A GC at a Series A startup described the moment to us with some precision. An employment question landed on a Tuesday afternoon, the kind that would have gone straight to external counsel six months earlier. She ran it through Habeas instead, traced the relevant authorities, formed a view, and answered the board before close of business. The invoice she did not send to a law firm was, she estimated, somewhere between three and eight thousand dollars. More than the cost of the subscription for the month.
The conventional story about in-house legal teams goes like this: budgets tightened, boards lost patience with external counsel invoices, and GCs responded by taking on more work themselves. Routine matters stayed internal. Complex, high-stakes work still went out. Everyone adapted, and the basic architecture of the in-house and external counsel relationship survived intact. Just a slightly different distribution of labour.
We think this understates what is actually happening. The surface story is accurate. The deeper one concerns something structural: in-house counsel drawing a new line, and the position of that line changing the nature of everything on both sides of it.
For most of the last thirty years, the default mental model was that external counsel operated as a kind of expert safety net. In-house teams handled the day-to-day. When something felt uncertain or high-stakes, it went out. The threshold for going out was set by a mix of risk tolerance and the simple fact that getting an authoritative answer in-house was slow and difficult. This gave external counsel a generalist function that went well beyond true specialism. A significant portion of what law firms billed for was confidence. The GC sent a question out, got an answer back on firm letterhead, and could say to the board or the regulator: we took advice. The invoice was partly for the analysis and partly for the indemnity that came with it.
Two very different things were being conflated: the need for genuine specialist expertise and the need for a quick, reliable answer to a question that was not, in itself, particularly hard. When those two things cost the same and come from the same source, nobody has much incentive to separate them.
What has changed is the cost of answering questions in-house. When an answer that would have taken half a day of trawling through legislation and case law can be assembled in minutes, with traceable citations and grounded analysis, the threshold for going out shifts accordingly. Foundational research that used to take a full morning can now be completed in minutes. That compression does not eliminate the need for specialist judgment. It eliminates the need to pay specialist rates to establish the basics before specialist judgment can begin.
Visibility matters as much as speed. A GC who can rapidly form a well-grounded view on a question, before deciding whether to escalate, is in a fundamentally different position than one who routes everything out because the alternative is too slow. Having actually read the landscape, she knows whether the question is genuinely hard or merely unfamiliar. She can engage external counsel at a level of precision that was previously difficult to achieve.
One GC we have spoken with put it plainly: "External counsel used to be the thought partner. Now Habeas and AI fill that role. Over time, I expect legal spend on external counsel to be a fraction of what it is today, reserved mainly for highly specialised or strategic input."
This is a practitioner describing a structural reconfiguration, not a budget adjustment. The thought-partnership function that law firms have historically provided is moving. What remains external is the work that genuinely cannot be done any other way: contested questions, positions taken under uncertainty, advocacy or strategic counsel that could not plausibly come from any other source. Some law firms are adjusting to this. Many are not. The billable-hour model was built on the assumption that clients needed a reliable route to authoritative answers and were willing to pay for the combination of answer and authority in a single invoice. That bundling is coming apart.
There is a version of this story that positions tools like Habeas as a cost-cutting measure: a way for in-house teams to squeeze the external counsel budget, to survive headcount constraints. We find this framing accurate but insufficient. A more interesting account concerns where intelligence sits in the legal function. When in-house counsel can answer a much wider range of questions confidently, with work grounded in Australian primary law and traceable citations rather than generated plausibility, the whole architecture of how a legal team operates changes. External counsel becomes a specialist resource rather than a general one. Internal judgment carries more weight because it is better informed.
Habeas was built for this. Semantic search across over 300,000 Australian cases and pieces of legislation, Research Assistants that return structured, cited analysis, Document Stores that let a GC query her own matter materials: these are features designed to change where the work actually happens, to shift which questions require external engagement and which can be resolved before that decision is even made. "It feels like having a law firm in your pocket," as another GC described it. "Not something you blindly bet the house on, but a powerful first-line legal intelligence tool."
A quiet revolution rarely announces itself in headlines. It accumulates in the changed nature of instructions going out, the changed nature of work coming back, and the shifted relationship between the two. Firms that understand this and adapt, becoming genuinely specialist rather than generalist-with-prestige, will be worth more. The GC who has already formed a grounded view, traced the authorities, and understood the lay of the land is asking a different question of external counsel than the one routing a query out because the alternative is too slow. That shift in the question changes everything that follows from it.
If you would like to see how Habeas operates as a first-line layer, you can book a demo at habeas.ai.
The legal research in this article was conducted and every citation verified using Habeas, the Australian legal AI research platform.
Hero image: Vitaly Gariev on Unsplash
