Consulting is dead!
The firms that keep selling advice as theatre will feel the pain from AI disruption. The firms that will win are deep in the deployment layer proving where AI changes the economics.
Another social post declaring “this one prompt will replace McKinsey” or my favourite, “do what takes (insert any management consulting firm here) 6 months in 6 mins with this prompt”.
AI can draft the deck, summarise the interview, search the market, build the benchmark and produce the steering committee pack before the junior consultant has found the client’s brand guidelines.
The obituary works because part of it is true. The market has now declined at a five-year annualised rate of 0.9%. Public-sector buyers are rebuilding internal capability after the PwC tax scandal and the broader backlash against outsourced advice. Corporate procurement teams are asking harder questions about rates, deliverables and measurable value.
The other part that is true is that the Australian management consulting market is A$45.8 billion in revenue, employs roughly 130,000 people and produces A$6.6 billion in profit.
The category is not disappearing anytime soon, but AI disruption is a problem, and it is better understood as a business model problem.
Management consulting serves the market through three primary business models: the Apex, the Multiplex and the Integrator. AI hits each model through a different part of its profit engine. Two are being squeezed, one is expanding.
The Apex
The Apex sells judgement under ambiguity. McKinsey, BCG and Bain earn their premium when the client faces a decision where the cost of being wrong dwarfs the cost of the engagement. The work might be CEO succession, market entry, major M&A, existential disruption or a board-level strategic choice where no one wants to own the answer alone.
AI weakens the analyst layer underneath that model. Research, synthesis, market mapping and scenario work all become faster and cheaper. Internal strategy teams with good tools can now do work they once outsourced because the external firm had better access to benchmarks, case libraries and junior labour.
The Apex survives where judgement remains scarce. A model can summarise the last fifty market entries. It cannot sit in the boardroom and absorb accountability for the next one. It cannot lend institutional credibility to a decision that will be challenged by investors, regulators or a divided executive team. The premium narrows as the analysis layer commoditises, but the model does not collapse while buyers still need trusted judgement on decisions they cannot easily reverse.
The Multiplex
The Multiplex has the harder problem. Deloitte, KPMG and EY sell breadth, risk transfer and access across audit, tax, risk, technology, workforce and advisory. Everything under one roof. The model works because partners hold senior relationships while junior-heavy pyramids produce the work: analysts research, consultants model, managers package and partners sell the next tower.
AI attacks that structure directly. If a large share of the work can be accelerated by software, buyers will ask why they are paying for the same pyramid. The firm can keep some productivity gains by moving to fixed fees or outcome-based pricing, but the client has the same tools and the procurement function has the same suspicion. Once the buyer believes the output costs less to produce, premium day rates become harder to defend.
Australia adds a second force. The public-sector reset did not come from AI. It came from trust, politics and procurement reform. Government buyers have narrowed the scope for external advice on core public service work, increased scrutiny of conflicts and rebuilt internal capability. That pressure lands hardest on firms that sold themselves as the default safe choice across everything.
Each decision inside the Multiplex still makes sense in isolation: automate the junior work, offshore where possible, defend partner relationships, and sell AI transformation to clients while using AI to improve internal margin. Together those decisions expose the contradiction. The model is trying to monetise AI outside the firm while using the same technology to rationalise the labour structure that makes the firm profitable.
The Integrator
The Integrator sits in a different position. Accenture, Capgemini, TCS, and IBM sell delivery certainty at scale. Their promise is less “we know what you should do” and more “we will take the delivery risk off your P&L.”
AI strengthens that model because enterprise AI value is not realised in the demo. It is realised in production: workflow redesign, data architecture, integration with legacy systems, security, change management, governance and measurement. The Integrator already lives in that layer. It has the platform partnerships, account relationships, delivery muscle and blended labour model to turn a board-level AI ambition into something that actually runs.
Accenture’s public numbers show the direction of travel. In FY2025 it reported US$69.7 billion in revenue, US$2.7 billion in generative and agentic AI revenue, and US$5.9 billion in generative AI bookings. BCG, an Apex firm with a serious build capability, reported that AI and technology-focused services now represent more than 40% of its 2025 revenue. OpenAI has also moved towards the deployment layer through partnerships with Accenture, BCG, Capgemini and McKinsey, and through the OpenAI Deployment Company, a majority-owned vehicle launched with more than US$4 billion of initial investment.
The deployment layer is where value concentrates
Most enterprises do not lack AI enthusiasm. They lack the operating conditions required to turn AI into value. Many pilots produce no measurable P&L impact because they sit on weak data foundations, vague success metrics and workflows that were never redesigned around the technology. A chatbot on top of a fragmented organisation does not become a transformation programme. It becomes a more fluent interface to the same confusion.
The deepest value will sit inside complex companies with vertical knowledge. A bank, insurer, telco, university or energy retailer does not run on public information. It runs on undocumented rules, exceptions, compliance habits, product quirks, customer histories, pricing decisions, manager judgement and systems that disagree about basic definitions.
Many large enterprises for example, cannot consistently answer simple questions like: who is this customer across our systems, what are they worth to us over time, what did we do to them, what should we do next, and did the intervention create value? Customer level data sits across CRM, billing, service, product, complaints, consent, marketing, risk and external sources. Each function carries a different definition of the customer.
The next valuable consulting model will not sell generic AI strategy. It will build the semantic foundation that lets a company use AI on its own operating environment. In the customer domain, that means a client-owned customer ontology, a knowledge graph that connects relationships and events, a customer value model that links behaviour to economics, and a measurement layer that can tell whether the work changed retention, cross-sell, cost-to-serve or risk.
This is where the Integrator model can compound, provided it does not reduce the work to technology installation. The first conversation still has to be a value conversation, not a platform conversation. The client does not need another AI roadmap that names the same use cases as every competitor. It needs to know where the customer base creates and destroys value and where AI can change those economics.
Consulting is not dead
But generic analysis is being repriced, pyramid labour is being compressed and trust without measurable outcomes is weakening. The firms that keep selling advice as theatre will feel the pain first.
The firms that win will combine domain judgement, owned data foundations, workflow integration and value measurement. By 2030, the interesting question will not be which firm wrote the best AI strategy. It will be which firm stayed inside the enterprise long enough to prove that the strategy changed the economics.
Grada publishes strategic analysis for executives navigating complex markets