Will AI Replace Business Analysts? A View From Both Sides

Business analysts have always been hard to pin down. "What do you actually do?" is a question most BAs have answered at family gatherings more times than they'd like. The answer usually involves a lot of hand-waving about requirements and stakeholders and processes.

That vagueness might actually be your greatest asset right now.

I was a data scientist before AI restructured my role out of existence. Now i consult on AI strategy and sit in the meetings where companies decide what to do with their teams. Business analysis comes up frequently. Here's what i'm seeing.

What AI can already do in business analysis

Requirements documentation. AI can take meeting notes, interview transcripts, and existing documentation and produce requirements specifications. Not perfect ones. But first drafts that are structured, comprehensive, and need human refinement rather than human creation. That's a significant change.

Process mapping. Describe a business process to an AI tool and it'll produce a flowchart, identify inefficiencies, and suggest improvements. Tools integrated into Visio, Miro, and specialist process mining platforms do this semi-automatically now. The process mapping that used to take a BA a week? A few hours with AI assistance.

Data analysis. Business analysts who leaned heavily on the "analysis" side of their role are in similar territory to data analysts. AI handles SQL queries, data visualisation, trend identification, and report generation. If your primary value was pulling data and presenting it, AI does that faster.

User story generation. AI can take a high-level requirement and break it down into user stories with acceptance criteria. They're formulaic. They follow the template perfectly. They lack nuance. But they're a starting point that didn't exist before.

Gap analysis, competitor research, market sizing. AI can gather and synthesise information from multiple sources and produce analysis that's... serviceable. Not insightful. But serviceable.

Impact assessment and change analysis. AI can model the downstream effects of a proposed change across systems and processes. It's not always right, but it's thorough in a way that humans sometimes aren't.

What AI still can't do

Right. Here's where BAs have more protection than they might think.

AI cannot sit in a room with five stakeholders who all want different things and facilitate a conversation that ends with alignment. That's not a technical skill. That's a human one. It's politics, diplomacy, listening, and the ability to find common ground between people who are convinced there isn't any.

Eliciting requirements that people don't know they have. The most important requirements are often the ones nobody mentions because they're so fundamental they've been assumed. The skill of asking the right question at the right time, of watching someone's face when they describe a process and noticing the hesitation that reveals a hidden problem... AI doesn't do that.

Organisational politics. Knowing that the real decision-maker isn't the person with the title. Understanding why the finance team always blocks IT projects. Recognising that the stated reason for a change request and the actual reason are completely different things. This is the dark art of business analysis and it's entirely human.

Change management. Not the Gantt chart kind. The human kind. Understanding why people resist a new system. Knowing how to bring sceptics on board. Being the person who translates between the technical team and the business users because you speak both languages. AI generates change plans. It doesn't manage change.

And the creative problem-solving. When the obvious solution doesn't work and you need to think laterally. When the constraint isn't technical, it's political. When the best answer is "actually, we shouldn't build this at all." That kind of strategic thinking requires understanding the business context in a way AI simply doesn't.

The honest assessment

In the restructuring meetings i attend, business analyst roles are being affected, but not as dramatically as some other professions. Which is interesting.

The reason, i think, is that BA work is inherently cross-functional and relationship-based. It's harder to automate a role that's primarily about connecting people and translating between different parts of an organisation.

That said, the roles being cut are the more technical BAs. The ones whose primary output was documentation, data analysis, or requirements specifications. If your job was mostly writing documents, AI can do that. If your job was mostly facilitating humans, you're safer.

The pattern i see: a BA team of eight becomes five or six. The remaining BAs are the ones who are embedded in the business, who know the stakeholders, who facilitate and influence. The ones who were primarily technical writers or data analysts under a BA job title are the ones being let go or redeployed.

Agile environments are interesting. The BA role was already under pressure in Agile teams where product owners do some of what BAs used to do. AI adds another pressure. But the BAs who've adapted to work as facilitators, user researchers, and domain experts within Agile teams are holding their ground.

The BCS and IIBA certifications still matter. They signal a breadth of capability that goes beyond any single automatable task. If you're a BA without formal certification, it's worth considering.

One thing i've noticed: BAs who can code, even a little, are more secure than those who can't. Not because they're doing development work, but because they can prototype solutions, understand technical constraints, and have more credible conversations with engineering teams. The hybrid BA-technical person is a harder role to automate.

What to do this week

1. Assess your role honestly. What percentage of your week is facilitation and stakeholder management versus documentation and analysis? If it's heavily skewed toward documentation, you need to rebalance. The facilitation is your career insurance.

2. Facilitate one meeting without preparing a document. Run a workshop, a requirements session, a stakeholder alignment meeting. Your value is in the room, not in the artefact. Practise being the person who makes the conversation productive.

3. Use AI to produce your next requirements document. Feed it your notes and see what it produces. Then refine it. If the refinement takes you thirty minutes instead of four hours, you've just freed up time for higher-value work. Use that time wisely.

4. Build a deeper relationship with one stakeholder. Not about a specific project. About their world. Understand their pressures, their goals, what keeps them up at night. The BA who understands the business is infinitely more valuable than the BA who understands the methodology. Project managers know this dynamic well.

5. Learn a bit of code. Python, SQL, whatever's relevant to your domain. Not to become a developer. But to be the BA who can prototype, query data directly, and speak the technical language credibly. It's a differentiator.

If the uncertainty is getting to you, AI replacement dysfunction is worth reading about. And if your company is showing signs of restructuring, better to know early.

The one thing to do today: think about the last time you resolved a disagreement between stakeholders. That moment, where two humans needed a third human to help them find common ground, is your entire career justification. AI doesn't do that. You do.