Over 80% of translation agencies are struggling to keep project margins. They serve more small projects, more language combinations, different file formats while keeping even tighter deadlines with a lower budget per unit. Can they still be profitable? How should they adapt, what changes should they introduce to become more effective?
Process automation is definitely the answer. Repeatable management related tasks have to be automated, so project managers can concentrate on what builds real value for customers rather than project file copying or PO and invoice formatting. Although the answer is simple, the implementation remains a challenge.
Process Automation Today
One may expect automation implementation to be a set of a few steps. You buy a system, define automations, implement it and the magic works. The reality happens to be a little bit more complex…
Regardless of the starting point, many companies usually follow the same pattern: find the automation area, consult the people involved, define the automations, persuade the stakeholders, implement, run and maintain it. These steps are repeated for subsequent automated areas of activity. One may start with automatic PO sending, the other with invoicing, payment reminders or project delivery – it does not matter, the schema used is the same. An automaton is introduced in the ‘entire company scope’ and expected to bring ‘company scope’ effects.
What companies actually do is they are looking for significant savings – so the entire automation implementation process pays off. That is a very good approach and many companies succeed in achieving it. The question is if we can go even further and whether we are not missing an important piece that could help to automate areas currently marked as ‘does not make sense/too complex to automate’.
Machine Project Management – the Vision
A project manager actually repeats many tasks. Let’s imagine a system that gently asks – “Dominik, aren’t you repeating yourself? You always modify the deadline from Friday evening to Monday morning while working with the translator named Smith. Should this logic be applied now and in future projects?” or “You are selecting Peter Smith for the verification step while working for customer ACME, should I send a job availability request to Peter Smith automatically for each new ACME project?” or “You are adding a surcharge of 20% to projects ordered on Friday afternoon with a deadline for Monday morning – should I add such a surcharge automatically now and in future projects?” etc.
The above examples of artificial intelligence are based on an analysis of previous behavior patterns. Each project – from its creation to final payment – can be described as a sequence of events. Events describe what happens as well as when and who did it – so they build up the entire history of project execution. Artificial intelligence algorithms can detect similarities in PM behavior (same sequence of events) and suggest automation in such situations.
The benefits of this approach include:
- less/no upfront automatization configuration required,
- shorter learning cycle for new PMs and users,
- digitalization of people’s operational knowledge,
- much more automation implemented,
- lower cost, faster implementation,
- entire organization committed to automation definition from day 0
- very low resistance from the organization,
- and finally automations are not statically defined but can be,
- adapted to actual user requirements that usually change over time.
We named the above approach Machine Project Management per the analogy to machine translation as automation rules are automatically discovered by analyzing historical project execution .
Machine Project Management – Give It to Me
What is needed to bring this vision to life? We need … events. The events that trace each project execution, the events that are lost when you use traditional TMS systems…
Traditional Technology Limits
Traditional project management systems are mostly focused on data processing. These systems efficiently collect data in central repositories (databases) enabling the collaboration of many people. They in no way interpret or understand the data processed, reducing all executed operations to the basic four i.e. create, read, update and delete (CRUD). More advanced solutions collect historical data stored exclusively for audit purposes, although they cannot answer why and by what business operation the data has been changed.
Modern System Are Based on Events
New generation systems process events and perform semantically rich actions in place of the simple CRUD operations. Based on an analysis of these actions not only can the current system (or project) state be calculated, but also one can analyze the chain of events and operations that led to this state. Besides the direct advantage of being able to restore the system to any state in the past, a rich base of users’ behavior is being collected.
Smart Technology Choice
The appropriate volume of events and actions combined with artificial intelligence algorithms can lead to the creation of autonomous management systems, implementing the idea of Machine Project Management. The approach of MPM will be similar to that of machine translation systems. After an initial phase of learning organization specific processes and operations, the system increases its autonomy, first suggesting and later only notifying the user of the action taken – increasing PMs’ efficiency.
The above is determined by the number of historical projects that can be analyzed. So if you are to change your TMS system choose one entirely based on events.
Regardless of the system you choose, please always support the culture of automation – thank you!