ModulesLabQuizAll courses

Workflow Mapping II: Designing the Future State

Decide, step by step, what a machine does, what a model judges and what a person keeps, then rank the candidates by value and risk and pick the first one to build.

Module 4 of 14 · Day 2 · Session 4 · AI Administrator: Agentic Workflows & Automation

Module 04 ~45 min read + lab No code

What you will learn

Prerequisites: Module 3: Workflow Mapping I and your confirmed current-state map.

1. From current state to future state

The future-state map is the same swimlane, redrawn as it will run once the automation exists. Usually a new lane appears, labelled Automation or Agent, and steps migrate into it from the human lanes. What stays in the human lanes is the point of the exercise.

Three rules keep the redesign honest:

The output of this module

A future-state map, a scoring sheet, and one chosen automation with its approval gate marked. Modules 5 and 6 build it.

2. The three-way split: rule, model, person

For every step in the map, ask one question: could I write the rule?

CategoryTestExamplesBuilt with
Rule-basedYou can write the complete rule in a sentence with no "usually"Copy form fields to a sheet; if amount > 500, notify finance; send reminder after 2 daysPlain automation nodes (Module 5)
AI-judgedA trained person does it in seconds from reading, but the rule has exceptions you cannot listClassify a request; extract the dates from an email; draft a first reply; summarise a threadAn AI step inside the automation (Module 6)
Human-onlySomeone must be accountable, or the consequence is hard to reverseApprove spend; decline a candidate; commit to a delivery date; anything with personal or legal weightAn approval gate or a task assigned to a person (Module 10)

Most steps tagged READ and DRAFT in Module 3 become AI-judged. Most REKEY, WAIT and CHASE steps become rule-based. JUDGE stays human. Where you hesitate, choose the more conservative category; you can promote a step later once the automation has a track record.

AI-judged is not AI-decided

An AI step proposes: a category, an extracted value, a draft. Whether that proposal is acted on automatically depends on the risk score in the next section. Low-risk proposals can flow straight through; higher-risk ones stop at a person.

3. Scoring value against risk

Every candidate gets two scores. Keep them simple enough to defend in a meeting.

Value: hours saved per month

Value = minutes saved per occurrence × occurrences per month ÷ 60. Use the touch times from your current-state map. Count only the minutes the automation actually removes; if a person still reviews the draft, subtract the review time.

Risk: four questions, each scored 0 to 3

Question03
Reversibility — if it goes wrong, how hard is it to undo?Internal draft, easily correctedSent to a customer, paid out, deleted
Money — what is the largest amount involved?NoneAbove the manager's own approval limit
People — does it affect someone's job, pay, access or wellbeing?NoDirectly and significantly
Legal and data — contracts, regulation, personal data?NonePersonal data or a regulatory duty

Total risk runs from 0 to 12. A working rule of thumb: 0 to 3 may run fully automatically; 4 to 7 runs automatically but a person is notified and can undo; 8 to 12 stops at an approval gate every time. Module 10 refines this into proper risk tiers.

4. The prioritisation matrix and a first ROI

Plot each candidate with value on the vertical axis and risk on the horizontal. Four regions, four decisions.

High value, low risk

Build first. These are the quick wins that pay for the course. Typical: rekeying, chasing, routing.

High value, high risk

Build with a gate. Automate the preparation, keep the decision human. Typical: refunds, offers, contract terms.

Low value, low risk

Batch later. Cheap but not worth a project on its own; add them once the platform exists.

Low value, high risk

Leave alone. Improve the human step with a better summary or checklist instead.

A one-line ROI

Monthly hours saved × loaded hourly cost, minus monthly tool cost and a maintenance allowance (one hour per automation per month is a fair start). Present it as a range, not a number, and state the assumptions next to it. Managers trust "8 to 12 hours a month, assuming 40 requests" more than "11.3 hours".

CandidateValue (h/month)Risk (0-12)Decision
Copy leave form to calendar and sheet61Build first, fully automatic
Classify and route support emails103Build first; "unsure" bucket to a person
Draft refund replies58Build with an approval gate
Auto-approve leave under 2 days26Later, notify manager with undo
Decline job applicants311Leave alone; give recruiter a summary

5. Placing the approval gate

In the future-state map, an approval gate is a diamond in a human lane whose "yes" arrow continues the automation and whose "no" arrow returns the item with a reason. Three placement rules:

  1. As late as possible, before the irreversible step. Let the automation do all the preparation, then stop just before sending, paying or deleting. The approver sees a finished proposal, not a half-done task.
  2. One gate per irreversible action. Two gates in a row is a RUBBER stamp waiting to happen.
  3. The gate carries the evidence. The approver must see the input, what the AI concluded and why, and a one-click approve, edit or reject. If they need to open three systems to decide, they will approve without looking.

