The Watertm Workforce Transformation Platform offers a clear Roadmap to Automation at scale

Automating work begins with identifying and prioritizing repetitive Action Items that Sidekicks can potentially perform autonomously.

The Steps of AI Agentic Transformation

1. 1. Make a list of your functional departments like Recruiting, Sales, Accounting, etc.
   2. Work with the supervisors of those departments to decompose repetitive projects into lists of tasks.
   3. Decompose those tasks into a list of microtasks, which we refer to as Action Items. An Action Item is a small task assigned to just one person.
   4. Analyze that list of Action Items and identify those that can potentially be automated today, given the current state of technology. We’ll call this the “A-List”.
   5. Prioritize the A-List Action Items for development based on criteria such as difficulty of development, cost of development, speed of development, expected repetition of use & expected ROI.
   6. Develop the automations in order of priority. For convenience, we’ll often refer to automations as ‘Agents’, however; automations might fall under a number of categories including:
      • • Repetitive Process Automations (RPAs); programmatic IF/THEN workflows, such as those developed with N8N, Make.com or Zapier.
        • MCP Servers, which are narrow tools or services.
        • Actual Agents, which are capable of thinking, reasoning and making human-like, qualitative decisions.
   7. Add the automations to Water as new ‘skills’
   8. Test the skills, refining the instructions given to the Sidekick until you are getting repeatable results.
   9. Selectively authorize Sidekicks to use those skills
   10. Create or update template tasks, indicating which Action Items should be performed by a Sidekick. Generally no employee training is required.
   11. Instruct the team; when starting a new project, to copy the relevant template project (which contains the template tasks).
   12. Measure outcomes. As the team utilizes the new automated workflows, data will flow immediately, including; estimated hours of labor saved, estimated delivery time saved (production is much faster) and the cost of operating the Sidekick’s.  ROI and proof of value can be measured from the very first day. These reports are definable by Region, Division, Department, Team and User.
   13. Start back at Step 6 and develop more automations!

Many organizations struggle to identify viable automation opportunities that come with the magic combination of rapid deployment, low risk, low implementation costs, and high ROI.

Below we reveal methods not just for identifying use cases, but also for prioritization based on the magic criteria mentioned above.

Identifying Use Case Candidates

A common-sense principle of automation is to keep the use case narrow. Narrow is easy. Narrow is reliable. Broad is not!

The Water platform provides an interface where existing workflows and tasks can be analyzed and documented into narrow bits and pieces, each of which become candidates for automation. Let’s refer to this as Task Decomposition. If an organization has existing SOP documentation, this decomposition process can be performed by a consultant or a specialist on the team.

Inside Water, the smallest component of a workflow is referred to as an Action Item, which consists of a single micro-task, to be assigned to one person (or Sidekick). It is these narrow Action Items that be the focus of your candidate evaluation.

Documenting the various components of a larger workflow in the smallest possible components makes it surprisingly easy to identify use cases for AI.

Example of Task Decomposition – Logo Development

Action Item #1 – Plan and document scope of task

Action Item #2 – Gather brand guidelines

Action Item #3 – Generate draft logos

Action Item #4 – Send logos to stakeholders for review

Action Item #5 – Make revisions

Action Item #6 – Get final approvals

Action Item #7 – Organize final files and wrap up task

Next, the transformation leader must review each Action Item, asking themselves “Can this be automated today, with the existing state of technologies, or must this wait for future developments?”

At the time of writing this article in September 2025, we see that Action Items #2, 3, 4, 6 & 7 are relatively easy candidates for automation. 

Those that can be automated today go on a list for evaluation, which will be referred to during assessment and evaluation, using some method of prioritization, such as those methods proposed below.

Eventually, all Action Items in the world might be automated. This might be ten years from now or perhaps fifty. The timing is uncertain, but we are all on that trajectory. So even if only one in ten microtasks can be automated today, eventually that will become two in ten, then three in ten, and so on.

There is no need to struggle with solving difficult automation challenges today, because there will be plenty of easy ones that reveal themselves.

Prioritization Methods

Below you’ll learn about planning methods made easier by using reliable models for prioritization. One method is focused on pre-development planning, the other on post-development planning (i.e. optimization).

The FVC Prioritization Method

This is a pre-development prioritization process, focused on the value of automating vs the cost of automating.

The basic formula is frequency*value/cost. Here is an example:

In this example, automating the task ‘Create daily post’ is clearly identified as the highest priority. 

Note: To make the score more readable, you can modify the basic formula so the score is a product of 1000 or even higher, like this: (f*v)/c*1000

Standardize on units of frequency and value

In the above example, we standardized on a monthly frequency, and a value measured in hours. You may alternatively standardize on a different frequency unit (weekly, daily, etc.), and the unit of value in dollars or per incident, or any other method of valuing the task. The formula f*v/c does not change.

Probability of Failure?

If you believe there is a risk of your future automation process failing due to unforeseen obstacles, such as reliability issues, then the formula can be adjusted to (f*v)/c*p where p=probability of failure.

Including a probability of failure can paint a different prioritization picture.

You may alternatively add a ‘time’ factor, where
t=time to deploy the new automation

However; for simplicity, cost can be a proxy for time, which may be sufficient for most cases.

Getting Task Frequency Data

The FVC method of prioritization requires the identification of tasks or microtasks performed repeatedly in an organization. 

Water uses historic data to provide frequency of every task and microtask across an organization. In the simplest of terms, if multiple tasks have the same name, they are grouped. However; many tasks, while similar in nature, have names which humans have chosen to customize, not always for good reason, but it happens. Water’s algorithm performs vector analysis so grouping take place, even if there are dissimilar task names.

The algorithm goes even further by identifying microtasks using the Action Item structure, which is built into water.

This data is available after an organization has been using Water for a minimum of two weeks. Water uses a familiar architecture to structure workflows in a format that is intuitive for humans, and highly readable by AI.

• Organizations – Contains a collection of related projects (such as those belonging to a client, customer, division, region, ministry or department)
• Projects – Contains a collection of related tasks
• Tasks – Contains a collection of related Action Items
• Action Items – The smallest granular event performed by a single user

Water allows dozens or even hundreds of automation opportunities to be swiftly identified, then narrowed down to those repetitive micro-tasks offering the quickest and largest payoffs.

The Adoption Gap Method

As the organization goes through the transformation process, you’ll get insights into adoption and its impact on productivity at every organizational layer, and for every user. 

Adoption-gap reports reveal which user, team, department or division is utilizing AI to complete work. The ‘gap’ reveals who is lagging behind.

The gap may be a result of suitable AI tools not being available yet. Or it can be a result of low adoption. In either case, the data can be used to reset priorities on future development and training.

Water provides transformation leaders insights into metrics that make planning the AI digital transformation process much faster, with more predictable outcomes.

Technology Decisions

Sidekicks can control agents and workflows developed on any platform, including A2A, MCP or even workflow tools such as N8N or Zapier.

These agents and workflows will be interacting with your data so consider how you will lock it down. Within Water you can designate which of your people will get access, so on this side it is secure.

Securely Roll Out your Agents to the Team

Inspira’s Water Platform has enterprise-grade security, allowing you make your agents available to select team members, while restricting others.

Agent’s are added through a conventent form, including authentication requirements. These newly registered agents become ‘skills’ that can be selectively assigned to one or more Sidekicks.

Prompts that will reliably invoke these new skills can be optimized and stored in Water’s built in prompt libary.