McKinsey now tests job candidates on AI tools in final round interviews. Job postings requiring AI skills have nearly doubled in two years. Here is exactly what is changing and what you should do about it.
TL;DR
AI skills are becoming a formal job requirement, not just a nice-to-have. McKinsey now tests final-round candidates on its internal AI tool. Job postings requiring AI skills have risen from 5% to 9% in just two years, and professionals with AI fluency earn 56% more than peers in the same role. The skills employers want are not technical coding skills. They are practical AI fluency: structured prompting, output verification, and workflow integration. This article walks through exactly what you need to build and how to start this week.
Let’s be honest about what is going on. For the past few years, having AI skills was a way to stand out. It made you look forward-thinking. Now it is becoming something else entirely: a baseline expectation.
When McKinsey, one of the most selective employers in the world, starts testing job candidates on AI tools in final round interviews, that is not a signal. That is a statement. And McKinsey is not alone. The broader hiring data tells the same story.
Job postings explicitly requiring AI skills rose from just over 5% of all postings in 2024 to more than 9% in 2025. [1] That is nearly double in twelve months. Unique postings calling for generative AI skills specifically grew from 55 in January 2021 to nearly 10,000 by mid-2025. [1] The trajectory is not slowing down.
And this is not just about technical roles. The growth is happening across marketing, operations, HR, finance, consulting, and general management. AI skills are being treated the way Excel fluency was treated fifteen years ago: first a competitive advantage, then a baseline requirement, then something that raises questions if you do not have it.
In January 2026, McKinsey introduced an AI-enabled exercise into select final-round interviews for graduate roles in the United States. [2] Candidates are asked to work with Lilli, McKinsey’s internal AI tool, to complete applied problem-solving tasks that closely mirror real client work.
This is not a test of technical knowledge. You are not asked to explain how large language models work or write code. You are asked to demonstrate reasoning and judgment while working alongside an AI tool: how well you prompt it, how critically you evaluate its output, and whether you can synthesize AI-generated insights into a coherent, well-structured answer. [3]
What McKinsey is actually measuring: The ability to guide AI effectively, evaluate its output critically, and own the final analysis. That is exactly the skill profile most non-technical professionals need to develop, not just for McKinsey interviews, but for their current jobs.
The same CEO announcement revealed something else worth sitting with: McKinsey now runs 20,000 AI agents alongside its 40,000 human employees. [4] That ratio, roughly one AI agent for every two humans, is where the firm is operating today. The implication for anyone building a career in professional services, management, or consulting is clear. The expectation of your ability to work alongside AI tools is no longer hypothetical. It is baked into the operating model.
McKinsey’s move matters not because it is unique, but because firms like McKinsey set precedents. Where they lead on talent strategy, others follow. Within two to three hiring cycles, expect similar assessments to appear across consulting, finance, and corporate graduate schemes more broadly.
The McKinsey story is a useful hook, but it is just one data point in a much larger trend. Here is what the broader landscape looks like.
The wage premium for AI skills is substantial and growing. Professionals with advanced AI fluency earn 56% more on average than peers in the same roles without those skills. [1] That premium is highest in finance, consulting, marketing, and legal roles, which are precisely the roles where AI tools are creating the most leverage.
Demand for formal degrees is also declining in AI-adjacent roles. The percentage of AI-augmented jobs requiring a degree dropped from 66% in 2019 to 59% in 2024. [5] Employers are increasingly willing to hire based on demonstrated skill rather than credentials. That is both a challenge and an opportunity: you cannot just point to a degree anymore, but you can build a credible AI skills profile without one.
The three-times growth in AI skill requirements over two years is particularly significant. It means the curve has not flattened. Companies that were cautious about AI requirements in 2023 and 2024 are now explicit about them. The question is no longer whether AI fluency will be required in your field. It is when.
Here is where a lot of professionals get this wrong. When they hear “AI skills,” they assume they need to learn Python, understand neural networks, or get a data science certification. That is not what most employers are looking for when they list AI skills as a requirement for non-technical roles.
The skills that create real value for business professionals are:
None of those require a technical background. All of them require practice. And all of them are genuinely learnable in weeks, not years, if you approach it the right way.
Some roles are more directly affected than others, and the impact is not uniformly distributed. Based on where AI skill requirements are growing fastest, the professionals who need to move most urgently are in these fields: consulting and strategy, marketing and content, finance and FP&A, HR and talent management, and legal and compliance.
These are all knowledge-intensive roles where AI tools can genuinely amplify output. They are also roles where the volume of written, analytical, and research work is high enough that AI fluency creates a measurable difference in what you can produce per hour. A consultant who can use AI to prepare a first draft analysis in 30 minutes versus 4 hours is simply more valuable.
