Our lawyers used Copilot to help identify recurring themes in our email interactions with counterparties and six things they kept bumping into with counterparties emerged.
We agree with many of them, but not all, entirely, but read together they describe one change rather than six. We then reviewed those patterns: where negotiations get stuck, what gives rise to the demands, and what could help move things forward.
What people negotiate now is mostly operational. Not what the service costs, but how it runs, who actually runs it, what happens when it breaks, and who gets to check. We think that reads less a commercial agreement and more like an operating manual.
Here’s what we found across the contracting lifecycle.
The old triage question was simple enough. Is this a technology purchase, or is this outsourcing? One goes down the fast route while the other gets the full treatment.
That line has blurred. Suppliers now deliver managed, hosted and AI-enabled services that look like software on the invoice and behave like outsourcing in practice. Buyers have noticed, and outsourcing-style controls are turning up on deals that would have gone through as a straightforward licence two years ago.
This means legal gets pulled in earlier, to look at regulatory, operational and outsourcing risk before the commercial terms are settled. Most lawyers have spent years asking to be involved sooner, so this is good news, arriving in the least convenient form available: more deals, earlier, with less information attached.
The fix sits at intake, not in legal. If your intake form still asks what you are buying, it is asking the wrong question. Ask what the supplier will actually do, what data they will touch, what breaks if they stop, and who they are using underneath.
Those four answers route a deal far better than the category does.
Standardisation has not failed, it has moved.
Global templates keep colliding with local regulatory requirements, and procurement teams are consistently surprised by how much bespoke drafting that forces. Even with a perfectly good global framework in place, local law, tax, data protection and regulatory rules drive real deviations, every time, in precisely the places the framework was meant to prevent them.
The response we are seeing is a sensible one. Hold the standard centrally, then flex locally through country schedules and call-off structures instead of reopening the master agreement in every jurisdiction. Ergo - one spine, local vertebrae.
It works, with one condition - someone has to own the spine. The weak point is a central template that, for argument’s sake, nobody has touched since the last reorganisation, surrounded by fourteen local schedules that contradict it.
That is worse than having no standard at all, because everybody assumes there is one.
Price still matters, but it is rarely the thing holding deals up.
Risk Allocation: Liability, indemnities, audit rights, compliance obligations. Who carries what when it goes wrong, and on what terms.
Intellectual property: Technology suppliers have got noticeably sharper here. The old two-way split of customer deliverables and supplier background IP is now a three-way split, with AI-generated outputs in the middle. Who owns the output when a tool produced it from your data and the supplier’s model? Can the supplier use what you put in to improve it? The second question is usually the real argument, appearing as an IP point.
Operational resilience: Business continuity, incident management, testing rights, subcontractor oversight. These used to live in regulated outsourcing agreements. They now appear across ordinary supplier contracts, because an outage does the same damage whether or not a regulator has an opinion about it.
Subcontracting: Customers want to see who sits behind the supplier, particularly where cloud or technology services are involved. They want notice when it changes, and sometimes a say in it.
The above are four topics with one fundamental underlying question: can I rely on this, and can I show somebody else that I can?
If a counterparty is pushing hard on audit rights, they rarely want to audit you - they want to be able to answer their own board. Give them a credible evidence route and the clause usually comes back to something sensible.
It is also why a mark-up on its own has stopped being much use. Anyone can flag that a clause is off-standard. The useful output is a view: here is what they are actually worried about, here is where we hold, here is what we trade, and here are the two things that need escalating before anyone agrees to them.
That judgement comes from volume.
A team seeing the same arguments across dozens of deals knows which positions counterparties always concede, which are genuinely immovable, and which only look immovable until somebody asks why. A review that comes back with forty comments and no recommendation has simply moved the work, not actually done it.
Procurement cycles are getting longer at the same time as everyone is being told to close faster.
Nobody is ever really being slow on purpose. More scrutiny of information security, resilience and data use means more approvers, each with a good reason to look, and together they add weeks.
It’s easy to forget that the other side has the same queue. Much of the delay you may beputting down to a difficult counterparty is their internal process, not their position.
Helping them through it is the fastest route to a signed contract.
E-signatures sorted out the last mile years ago. The delay lies in the approval chain in front of it, which makes one number worth measuring:
how long does a contract spend waiting for somebody who is not a lawyer?
Testing rights, incident reporting, subcontractor change notices, audit windows and exit plans. Every one of these creates work after signature.
Somebody has to run the test, read the notice - keep the register, chase the evidence.
If nobody owns that work, the resilience exists only in the drafting.
So the question during negotiation is not only whether a clause is acceptable, but who is going to do this and whether they will know they have to. A testing right nobody exercises is something you paid for in concession and never collected.
The same logic applies back up the chain. Every local carve-out is a piece of data telling you where your standard is wrong. If local counsel has made the same amendment in Germany four times, that is a template update nobody has done yet. Track the deviations, look at them quarterly, and feed the recurring ones back into the spine.
The pattern is invisible from inside a single negotiation and obvious from above it, which means somebody has to be reading across the deals rather than closing them one at a time.
When the argument was about price, the other side was a counterparty in the simple sense. One of you was going to end up with the money. Now that it is about resilience, subcontractors, testing and audit, both sides are negotiating obligations they will both live with for three or five years.
A clause you win by grinding somebody down is a clause their delivery team resents on day one and under-delivers on by month six.
Being reasonable is not the same as being soft.
It means knowing which points actually protect you, holding those firmly, and not spending goodwill on the ones that only look like wins. It is quicker, too, because the other side stops bracing for the next unreasonable demand and starts helping you close.
It's worth remembering - the market is smaller than it looks. The person across the table this quarter may in two years be sitting on your side of a deal, or recommending who to call. People remember who was straightforward to deal with, and they tend to mention it.
This is fundamentally what inspired us to write this piece - to help legal teams manage counterparty issues more effectively so that contracting works better - for both sides of the table, ideally.
You might be wondering what your next steps should be. Let us guide you with three easy options:
We use cookies to monitor the performance of our website and to improve user experience. By continuing to browse our site, you agree to our use of cookies.
