2. Use case definition
A good AI use case is not simply “use AI for X”. It is a workflow improvement with a named owner, measurable outcome and safe control model.
Assessment context *
New idea
Existing workflow improvement
Prototype candidate
Pilot already in progress
Scale candidate
Unsure / early exploration
Guidance: Choose where this opportunity currently sits: idea, existing workflow improvement, prototype, pilot, scale candidate or early exploration.Format: Select one option. Choose Unsure / early exploration only if the use case is genuinely not yet formed.
Priority zone *
Labor Bottlenecks
Operational Efficiency
Knowledge Advantage
Compliance / Risk
Competition / Market
Customer Experience
Revenue Growth
Cost Reduction
Quality / Consistency
Onboarding / Enablement
Executive Decision Support
Other
Guidance: Select the main business value driver. Choose the strongest driver, not every possible benefit.Format: Select one option.
Use case category / solution pattern signal *
Select…
Document processing / extraction
Knowledge assistant / SOPs / policies
Decision support / case review
Reporting / executive signal digest
Workflow automation
Legacy system / RPA workflow
Customer service / ticket triage
Sales / pipeline intelligence
Compliance / risk review
Agentic task execution
Other / unsure
Guidance: Select the closest category. This drives the recommended AI solution pattern, candidate tooling, risk library and delivery model.Format: Select one category. Choose Other / unsure only if no option fits.
Function / team *
Guidance: Name the team or function that owns or runs the workflow. Avoid broad labels such as “the business”.Format: Text. 2-190 characters.
Workflow / use case title *
Guidance: Use a concise, specific title that describes the workflow being improved. Do not use a tool name as the title.Format: Text. 5-255 characters.
Problem / opportunity *
Guidance: Describe the current problem using measurable facts where possible: volume, time, errors, delays, cost, customer impact, compliance exposure or revenue leakage.Format: Paragraph. Minimum 40 characters. Include numbers where available. Do not paste confidential case data.
So That statement *
Guidance: Write the business outcome in “So that…” form. State what changes for users, managers, customers, risk or financial performance.Format: Paragraph. Minimum 30 characters. Start with “So that…” where possible.
North Star metric *
Guidance: Enter the single most important measurable outcome. Good examples include time saved, cost reduction, error reduction, throughput, conversion rate or SLA improvement.Format: Specific metric with target value and unit. Example: “Reduce error rate from 8% to below 1%”.
Business owner *
Guidance: Enter the role accountable for business value and workflow ownership. Use “TBC” only if the owner has not yet been confirmed.Format: Role name or person/role. Minimum 2 characters.
Executive sponsor *
Pilot owner *
Guidance: Enter the person or role who would coordinate the pilot, manage evidence collection and drive follow-up actions. Use “TBC” if not yet assigned.Format: Role name or person/role. Minimum 2 characters.
Current status *
Idea
Discovery
Prioritised
Prototype
Pilot
Scale
Sunset
Guidance: Select the current delivery status. Choose Idea if no formal discovery or delivery work has started.Format: Select one status.
Data sensitivity *
Internal only
Confidential business data
Personal data
Special category / regulated data
Public or low sensitivity
Unsure
Guidance: Select the highest sensitivity level involved in the workflow. If unsure, choose Unsure rather than downgrading the risk.Format: Select one option.
AI action level *
Drafts or summaries only
Recommends decisions
Proposes system updates
Takes action after human approval
Takes action automatically
Unsure
Guidance: Select the most advanced action the AI would take in the target workflow. Higher autonomy increases governance requirements.Format: Select one option.
Impact if the AI is wrong *
Low operational impact
Medium operational impact
High customer / employee impact
Financial / legal / compliance impact
Safety, clinical or regulated decision impact
Unsure
Guidance: Select the highest realistic consequence if the AI output or automated action is wrong.Format: Select one option.
3. Current-state workflow baseline
Capture how the workflow actually operates today. These fields drive the current-state baseline, architecture, risk register and investment-readiness report sections.
Workflow trigger *
Guidance: State the event that starts the workflow. Be specific.Format: Text. Minimum 8 characters. Example: “Customer submits web enquiry”.
