The client
The client is a high-growth international services company. It delivers regulated, compliance-heavy services to thousands of business customers across dozens of countries.
The work has extraordinary process complexity, which is a major factor in achieving a successful AI transformation.
The problem
To run an agentic transformation, you have to understand the work first
The client is strongly AI-first and is deploying agentic automation across its operations and customer support teams.
But before the work is automated, it has to be understood.
The operational complexity is genuinely extreme
The client’s core operations run across several distinct service models, each with different ownership, responsibilities and compliance obligations. Each model then behaves differently in every country: different regulatory requirements, different local tools, different deadlines, different partner networks. Onboarding, delivery, changes, exceptions and reconciliation each fork again underneath that.
Manual SOPs are difficult
The conventional answer is to ask process owners to write it down. At the client’s scale and speed, that becomes a challenge:
- Capacity. The people who know the process are the people working under demanding SLAs.
- Standardization. Written SOPs vary by author.
- Completeness. What is obvious to an expert is not obvious to a new joiner, or to an AI agent.
- Decay. A tool UI changes, a regulation is updated, a partner process shifts. Some SOPs are out of date the day they are written.
What the client tried first
The team experimented with homegrown approaches and found them difficult to operate at scale. It then evaluated the market for a proven solution and selected Gralio.
The solution
For the AI & Knowledge team
We deployed Gralio’s flagship product: a Mac and Windows desktop app that runs alongside a team member’s normal work and captures the context of what they actually do.
From that raw, messy reality, Gralio:
- Identifies discrete processes, separating one job from the next inside an unbroken workday of tab-switching, chat interruptions and half-finished tasks.
- Writes the SOP: a complete, step-by-step document with tools, systems and decision points named, based on observed behaviour rather than recollection.
- Detects variants, recognising that the same process run in two countries under two service models is one process with significant variation.
- Detects and quantifies automation opportunities, recommending tools the client already has in place, including its internal AI agents.
- Surfaces it to managers through a dashboard of every process map generated, with attributed hours, so leaders can see how work is actually distributed and where automation is worth building.
For the people doing the work
SOPs where people already look. Every generated SOP is published into the client’s internal knowledge system. The Gralio interface adds the ability to explore variants flexibly.
Insights. When Gralio observes one team member doing something particularly efficiently (a macro, a script, a shortcut, a better sequence), it surfaces that as concrete advice to others. Not abstract best practice, but “Ask Bob about the macro for this process.” Expertise moves sideways across the organisation.
Real-time help. When someone opens the app while stuck on an edge case, Gralio reads what they have been doing, searches the knowledge base for what is relevant to that situation, and suggests the answer at the point of work.
Results
- 1,000People covered
- 3,000SOPs generated
- 90%Of processes with automation opportunities identified
- ~0Time for a process owner to create a new SOP
- 90%Employee approval rate of Gralio-generated insights