The Problem

‍

The Global Payroll space is full of inefficiencies, manual processes and inaccuracies. We have experienced this first hand, at large global Payroll players and EOR providers alike. This is driven by the reality of global payroll: Each country, and often states and jurisdictions in a country, define their own rules. How is a HCM player or an EOR provider that needs to help customers with a presence in multiple countries to manage this complexity?

The current answer is to manage a large number of 3rd parties, each of which manage payroll for a given country (and sometimes, a region in a country). Running payroll means endless back and forths with all the 3rd parties, requiring a limitless supply of spreadsheets and emails. Lost in the noise is accuracy, timeliness and the abillity to service customers. Estimates are that 21% of employees complain that they are paid either inaccurately or late: even worse, the problem seems to be getting worse.

‍

‍

The rest of the HCM has progressed on to much great adoption of technology: Why is Global Payroll lagging?

‍

Managing the complexity of payroll

Payroll is just complex. The US has 10,0000+ Tax Jurisdictions. Each country has its own nuances. Conventional wisdom in payroll has been that 'local payroll always wins' - and that is true thus far. Each country has a handful (or smaller) number of dominant payroll providers, and all global companies must work with them. How is an HCM company or an EOR company going to tackle the complexity of payroll, with all its myriad regulations - and the changes that happen to these regulations?

‍

Intermezzo's Approach

At Intermezzo, we understand why Global payroll has been such a challenge. We also believe it is time to build a single, global API that can serve the needs of customers by servicing their payroll needs in multiple countries, all with this single API. Our leading edge platform - unique in its construction - enables us to quickly assimilate new countries, all while keeping the API consistent with both customer needs and changing country regulations. How do we do this? A quick meeting is in order :)

‍

The judicious use of AI

AI (conflated with LLM's nowadays) is all the rage in the past year, and it certainly has a significant role to play in our platform. It provides a significant lift to our ability to quickly ingest content and convert them into systemic rules. We use an Agentic AI system with human feedback, along with various related techniques. But AI is delegated to the offiline processing - payroll construction is done using other leading edge techniques. Computational Graphs play a big role in our technology stack in particular.

 Payroll systems are deterministic. LLM's are probabilistic. Payroll computation is not ideal to do in a probabalistic system, where each iteration may result in a different outcome.

‍

Simplifying support

A vexing problem in Global payroll is support. If an employee asks 'hey why have you deducted X amount from my paycheck' - the current resultion for a provider is to turn around to their local provider, who will then examine their system (spreadsheets, typically), send an email back, and two weeks later, the unhappy customer has an answer - which might not even suffice.

‍

We are proud of how we handle customer support - our system allows accurate and specific answer to customers , germane to their situation, using our (patent pending) support copilot technology. Ask us about it! We are an email away - mailto:hello@intermezzo.ai

‍

‍