A procurement lead or a project sponsor now opens ChatGPT and asks which firms in India handle a specific problem, and gets a list. That list is often the first version of the shortlist, assembled before any RFP is written and before anyone visits a website. For consulting firms this is a sharper threat and a sharper opportunity than in most categories, because the material that would make you a credible answer is usually sitting behind a lead capture form where no model can read it. Meanwhile the Big Four publish enough ungated material to be named by default. The work is making your firm and your partners legible as expert entities, and putting your substance somewhere a machine can actually reach it.
Consulting is one of the categories where assistant-mediated discovery has moved fastest, and the reason is that the questions map neatly onto what these systems do well. Someone asking which firms in India have depth in a specific regulatory change, or what a reasonable engagement model looks like for a particular kind of programme, is asking a synthesis question that a results page answers poorly. The buyer using the assistant is also frequently not the ultimate decision-maker but an analyst or a project manager tasked with building an initial list, which means the assistant's answer shapes who is even considered. Against that, professional services websites are structurally hostile to extraction. The substance sits in gated PDFs, the capability pages carry positioning language rather than facts, the practice names are inconsistent, and the partners who constitute the actual expertise are represented by a photograph and two sentences. The models fall back on Big Four material, which is ungated and voluminous, on directories and rankings, and on trade press. That creates a genuine opening for mid-sized and boutique firms, because depth on a narrow problem is easier for a model to attach to a specific query than breadth across everything, and a boutique with genuine specialisation can be named alongside far larger firms on a sufficiently specific question. Two honest caveats. Attribution is poor, since a committee that finds you through an assistant typically arrives later via branded search or direct. And the field is young enough that nobody has long-run data, so anyone selling certainty is selling something else. We run this from Powai, Mumbai.
