“Organisations are moving away from the strict rule-bound organisational structure, to one that can turn on a dime and react to a rapidly changing environment.”
– Lyle Cooper
The origin of the organisation as a human endeavour is murky. But we do know that the modern organisation finds its birth in the late 18th and early 19th century off the back of the industrial revolution. The modern descendants of these enterprises have changed significantly from their ancestors. The impact of new management approaches such as Agile, DevOps, Daily Stand-ups and project-driven organisations has resulted in the slow (and sometimes not so slow) dismantling of the once familiar world of clearly structured organisational hierarchies and reporting lines, as well as the step-by-step business processes.
Organisational hierarchies have a direct impact on how work is executed. To fully appreciate quite how far we have come, it is helpful to take a look at the history of the organisational hierarchy. The industrial era borrowed from the military doctrines of the time to structure organisations that capitalised on the dominant paradigm and technological advances.
This mechanistic approach focussed on the logical structuring of the workforce into layers, with each department being led by a single manager. This approach was appropriate when the nature of the work and tasks to be executed lent themselves to segmentation (as was the case with the motor car production line, or the accounting firm).
But the world of work has moved on. Competitive advantage was no longer a function of the size of an organisation, but a function of its agility. The efficacy of manager-led teams started to wane. Less structured organisations began to appear and while these matrix-based organisations still had managers, employees also had to contend with other people who wanted to have influence over their daily tasks (project managers, operational managers, etc.).
Organisations are moving away from the strict rule-bound organisational structure, to one that can turn on a dime and react to a rapidly changing environment. In a similar way, the regimented process is starting to be eroded. Processes are becoming shorter, less hierarchical and more representative of the dynamic landscape they take place in.
“It seems that we have moved from a strict regiment based structure to small commando based teams that take on tasks in a fluid manner.”
This shift has continued to deconstruct the hierarchy, with many organisations moving towards pure team-based structures some even drive the concept of self-organising teams that coalesce around a task or project and then disperse when the work is done.
A person’s role in any of these teams may vary the leader on one team might be the analyst on another, depending on the individual’s ability to contribute to the task at hand. Sticking with the military analogy, it seems that we have moved from a strict regiment based structure to small commando based teams that take on tasks in a fluid manner.
It’s no secret, computers like structure;
Computers are binary, they like things to be 1 or 0 they are not keen on fluidity or unpredictability. Computers need things to fit into neat “if-then” statements. We can construct these simple binary choices into incredibly complex applications that mimic many chaotic functions (though they still struggle with true chaos.
Most Human Capital Management softwares have not yet reached this level of complexity, and it is still constrained by the structured regimental thinking of the past. From a workforce structure and process design perspective, most systems require you to identify who reports to whom (computers need a basic one-to-one or one-to-many relationship that remains fairly static).
Systems then use these reporting structures to manage workflows and financial rollups (along with a myriad of other tasks). This approach to structuring employee information aligns well with traditional HR and Finance functions, which need to know how to group people or costs together.
“HR cannot escape the reality that most HR processes require some form of an approval for the request at hand.”
The reality of the world of work is decidedly unstructured;
This approach was perfectly functional when this very structured approach to employee hierarchies matched the realities on the ground in the organisation.
But the reality on the ground is changing rapidly: In many recent engagements, my clients have raised the fact that this single-reporting-line structure is too rigid for the realities of their operations. For the majority of them, some or all of the employees permanently work in teams or on projects.
These teams often stretch across functions. To complicate the matter further, individuals will work on multiple projects, while their titular manager might not have any day-to-day contact with them at all. Finance is usually less impacted by this, as cost centres and project codes can be used to track costs and profit created, but for HR, this creates a nightmare.
HR cannot escape the reality that most HR processes require some form of an approval for the request at hand (either for legal or risk management reasons). Even the lowly leave application requires approval, as it has financial and operational implications for an organisation. This approval needs to be provided by a person in the organisation who is duly authorised to allow this request. But the challenge in the modern organisation is, ‘Who’? If the organisation is made up of shifting teams and projects, how is the request routed to the person who needs to approve it?
Enter the Human Post Box;
In order to get around this complexity, many organisations create roles for a person who acts as a manual re-allocator of processes. Staying with our leave example, all leave requests go to this one person and they then determine who the requester is, whom the requester reports to, which project the requester is on, who manages that project and then forwards the request on to (hopefully) the person who can indeed grant or deny the request.
This entire process is ‘off system’ and based on the persons’ knowledge of the current projects and scheduling in the organisation. This is clunky and inefficient, but it often produces better results than just sending the request to the person noted as the line manager on the HRMS.
Maybe it’s time to let the robot drive;
As we know, humans make mistakes or do not have all the information to make the correct decision. If you think about a fast-paced project-based organisation; the factors that impact that should see and approve a specific transaction could include the project sponsor, the nature of the project, the project plans (as there are dependencies on tasks), the project manager or team leader, and more.
Most agile organisations track this information on a number of systems (not all of them are the ERP), in order to shepherd all these projects towards a strategic goal. The information needed to get a request to the right person is often captured, but it is not always easily consumed or accessed. A further complication is that this information changes quickly, making it impossible to capture on the ERP.
“Given the correct training, a basic AI system could take over from the workflow engine and route work to the people who care about a transaction or approval.”
A further complication is a reality that people fulfil different roles on different projects, and their requests may need to be seen by different people. This is where AI potentially has a huge role to play. Given the correct training, a basic AI system could take over from the workflow engine and route work to the people who care about a transaction or approval.
This ‘process AI’, would be able to consume the data across the various management tools (Projects, Finance, HR, Policies, Compliance, etc.) and use the data points to ensure that the correct person or people are presented with an approval request that actually relates to their current responsibilities for a specific task. This would mean a radical shift in our thinking on processes; from a fairly linear step-by-step view to one that is more rule-bound than a set of steps.
In this scenario, the rules will define how to get approvals, but not specifically who has to approve. The outcomes of the transaction could then be stored in an application of the blockchain technology, containing all the pertinent information about the application and approvals obtained.
This on-the-fly-calculation and routing of work is exactly the type of work that AI is great at and seems to suit the decentralised fluid world we are moving into. Maybe it is time to move over and let the robot drive.