Consulting Is Evolving - And Its Destination Is Software
There’s this FT interview with IBM Consulting’s head Mohamad Ali that landed on my desk recently, and honestly, it crystallized something I’ve been watching unfold across the consulting world.
Ali doesn’t mince words: “The future of consulting is going to be a hybrid of people plus software - like, a lot of software. And I think that consulting companies that can’t do that are going to fall away.”
The Numbers Tell the Real Story
IBM brought Ali back two years ago with one clear mandate: transform consulting into a “service as a software” business. And they’re actually executing on it. They’ve built thousands of digital agents that perform tasks with minimal human oversight. The result? $3.5 billion in savings over two years by deploying automation across 70 corporate functions.
That’s not a pilot program. That’s not “experimenting with AI.” That’s a wholesale transformation of how work gets done.
And it’s happening everywhere. I’ve been watching Globant with their AI Pods, McKinsey setting records for token consumption that OpenAI actually recognized publicly. From Deloitte and KPMG to McKinsey, everyone’s racing to build agentic AI platforms.
Meanwhile, Accenture’s shares are down 30% this year because investors are genuinely worried about what AI means for managed services.
The Competitive Pressure Is Getting Real
Here’s what Ali said that I think captures the urgency: “The software companies are going to get into this business as well, and are going to be delivering digitally through agents. So our strategy here is to run as fast as possible to deliver as much of this through these agents as possible and convert a big part of our business into a software business.”
Think about that for a second. The head of IBM Consulting is basically saying: we need to become a software company before the software companies eat our lunch. That’s not a future threat. That’s a present-tense competitive reality.
And you can see why. Consulting has been IBM’s weakest division this year with zero revenue growth, while companies spend heavily figuring out AI in-house. Why pay consultants when you think you can build the AI yourself?
What This Actually Looks Like in Practice
Let me give you some concrete examples from what we’re building, because I think the abstract conversation about “AI transformation” misses the point.
We recently built an SAP architect agent for a client. This thing analyzes two decades of system configurations and reasons about consequences of new implementations and upgrades. It’s not just searching through documentation. It’s actually thinking through the cascade effects of changes across a massive, interconnected system.
This saves hundreds of architect hours, sure. But that’s almost beside the point. What it really does is surface insights that were physically inaccessible before because of the depth and multidimensional nature of the context. No human architect, no matter how brilliant, could hold that much system state in their head at once.
We’re also building what I call an AI project factory. All documents, meetings, artifacts - everything exists in a living, dynamic AI environment. Teams can orient rapidly in project context, create detailed client instructions, configure business applications at speeds and precision levels that were just impossible before.
This is what Ali means when he says consulting is becoming software. The consulting work itself is being encoded into systems that can reason, analyze, and execute.
The Implementation Paradox
Here’s something interesting Ali pointed out: an MIT study showed that 95% of generative AI projects failed to deliver financial returns. 95%! That’s catastrophic.
Yet Ali predicts companies will ultimately need consulting firms to implement AI strategies successfully. Why? Because, as he puts it: “It’s less about the LLMs and now about the applications that you’re building on top.”
That’s the paradox right there. AI is simultaneously threatening to eliminate consulting and creating massive demand for the expertise to implement it properly. The companies that figure out how to build those applications - not just use ChatGPT, but build real business applications on top of these models - are the ones that survive.
The Existential Question
I keep thinking about what this transformation really means. Consulting has always been about selling expertise, right? You hire McKinsey or IBM because they know things you don’t, they’ve seen patterns you haven’t, they can bring best practices from across industries.
But what happens when that expertise gets encoded into software? When does pattern recognition happen algorithmically? When the “best practices” are embedded in agents that can configure themselves?
The consulting firms that don’t transform into software businesses - and I mean really transform, not just add an AI practice - are going to face serious challenges. Because AI is shifting from competitive advantage to table stakes. It’s becoming a necessary condition for doing business, not a differentiator.
What I’m Watching For
I’m curious what patterns other people are seeing in professional services. Because from where I sit, the transformation feels like it’s accelerating fast. The firms that are treating this as “let’s add some AI tools” versus “let’s rebuild our entire value delivery model as software” - that gap is going to be impossible to close once it gets too wide.
Ali talks about running as fast as possible. That might be the most honest assessment of where things stand. Not careful experimentation. Not measured adoption. Running. Because the alternative is falling away.
The future of consulting isn’t hybrid human-AI work. The future of consulting is software that happens to involve some humans. That’s a fundamentally different thing.
And I think we’re going to see which firms understood that distinction pretty clearly over the next couple years.