Roles in operations, customer service, and project management are close behind. The pattern is the same: any role that involves processing information, drafting communications, or synthesizing data is being transformed by AI tools, and the people who know how to use those tools well are getting ahead of those who do not.
There is a problem with how most professionals approach AI learning. They watch a few YouTube videos, maybe read some articles (fair enough), sign up for a course, and then never quite get around to integrating it into their actual work. The learning stays theoretical. The skills do not stick.
The only approach that works is learning on real tasks. Here is a practical structure that works for busy professionals.
Do not sign up for five different AI tools. Pick one you already have access to: ChatGPT, Microsoft Copilot in your Office suite, or Google Gemini in Workspace. Use it every day on a real task from your actual job. Write emails with it. Summarize meeting notes. Draft sections of a report. The point is repetition on tasks that matter to you.
Good prompting is not magic. It follows a structure. At minimum: role + task + context + format. Here is an example: “You are a senior HR consultant. I need to draft a 3-paragraph performance review for a marketing manager who exceeded targets by 20% but had friction with two team members. Write in a professional but warm tone, using specific language about growth and development.” That prompt gets a useful output. “Write me a performance review” does not.
Prompt template to save: “You are [role]. I need [specific output]. Here is the context: [details]. Please format it as [format]. If you are uncertain about anything, flag it rather than guessing.”
Once you are using AI daily on real tasks, add a structured learning layer: a focused bootcamp, a role-specific course, or a team workshop. The learning will land much better when you already have a foundation of practical experience. You can see our AI Bootcamps and Corporate Workshops for options designed specifically for non-technical professionals.
Build one non-negotiable habit: always verify outputs that involve specific facts, numbers, or names. AI tools still get these wrong. The professionals who embarrass themselves with AI are almost always the ones who trusted an output without checking it.
This is the “what to do Monday morning” section. Here are three specific actions:
1. Identify your highest-volume repetitive task. What do you do most often that involves writing, analysis, or research? That is your starting point for AI integration. Not the most complex task. The most frequent one. Frequency builds skill faster than complexity.
2. Use AI for that task three times this week. Not once. Three times. The first attempt will probably be mediocre. The third will be noticeably better. That improvement is what builds the intuition for when and how AI adds value.
3. Add one verification step to your workflow. Before you send any AI-generated content, read it once specifically looking for factual claims, specific numbers, or names. Spot-check at least one. This habit alone will prevent the most common AI mistakes. See our piece on enterprise AI adoption trends for more on how leading companies are building AI fluency across teams.
What AI skills do employers actually want in 2026?
Employers primarily want practical AI fluency: structured prompting, AI-assisted research and analysis, output verification, and workflow integration. Technical coding skills are not required for most non-technical professional roles. Judgment and the ability to critically evaluate AI outputs are valued more than technical depth.
What is McKinsey’s Lilli AI interview?
McKinsey has introduced an AI-enabled exercise in select final-round interviews where candidates work with the firm’s internal AI tool, Lilli, on applied problem-solving tasks. The goal is to assess reasoning, judgment, and the ability to guide and synthesize AI outputs, not specialist AI knowledge. It currently sits alongside the standard case and behavioral interview, not replacing them.
How much more do professionals with AI skills earn?
According to Gloat’s 2026 AI Skills Demand report, professionals with advanced AI skills earn 56% more on average than peers in the same roles without those skills. The premium is highest in finance, consulting, marketing, and legal roles.
How can a non-technical professional build AI skills quickly?
Start by using one AI tool you already have access to on your most frequent work tasks, every day for two weeks. Real-task practice builds more useful skill faster than any course alone. After two weeks, add structured learning through a bootcamp or workshop focused on your role. The combination of practical experience and structured learning accelerates skill development significantly.
Is a technical background required to develop valuable AI skills?
No. The most in-demand AI skills for business professionals are non-technical: structured prompting, critical evaluation of AI output, workflow integration, and judgment about when to use AI and when not to. Employers in 2026 are prioritizing AI literacy and analytical thinking over formal technical credentials for non-technical roles.
About This Guide
Written by Sana Mian, co-founder of Future Factors AI and AI educator who has trained 2,000+ non-technical professionals. This article draws on publicly available hiring data, published research, and reported news about McKinsey’s hiring practices. Future Factors AI runs AI Bootcamps and Corporate Workshops designed specifically for non-technical professionals.
Sources
Weekly tips. Real examples. Practical help for busy professionals.
We care about your data, check out our privacy policy.
