By now you’ll have heard most of it. T+1 arrives in Europe and the UK on 11 October 2027. It will cut counterparty risk, but will be highly unforgiving of anyone whose reference data isn’t clean. It will be harder here than it was in North America, because Europe brings more currencies, more venues, more settlement locations and more tax regimes into the same two-hour window. All true and all repeated at every conference this year.

The directive?  Automate, automate, automate.  But what does that even mean?

Orchestration, not automation, is the real challenge

You can automate all the processes you want and in reality, most operations teams already have, so there’s not a lot of fruit left on those trees – let alone the low hanging stuff. What’s not already automated is often classed as ‘too complex’.

But even successfully automated reconciliation processes have exceptions, which have to be handled by humans. These humans get into the office at 8am, and then need to review and classify exceptions, identify fixes, shepherd to teams, confirm updates and re-run.

This isn’t automation, it’s orchestration, which requires more than just what is known as a ‘cron job’, which is simple time-based task scheduling.

Cron jobs can run overnight, but in order for firms to gain real efficiency when it comes to exceptions and have any hope of hitting T+1 timelines, they need tools that can assess breaks, identify root causes, filter out the noise, fix what can be fixed easily, suggest next steps and have it all served up on a platter by the time a human arrives in the office at 8am.

Without these tools, firms face a potential 18% bump in headcount costs. When volatility spikes trading volumes or bad files create a big spike in exceptions, firms that only have simple automation in place will need more humans in order to handle this, or risk face rising settlement fails. This is true now, let alone under T+1 which compresses significantly the time that firms have to fix issues.

This is possibly the best lesson we can take from the North American transition to T+1.  The UK and Europe move is going to be substantially more complicated because of the variability. In fact, an 18% increase in costs should be looked at as the floor for this region, rather than the ceiling or a likely average.

Scalability through agentic orchestration

This is why firms need to start thinking in terms of orchestration, and not purely automation. In order to solve the spike problem. Imagine a bad file comes in and the automated process runs after close of business.  This process usually has a break rate of 7%, but with a bad file it is now 27%. A solution based purely on automation does nothing here because it can only work with good data.

An agentic solution focused on orchestration, however, can take next steps, by analysing the cause of the breaks and identifying the correct counterparty or team to notify.

This does two things.  Firstly, it moves the work up the timeline so that the evaluation of the problem can happen at 11pm by an agent, not 8am by a human.  Then, notification to counterparties can happen as soon as the team validates the problem in the morning, saving hours and accelerating resolution.

Secondly, agents enable teams to handle increased volumes without additional people. With agents, operations teams don’t need to spend hours wading through the data to figure out the problem, they are able to take action, quickly.

Are we ready for agentic orchestration?

Realistically? The jury is still out. Cambridge’s Centre for Alternative Finance found that 52% of firms are deploying agentic AI in 2026, with 23% at deploying it at scale. However, the Bank of England and FCA found that only 2% of AI use cases are fully autonomous.

No independent research yet quantifies what AI agents can do to break volumes or touch time in post-trade, so treat any vendor number with care.

Despite this, the shape of the problem is clear. Zero exceptions is an admirable goal but highly unrealistic, so firms need to go beyond automating post-trade activities like allocation and address the question of who, or what, is in charge when something doesn’t match up at 22:00 on trade date, and how fast they can act.

Firms are already starting to leverage 24/7 AI agents here that can work right across a firm’s exception management workflows and, in some cases, solve an exception. This reduces the risk of activity spikes, enabling firms to handle exceptions successfully at scale.

T+1 will not be the end point either. The work that firms do now to build smarter post-trade operations will also put them in a much stronger position for any future move towards T0.