural networks have already learned to do faster, cheaper and almost without a queue.