AI can produce requirements documents. Therefore, the argument goes, what is left for the BA to do?
I want to engage with that question seriously, because it deserves a serious answer. Not a defensive one, not a cheerleading one, but an honest look at what business analysts actually do across the full lifecycle of a project and why that work is not going away.
Here is my answer: if you genuinely believe AI can replace a skilled business analyst, you have a narrow and outdated picture of what business analysis is. You are looking at the visible, documentable surface of the role and missing the analytical, relational, and judgment-intensive work underneath it. That work is not only surviving in the age of AI. In many cases, it is becoming more critical, not less.
Let me show you what I mean.
The Mistake Everyone Makes About the BA Role
The confusion about AI and business analysis comes from a real place. AI tools are genuinely capable of performing many of the tasks that BAs have historically spent significant time on. Writing user stories from meeting notes. Drafting acceptance criteria. Producing first versions of process documentation. Generating business rules from policy documents. These are real capabilities, and they are changing how BAs work.
But here is the mistake: conflating the tasks that occupy BA time with the work that creates BA value.
A significant portion of what BAs have historically spent time on is: Capturing, organizing, formatting, and documenting information so that delivery teams can act on it. AI can do a lot of that work. That is a genuine shift, and BAs who have not adapted to it yet will feel pressure.
But the analytical, strategic, relational, and judgment-intensive work that makes business analysis genuinely valuable, the work that determines whether an organization solves the right problem, builds the right solution, and actually achieves the outcomes it set out to achieve, that work is not production work. It cannot be automated. And it spans every phase of the project lifecycle, from the first conversation about a business need all the way through to evaluating whether the delivered solution is working.
Let me walk you through that spectrum.
Early: Defining the Right Problem Before Anyone Builds Anything
The most expensive mistake in technology delivery is building the wrong thing. It happens constantly, in organizations of every size and maturity level. A team spends months building a solution to a problem that was not correctly understood, that was the wrong priority, or that could have been solved more simply with less investment. By the time the gap between what was built and what was actually needed becomes visible, the cost of correction is enormous.
The BA’s work at the front of a project lifecycle is precisely about preventing this. It is about asking hard questions before anyone starts building. What problem are we actually solving? How do we know it is the right problem to solve right now? Who is affected by this problem, and do we understand their needs accurately? What does success look like, and how will we measure it? What assumptions are embedded in the proposed solution, and have we tested them?
These are not questions AI can answer. AI can help research context, generate hypotheses, and surface relevant data. But deciding which problem is worth solving, navigating the organizational politics of prioritization, challenging assumptions that powerful stakeholders have made, and building the shared understanding of purpose that a team needs to stay aligned through a long delivery effort, that requires human judgment, organizational knowledge, and relational credibility.
It also requires the willingness to slow things down when slowing down is the right call. When the urgency to start building is high and the problem definition is still fuzzy, the BA’s job is sometimes to push back and insist on clarity before momentum builds. AI does not do that. People do.
The strategic, upfront work of business analysis is where organizations avoid their most expensive mistakes. A skilled BA doing this work well is one of the highest-return investments an organization can make in any project. That is not changing. It is getting more important as AI accelerates delivery timelines and compresses the window for course correction.
During Delivery: Judgment, Facilitation, and Human-AI Workflow Design
Once delivery begins, the BA’s role shifts but does not diminish. This is where the visible production work has historically lived, and it is where AI assistance is creating the most obvious change. Requirements documents, process models, user stories, acceptance criteria. AI can generate first drafts of all of these, and that genuinely changes how BAs spend their time during delivery.
But look at what does not change.
Stakeholder alignment during delivery is a continuous, active effort. Stakeholders change their minds. New information surfaces. Competing priorities emerge. Assumptions that seemed solid turn out to be wrong. The BA’s job is to keep the delivery team and the business stakeholders aligned through all of that, facilitating the conversations that resolve ambiguity, escalating the decisions that need leadership attention, and redesigning scope or approach when the evidence points that way.