Draw the gate now, even if the first version of the automation stops at a notification. Module 10 turns it into a working approval step; Module 11 adds the audit trail behind it.

Practical lab

You will redraw your Module 3 map as a future state, score every candidate step, and choose the first automation. You need your draw.io file and a spreadsheet (Google Sheets or Excel).

1

Classify each step

Duplicate the current-state page in draw.io. Colour each step: green for rule-based, amber for AI-judged, red for human-only. Write one sentence per amber or red step explaining why it is not green.

2

Remove before automating

Delete any RUBBER approvals and any REKEY step that disappears once two systems are connected. Note what you removed in a text box on the page.

3

Add the Automation lane

Insert a new lane called Automation. Move the green and amber steps into it, keeping their order. What remains in human lanes should be red steps and the gate.

4

Build the scoring sheet

Create M4-scoring with these columns and one row per candidate (a candidate is one or more adjacent green/amber steps that would form a single automation):

Candidate | Steps included | Minutes saved per run | Runs per month | Value (hours/month) | Reversibility 0-3 | Money 0-3 | People 0-3 | Legal/data 0-3 | Risk total | Quadrant | Decision
5

Score with a colleague

Fill the sheet with the colleague from Module 3. Where you disagree on a risk score, take the higher one and write why in a Notes column.

6

Plot and choose

Sort by quadrant. Pick one high value, low risk candidate as your first build. If none exists, pick the highest-value candidate with risk 7 or below and plan the gate.

7

Mark the gate and write the ROI

On the future-state map, place the approval gate diamond immediately before the first irreversible action of your chosen candidate. In the sheet, add an ROI line as a range with its assumptions.

8

Stress-test with your assistant

Paste the scoring sheet into your assistant:

Here is a scoring sheet of automation candidates for one business process. For each row, challenge the risk score: name one failure that would make it worse than scored. Then tell me which single candidate you would build first and why, in under 100 words.

Deliverable

The future-state map (with Automation lane, colours and the gate), the M4-scoring sheet with ROI range, and one sentence naming the first build and its approval rule.

Knowledge check

Pick one answer per question, then check your score. These mirror the style of the final exam.

1. A step should be classed as rule-based when...

Why: Rule-based means the logic is fully specifiable. Anything with 'usually' or unlisted exceptions is AI-judged or human-only.

2. What should happen to an approval that is always given without reading (a RUBBER step)?

Why: Automating a meaningless approval preserves the waste. Remove it, or keep only the exception cases for a person.

3. Which factor is NOT one of the four risk questions in this module?

Why: Risk is scored on reversibility, money, people and legal/data. Popularity may affect adoption but not risk.

4. Where should an approval gate be placed?

Why: Late placement lets the automation prepare a complete proposal and stops exactly before the action that cannot be undone.

5. A candidate saves ten hours a month and scores 9 on risk. The right decision is...

Why: High value, high risk means automate the preparation and keep the decision human. That is the 'build with a gate' quadrant.

Self-check

Answer in your own words first, then open the model answer.

1. Explain 'AI-judged is not AI-decided' in your own words.

An AI step produces a proposal such as a category or a draft. Whether that proposal is acted on without a person depends on the risk score. Low risk flows through; higher risk stops at a notification with undo or at an approval gate.

2. Why present ROI as a range with assumptions?

The inputs (minutes saved, runs per month, review time) are estimates. A range with stated assumptions is honest, survives challenge, and can be checked against real numbers in Module 12.

3. Your chosen first build has no irreversible step. Do you still draw a gate?

Draw at least a notification point where a person can see what ran and undo it. If truly nothing can go wrong, note that explicitly; the decision to have no gate should be a decision, not an omission.

Summary

Key takeaways

  • The future state is the same map with an Automation lane; what stays in human lanes is the point.
  • Split every step: rule-based (write the rule), AI-judged (a person reads and decides in seconds), human-only (accountability or irreversibility).
  • Score value in hours per month and risk 0 to 12 on reversibility, money, people, legal/data.
  • Build high value and low risk first; high value and high risk with a gate; leave low value and high risk alone.
  • Place the gate as late as possible before the irreversible step, once per action, carrying the evidence.

Further reading: n8n: workflow design patterns · Estimation (Scrum course, Module 6)