Input channels *
Guidance: List every channel where work arrives and include percentages if known.Format: Comma-separated text. Include percentages where possible.
Systems currently touched *
Guidance: List all systems, spreadsheets, portals and databases touched in the current process.Format: Comma-separated system names.
System of record *
Guidance: Name the authoritative system where the final record, transaction or decision must live.Format: Text. Use TBC only if not confirmed.
Roles or teams involved today *
Guidance: List the roles that touch the current workflow.Format: Comma-separated roles or teams.
Approximate number of people touching the workflow *
Guidance: Enter the approximate number of people who touch, approve, monitor or correct this workflow today. Use the closest realistic number.Format: Whole number. Must be greater than 0.
Number of handoffs / approval transitions *
Guidance: Count transfers between people, teams, queues or systems. Include approval routing and exception escalation.Format: Whole number. Enter 0 only if there are no handoffs.
Number of systems, files or tools touched *
Guidance: Count every core system, spreadsheet, mailbox, portal, document store or reporting tool used in the current process.Format: Whole number. Must be greater than 0.
Handoffs and approvals *
Guidance: Describe who hands work to whom and where approvals occur.Format: Paragraph. Minimum 30 characters.
Exception types *
Guidance: List common exception types that interrupt the standard flow.Format: Paragraph or comma-separated list. Minimum 20 characters.
Primary bottleneck step *
Guidance: Name the slowest, most expensive or highest-risk step.Format: Text. Minimum 5 characters.
Workflow step summary *
Guidance: Enter a concise step table in text form. Include owner, system, time and pain for each major step.Format: Structured text. Minimum 60 characters. Use semicolons to separate steps.
Manual copy/paste involved? *
Select…
Yes
No
Unsure
Guidance: Select whether users manually transfer data between tools or documents.Format: Select one option.
Email dependency? *
Select…
Yes
No
Unsure
Guidance: Select whether the workflow relies materially on email chains, inboxes or manual chasing.Format: Select one option.
Spreadsheet dependency? *
Select…
Yes
No
Unsure
Guidance: Select whether spreadsheets are used as trackers, control logs, calculators or unofficial systems of record.Format: Select one option.
Current audit trail quality *
Select…
Poor – little or no audit trail
Partial – some trail but inconsistent
Good – most steps traceable
Strong – complete and reviewable audit trail
Unsure
Guidance: Assess whether the current workflow produces a complete, reviewable trail of actions, approvals and evidence.Format: Select one option.
Structured workflow step builder
Complete at least 1 step for Client Guided Assessment and at least 3 steps for Consultant Advanced Assessment. Each used row must be complete. This creates the workflow heatmap and strengthens evidence quality.
Workflow step 1
Step 1 – activity / decision
Guidance: Enter one discrete workflow activity, decision, review, handoff or system action.Format: Text. Complete all fields in a row if this step is used.
Step 1 – owner / role
Guidance: Enter the role that performs or owns this step today.Format: Text role name.
Step 1 – system / tool
Guidance: Enter the main system, spreadsheet, inbox or tool used for this step.Format: Text system/tool name.
Step 1 – input
Guidance: Enter what information or artefact enters this step.Format: Text input description.
Step 1 – output
Guidance: Enter what this step produces or passes onward.Format: Text output description.
Step 1 – average minutes
Guidance: Estimate minutes spent on this step for one typical case. Use 0 only where instantaneous/system-driven.Format: Number in minutes.
Step 1 – pain level
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score the operational pain in this step: delay, rework, frustration, cost, dependency or inconsistency.Format: Select 1-5 if this step is used.
Step 1 – AI interpretation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether AI can help interpret, classify, extract, summarise, draft or recommend at this step.Format: Select 1-5 if this step is used.
Step 1 – automation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether this step can be executed or routed by automation once the logic and controls are clear.Format: Select 1-5 if this step is used.
Step 1 – risk if automated incorrectly
Select…
1 – Low consequence
2 – Manageable consequence
3 – Moderate operational consequence
4 – High financial/customer/compliance consequence
5 – Critical or regulated consequence
Guidance: Score the consequence if this step is automated badly or without the right human review.Format: Select 1-5 if this step is used.