AI cannot have those conversations. AI cannot read the room when a key stakeholder is losing confidence and the project needs a different kind of engagement. AI cannot navigate the interpersonal dynamics of a disagreement between a business owner and a technical lead and find a path forward that both can commit to. Those are human skills, and they do not become less necessary when AI is handling the documentation.
There is also a genuinely new category of delivery work that is appearing in AI-enabled projects: the design of human-AI workflows. When AI is part of what is being built, someone needs to define how humans and AI systems will work together. Where the AI acts. Where humans review. What the handoffs look like. What the guardrails are. What happens when the AI produces something unexpected. This is analytical, design-intensive work that requires both business understanding and an ability to think about system behavior in new ways. It is BA work, and it is work that has barely existed until recently.
The delivery phase of a project in the age of AI requires less production work from BAs and more of the analytical, facilitation, and design work that has always been the most valuable part of the role. That is a good trade for the profession.
After Delivery: Evaluation, Governance, and Continuous Improvement
This is the phase where BA involvement has historically been most inconsistent, and where the opportunity in the age of AI is perhaps most underutilized.
Most project teams declare victory at launch and move on. The BA’s involvement ends when the solution ships. But whether the solution actually delivers the business value it was supposed to deliver, whether the outcomes promised in the business case are materializing in practice, whether the process changes intended by the project are actually sticking with real users in real conditions, all of that goes unmeasured or gets measured by someone else who was not involved in defining what success was supposed to look like.
In an AI-enabled environment, this gap is not just a missed opportunity. It is a governance risk.
AI systems behave in production in ways that are not always predictable from testing. They encounter edge cases that the design did not anticipate. They drift toward outcomes that were not intended. The business rules they apply evolve as the business context changes, but the system does not automatically update. Without someone actively monitoring whether the AI-enabled process is producing the outcomes it was designed for, problems accumulate quietly until they become visible crises.
The BA is the person best positioned to own this evaluation and governance work. They were there when the success criteria were defined. They understand the business intent behind the solution. They have the relationships with stakeholders to surface concerns early and the analytical skills to evaluate whether what is being observed in production reflects the design that was intended.
This includes asking hard questions after launch. Is the solution being used the way it was intended? Are the business outcomes we promised materializing? Where is the AI-enabled process producing unexpected results, and what does that tell us about where the design needs to evolve? What should be adjusted, and how do we make the case for that adjustment to the people who control resources and priorities?
That is continuous improvement work, and it requires exactly the same skills as the upfront work: analytical rigor, stakeholder relationship, and the ability to translate evidence into decisions. The lifecycle of a BA’s work does not end at launch. In an age of AI, it arguably should not end at all.
The Full Picture
When you look at the full spectrum of what business analysts do, from defining the right problem before delivery begins, through aligning stakeholders and designing human-AI workflows during delivery, to evaluating outcomes and governing AI-enabled systems after launch, the replacement argument falls apart.
AI is changing the production layer of BA work. It is not touching the judgment layer, the relational layer, or the strategic layer. And it is creating entirely new categories of BA work around human-AI collaboration design and governance that did not exist before.
The BA who understands this is not anxious about AI. They are paying attention to it carefully, learning how to work with it effectively, and positioning their value in the places that AI cannot reach. Those places are not narrow or marginal. They span the entire project lifecycle and represent the highest-leverage work in every phase.
The real risk for BAs is not AI replacement. It is being so occupied with production work that has become automatable that there is no time left for the work that is not. That is a choice, and it is one that individual BAs and the organizations they work in can make differently.
Build the Skills That Matter
The BA skills that thrive in the age of AI are learnable. Strategic analysis, stakeholder facilitation, human-AI workflow design, outcome evaluation, and continuous improvement are all disciplines that can be developed with the right guidance and practice.
I teach this work in my Maven course series, with practical frameworks and real-world scenarios designed for BAs who want to lead in an AI-enabled environment across the full project lifecycle.
Visit www.maven.com/angela-wick to explore current courses and upcoming cohorts.
AI is not coming for the BA role. It is coming for the parts of the BA role that were never the most valuable parts to begin with. The rest is yours.