From manual schedules to 95% schedule automation: How Marshmallow rebuilt WFM in 4 months

Marshmallow is a UK insurtech on a mission to make car insurance fairer for people who have often been underserved by traditional insurers, including many customers who are new to the UK. As the business grew, so did the complexity of its support operation, and the pressure on workforce management grew with it.

The challenge

When Lauren Gatt, known as Loz, joined Marshmallow as Senior WFM Manager in August 2025, she inherited a workforce management setup that was not built to scale.

Assembled was already in place, but the team was not using it as a true WFM platform. Key planning inputs like handle times, SLA targets, and shrinkage had never been properly configured. Scheduling was highly manual, ownership was fragmented, and forecasting decisions were difficult to trust because the underlying setup was incomplete.

As Loz put it, the team was using Assembled as a “glorified spreadsheet.”

That gap showed up most clearly in scheduling. Two planners were spending an entire week to publish four weeks of schedules for just 80 FTE. Even then, the process relied heavily on repeating what had worked before, rather than generating schedules from a reliable planning model.

The operational setup made things harder. Marshmallow’s chat team worked set daytime hours, but customer messages continued to come in overnight, creating pressure on service levels before the day had even properly started. WFM ownership was split across teams, institutional knowledge had eroded over time, and the business lacked the foundations needed for confident capacity planning.

Marshmallow did not need a few process tweaks. It needed to rebuild the basics of workforce management.

Why Assembled

Loz had hands-on exposure to Assembled during her time at Monzo, so she already knew what the platform could do when it was used well. When she arrived at Marshmallow, she quickly recognized that the issue was not the tool itself. It was the way the foundation had been set up around it.

That mattered because Marshmallow needed a platform that matched the realities of a fast-moving startup environment. Ease of use was a major factor. So was flexibility. So was native compatibility with the tools Marshmallow already relied on.

In Loz’s words, “It was the ease of use, I think, more than anything. It aligns quite nicely with the startup environment.”

She also had a clear perspective on the broader WFM market. After working across a wide range of workforce management tools, what stood out about Assembled was not just functionality, but usability.

That was more than a design preference. For a lean planning team, interface quality directly affects adoption. Marshmallow needed a platform planners would actually want to work in every day, one that made it easy to build schedules, copy and adjust shifts, and manage change without adding friction.

Integration mattered too. As Loz noted, “NICE, Genesys, they don’t integrate with Intercom.” For a business running customer conversations through Intercom, that was a meaningful differentiator.

Rebuilding for schedule automation

With the right leader in place, Marshmallow and Assembled focused on rebuilding the WFM foundation from the ground up.

That meant reworking queue structure, configuring the core planning inputs that had been missing, centralizing WFM ownership, and replacing the manual scheduling process with AI-powered schedule generation. Instead of treating Assembled like a spreadsheet, Marshmallow began using it as the scheduling and planning platform it was meant to be.

This foundation-first approach is what made schedule automation possible. Once the underlying setup improved, Marshmallow could move from manual effort and guesswork to a much more scalable model.

The transformation

Before the rebuild, two planners were spending an entire week to publish four weeks of schedules for 80 FTE. After implementing schedule generation, that same team was able to support roughly 370 agents while planning up to two months ahead.

As Loz explained, “Technically speaking, it’s a very minimal amount of time that they actually spend on it now, which is super rewarding.”

Schedule automation rose from 0% to 95%. What had once been a manual, spreadsheet-heavy workflow became a far more efficient and scalable process.

The operational impact went beyond speed. Marshmallow now had clearer WFM ownership, better planning inputs, and a stronger structure for forecasting and staffing decisions. Confidence in the platform grew along with its footprint across the business, and that renewed confidence showed up commercially when Marshmallow moved from a 6-month commitment to a 12-month term.

What comes next

Now that the core WFM foundation is in place, Marshmallow is using it to support a broader and more complex operation.

The business is balancing human support with growing AI usage in parts of the operation, which makes workforce planning even more strategic. WFM is not just about filling shifts. It is helping Marshmallow think through where automation should expand, where human expertise matters most, and how to scale without overloading the team.

For Marshmallow, reaching 95% schedule automation was not just an efficiency win. It was proof that with the right foundations in place, workforce management could finally keep pace with the business and help it scale.