Kully Thandi, global GC at Allianz Technology Group in Germany, explores whether legal teams are shaping the future of the legal function or simply inheriting it, examines the balance between guardianship and enablement in a regulated environment, and considers what GenAI reveals about modern legal leadership
Every legal leader reaches a point where the honest question is: am I shaping the Legal Function or am I simply maintaining it?
I have come to believe that how you answer this question is one of the most revealing things a legal leader can offer. My answer begins not with what we built, but with what no amount of building can do for you.
Allianz Technology is the captive provider of IT services to the Allianz Group — and the engine of its digital transformation. Across Allianz Group’s global footprint, Allianz Technology is the means by which the Allianz Group modernises: standardising platforms, industrialising services and building the technology infrastructure that allows a global financial services organisation to operate at pace in a rapidly changing world. The legal function sits at the centre of that, not as a buffer between commercial ambition and regulatory scrutiny, but as an active player in that agenda.
That participation demands two things simultaneously: guardianship and enablement. Finding the right balance between them is, I think, the central and never fully resolved challenge of the role.
A well-designed legal function can hold both guardianship and enablement in its architecture: in its mandate, in its structure and in the way it deploys its people. That architecture handles a great deal, and it is worth building carefully. But it has a limit that every legal leader eventually confronts. The clear cases — where the answer is unambiguously yes or no, protect or proceed —are rarely the difficult ones. The work lives mostly in the grey: situations where the risk is real but the right response is genuinely contestable, where protection and facilitation pull in different directions or where the blueprint points towards a principle but cannot resolve the specific.
Those readings require something the design cannot supply —experience, judgement and a quality of contextual intelligence that only develops over time and under pressure.
In this context, the gap between the blueprint and the build is where judgement does its real work. A company driving transformation at this scale, inside one of the world’s most closely supervised financial groups, does not accommodate a function that conflates its design with its execution. The pace and ambition of that agenda demand something more than procedural correctness. The blueprint creates the conditions in which good decisions can be made, it cannot make them.
What does GenAI reveals about the Paradox?
Like most legal functions, we are already working with GenAl in parts of what we do —and approaching that with genuine curiosity about where it can add real value. The question of where GenAl genuinely helps, and where it does not, is one worth working through carefully rather than answering too quickly. That process has, if anything, sharpened a conviction about what technology cannot do —and why that matters more rather than less as the tools improve.
What technology cannot do is sit in a room, or on a call, and tell a senior executive that the path they are pursuing carries a risk they have not fully accounted for. It cannot read competing interests across stakeholders with different agendas and different tolerances for risk. It cannot calibrate what to say, to whom, and when, under conditions of genuine uncertainty. And it cannot make a clear, accountable call when the answer is unclear and time is pressing. These capabilities —contextual, relational, deeply human — sit at the heart of what the role of the function requires. That does not appear to be changing.
What is changing is the context in which they are exercised. As GenAl takes on more of the routine work, what remains for the legal function is increasingly the non-routine: the structuring, the regulatory navigation and the moments where the stakes are real and the answer is not in the blueprint. In an environment like this one — where the pace of technological transformation routinely outstrips the frameworks designed to govern it — that is not an occasional condition. It is a permanent one.
The legal function operates, by design, at the edge of what the existing blueprint covers. A function that embraces GenAl does not reduce its exposure to those moments, it concentrates towards them. The paradox sharpens as the tools improve. The more a function evolves in this way, the more precisely it must know which moments still require judgement — and the more consequential it becomes when that reading is wrong.
Is your legal function shaping outcomes, or simply maintaining control?
This is, ultimately, what I mean when I return to the opening question. Shaping a legal function is not principally about what you build. It is about the quality of judgement the architecture is designed to enable — and whether that judgement is being exercised, consistently and at the right moments, across the function.
In a regulated financial services environment, the guardianship responsibility is not optional or negotiable. The business must be protected from regulatory risk, from contractual exposure, from the reputational consequences of getting complex things wrong. That is the foundation. But a legal function that only protects — that defaults to oversight and control as its primary expression of value — will, over time, find its contribution diminished. Enabling the business to move, grow and respond to a market that does not stand still is an equally important part of the mandate. The two are not in tension, they are complementary.
A bridge holds tension and compression simultaneously in every element — remove either and the structure fails. The discipline is in holding both simultaneously: not alternating between guardian and enabler, but being both at once, and knowing only where to place the weight. When that balance is genuinely struck, something changes in the relationship with the business.
Legal is no longer a function the business comes to because it has to. It becomes one the business chooses to involve early, because experience has shown that doing so produces better outcomes — not just safer ones. The difference between a checkpoint and a trusted partner.
Reaching that point requires the legal function to concentrate its energy on what genuinely demands human judgement — and that is precisely where GenAl plays its part, not by replacing that judgement but by clearing the ground for it. A function that has embraced that discipline is one better placed to hold guardianship and enablement together, rather than trading one off against the other. That reliability, demonstrated through the quality of judgement exercised day after day in the grey areas where blueprints run out, builds something more durable than any architecture: it builds trust.
