System redesign

We’re living through a strange paradox. We have entered the era of the AI agent, where Large Language Models can draft contracts, code apps and execute complex workflows in minutes. Yet in 2026, professionals still spend Sunday nights buried in ‘shadow work’ – the manual, invisible tasks of rebooking flights, reconciling receipts and hunting for trip itineraries.

Research from Forrester Consulting, commissioned by Perk, found shadow work costs UK businesses £95 billion annually in lost productivity. This isn’t just an administrative annoyance; it’s a failure of a system that is burning out our best people.

The automation gap

In digital transformation, we historically fixed the ‘stable’ processes first. Take payroll, for example, a success story because it deals with linear data. Your salary and tax code rarely change mid-cycle. It’s a clean mathematical equation that software solved a decade ago.

Travel and spend remain the final frontier for automation because the data is ‘unstable’. It is nuanced and in flux. An itinerary may change five times before a trip starts, while a London hotel receipt looks nothing like a Tokyo taxi invoice.

Because this data is messy, the industry settled for fragmented tools, forcing employees to act as a bridge between disconnected systems. We filled the gap with human effort because, until now, AI wasn’t sophisticated enough to handle it.

Beyond standalone tools

For years, the gold standard was a best-of-breed stack: a booking tool, expense software, and a card provider. But standalone tools, no matter how good they are, hit an intelligence ceiling. When tools are built in silos, their data is trapped.

The result is a fragmentation tax. The average traveller navigates four different systems to tackle shadow work, costing them 11 minutes of focus every time they switch tasks. True excellence today isn’t defined by a standalone feature, but by a natively AI-integrated architecture, where every part of the journey – from booking to spend and everything in between – is aware of the other elements in real-time.

Designing for the human journey

Closing the automation gap requires a move away from basic tools like legacy OCR (Optical Character Recognition) towards true system-level intelligence. 

To do that, a modern system must be built on three non-negotiables:

• Intelligence that handles nuance: We don’t need more data capture; we need AI that puts data to work in the right ways to remove shadow work. Our systems, for example, recognise that a flight delay in an itinerary should automatically trigger an adjustment in the expense report and tax reconciliation without a human lifting a finger.

• A single policy engine: Rules shouldn’t be fragmented. A unified architecture allows AI to enforce policy across travel, spend and events at once, ensuring compliance without manual audits.

• Zero-training simplicity: AI should make software invisible. If an employee needs a manual, the design has failed.

Architect of impact

When we get the AI architecture right, the travel manager’s role shifts. You move from being a ‘receipt chaser’ to an architect of ‘real work’. You stop managing the friction of broken processes and start leveraging a system that empowers people to do the impactful work they were actually hired for.

It’s time to stop patching a fragmented system with yesterday’s tech. It’s time to redesign the process with AI at the core and finally close the automation gap.

perk.com