AI Is Forcing Consulting to Reinvent Itself — Or Be Left Behind

The management and technology consulting business is facing tremendous pressure as artificial intelligence picks up steam. While the initial wave of AI led to huge contracts to consultants to help companies figure out how to use and deploy new intelligence technologies, clients are asking harder questions now. Questions like “Why do you need so many junior consultants or business process workers when AI can automate much of their work?”.  

These customers point to soundbites from Satya Nadela of Microsoft saying that he expects AI to replace big chunks of knowledge work. Nadela notes his employees on CoPilot are far more productive than ever before, and that he anticipates a future where each worker is AI conversant and generates significantly higher levels of revenues. The revenue per employee question is a critical one. A group of AI-first startups have broken out with eye-popping metrics. For example, Midjourney, which makes an image generation and editing platform, reported annual revenues of $500 million while employing only 40 people. In other words, AI-first businesses can generate revenues far greater than previous business models and technology stacks.  

Forward-leaning companies like fintech startup Klarna have moved to leverage AI customer support and AI coding tools to make internal software development more efficient (and to whittle down its workforce while continuing to grow revenues). This has recast the build/buy equation, a calculation that will become more and more competitive as AI coding tools continue to improve. The rise of software agents, too, is calling into question the entire cost structure and value of business process outsourcing. Already under attack by robotic process automation (RPA), business process outsourcing (BPO) giants now face an uncertain future where they are competing with infinite numbers of AI agents that cost a fraction of a human operator. These fleets can benefit immediately across the entire fleet from upgrades to capabilities and data access via AI and AI tools.  

This doesn’t mean BPO from consultants disappears. Rather, BPO (in any form) must pivot from people‑focused task execution to transformation‑focused process automation and improvement — which requires entirely new metrics. By radically shifting economics, SLAs, and delivery models, AI-native providers that productize and seamlessly integrate intelligent automation will outpace incumbents who resist or struggle under legacy constraints. Success in this new era requires a holistic reinvention of end‑to‑end processes — far beyond IT upgrades or adding RPA to rote tasks — to deliver genuine business process transformation rather than mere digital interventions on specific tasks or cost‑cutting initiatives. 

Not surprisingly, based on all this, clients want more: faster results, smarter insights, real impact. Delivering the same consulting outputs is no longer sufficient. Cost-cutting mandates for technology contracts from multiple agencies in the U.S. government have underscored the brutal new reality. In speaking with peers, every organization that does consulting is re-evaluating its future. They recognize that the old consulting playbook no longer applies. What’s emerging is a new model built on speed, outcomes, and automation. Consulting firms that fail to adapt will be eclipsed by leaner, AI-native competitors. 

From “Hours Billed” to “Outcomes Delivered” 

For decades, consulting firms sent in fresh MBAs to write reports and build slides. AI tools today can accomplish more than 50% of what these junior analysts delivered. Similarly, AI coding tools can automate significant portions of software development, specifically lower-value jobs. In BPO, the problem is even worse. Stability AI founder Emad Mostaque put it bluntly: AI will “completely destroy” the BPO market. Andreessen Horowitz describes this shift as "unbundling the BPO" — not just automating it, but productizing it into vertical AI agents that specialize in customer support, claims processing, invoice reconciliation, and more. What used to take dozens or even hundreds of humans now take one AI system, running 24/7, trained on deep context. 

As AI absorbs the bottom of the consulting pyramid, the economics of the industry are being inverted. Customers are asking to pay per result — as in, number of customer support calls processed, number of ServiceNow tickets resolved, number of credit card transactions processed. They are responding to new AI-driven pricing models in SaaS that are destroying older per-seat, all-you-can-eat economics. This is an existential moment for consulting firms. 

AI-Native Consulting 

In many ways, the past five years has steadily redefined what clients should expect from enterprise software and consulting. The past model of legacy firms billing for six months of discovery before delivering anything tangible is over. After seeing their own internal technology development velocity increase, customers expect functional prototypes and measurable value in weeks. Not quarters.  

This speed is not a stunt. It’s a strategic differentiator. This requires embedding development engineers directly into client operations not only to learn but to immediately start building code, blurring the line between software vendor and consultant. AI remains at the core of everything, the special sauce that makes this new approach possible. The change in expectations is irreversible.  

So, What Can Consulting Firms Do to Thrive? 

Not all is dark. Consulting domains better aligned to outcomes-based commercials (e.g. BPOs with a pay per transaction model) are already seeing greater adoption and disruption with AI and they are competing effectively against upstarts. Consulting firms that effectively transition to a pay-per-performance model with BPO will be more like product companies, as A16z proposes, and will fully re-platform to put AI at the center of all their offerings. 

While AI has already proven to boost productivity across a wide range of consulting roles in experiments, most clients do not have in place measurable KPIs to quantify benefits. This makes it difficult to pivot to different commercial models, so this transition needs to be an iterative journey and partnership between clients and consulting shops, with both open to experimentation.  

Even though clear cost-per-result models can deliver demonstrable value, the start-up costs tend to be higher because of dependencies on high-quality data, well-trained and tuned models, and large infrastructure outlays. In other words, context matters and that last-mile of context is where much of the value in AI lies. For this reason, we hear from our partners at Microsoft, for example, that there is an advantage in AI chief building AI models ‘3 or 6 months behind’ the cutting edge because the context and not the core model is the true source of customer value. 

Above all, we can expect further fundamental disruption in the “Ways of Working” and “Delivery model” driven by next generation “agentic AI” consulting workflows. This will result in unprecedented levels of automation, particularly for relatively forgiving tasks that are easy to check and error-correct. Given the above dynamics, “gain-share” commercial models are gaining traction. This allows the benefits of AI to be split 50-50% between clients and consulting firms. In other words, shared risk, shared reward. 

The Path Forward 

To survive, consulting firms must move beyond legacy billing structures, outdated hierarchies, and slow delivery cycles. This will be uncomfortable with significant disruption but will result in leaner, more outcome focused with the economics of consulting firms closer to software businesses. 

Consulting firms need to help organizations get prepared with an "AI for Business" foundational framework (implementing KPIs to track benefits, changing Ways of Working etc.) In this future, consulting firms will be judged and compensated on four attributes, what we call the 3VQ Commercial Model: 

·       Value - Cost 

·       Velocity - Speed 

·       Volume - Measurable Outputs 

·       Quality - Accuracy and usefulness 

Context will become ever more important. Consulting will be re-imagined with industry or domain-aligned Agentic AI Workflows, and the returns will be highly disruptive with greater returns to the highest performers. To survive and succeed, next generation consultants will need to be “multi-model” proficient, as clients will select from a variety of AI platforms, models and tech stacks, based on preferences, location and use cases. 

Ismail Amla is Senior Vice President and leader of Kyndryl Consult.

Views expressed here are my own.