Workflow step 2
Step 2 – activity / decision
Guidance: Enter one discrete workflow activity, decision, review, handoff or system action.Format: Text. Complete all fields in a row if this step is used.
Step 2 – owner / role
Guidance: Enter the role that performs or owns this step today.Format: Text role name.
Step 2 – system / tool
Guidance: Enter the main system, spreadsheet, inbox or tool used for this step.Format: Text system/tool name.
Step 2 – input
Guidance: Enter what information or artefact enters this step.Format: Text input description.
Step 2 – output
Guidance: Enter what this step produces or passes onward.Format: Text output description.
Step 2 – average minutes
Guidance: Estimate minutes spent on this step for one typical case. Use 0 only where instantaneous/system-driven.Format: Number in minutes.
Step 2 – pain level
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score the operational pain in this step: delay, rework, frustration, cost, dependency or inconsistency.Format: Select 1-5 if this step is used.
Step 2 – AI interpretation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether AI can help interpret, classify, extract, summarise, draft or recommend at this step.Format: Select 1-5 if this step is used.
Step 2 – automation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether this step can be executed or routed by automation once the logic and controls are clear.Format: Select 1-5 if this step is used.
Step 2 – risk if automated incorrectly
Select…
1 – Low consequence
2 – Manageable consequence
3 – Moderate operational consequence
4 – High financial/customer/compliance consequence
5 – Critical or regulated consequence
Guidance: Score the consequence if this step is automated badly or without the right human review.Format: Select 1-5 if this step is used.
Workflow step 3
Step 3 – activity / decision
Guidance: Enter one discrete workflow activity, decision, review, handoff or system action.Format: Text. Complete all fields in a row if this step is used.
Step 3 – owner / role
Guidance: Enter the role that performs or owns this step today.Format: Text role name.
Step 3 – system / tool
Guidance: Enter the main system, spreadsheet, inbox or tool used for this step.Format: Text system/tool name.
Step 3 – input
Guidance: Enter what information or artefact enters this step.Format: Text input description.
Step 3 – output
Guidance: Enter what this step produces or passes onward.Format: Text output description.
Step 3 – average minutes
Guidance: Estimate minutes spent on this step for one typical case. Use 0 only where instantaneous/system-driven.Format: Number in minutes.
Step 3 – pain level
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score the operational pain in this step: delay, rework, frustration, cost, dependency or inconsistency.Format: Select 1-5 if this step is used.
Step 3 – AI interpretation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether AI can help interpret, classify, extract, summarise, draft or recommend at this step.Format: Select 1-5 if this step is used.
Step 3 – automation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether this step can be executed or routed by automation once the logic and controls are clear.Format: Select 1-5 if this step is used.
Step 3 – risk if automated incorrectly
Select…
1 – Low consequence
2 – Manageable consequence
3 – Moderate operational consequence
4 – High financial/customer/compliance consequence
5 – Critical or regulated consequence
Guidance: Score the consequence if this step is automated badly or without the right human review.Format: Select 1-5 if this step is used.
Workflow step 4
Step 4 – activity / decision
Guidance: Enter one discrete workflow activity, decision, review, handoff or system action.Format: Text. Complete all fields in a row if this step is used.
Step 4 – owner / role
Guidance: Enter the role that performs or owns this step today.Format: Text role name.
Step 4 – system / tool
Guidance: Enter the main system, spreadsheet, inbox or tool used for this step.Format: Text system/tool name.
Step 4 – input
Guidance: Enter what information or artefact enters this step.Format: Text input description.
Step 4 – output
Guidance: Enter what this step produces or passes onward.Format: Text output description.
Step 4 – average minutes
Guidance: Estimate minutes spent on this step for one typical case. Use 0 only where instantaneous/system-driven.Format: Number in minutes.
Step 4 – pain level
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score the operational pain in this step: delay, rework, frustration, cost, dependency or inconsistency.Format: Select 1-5 if this step is used.
Step 4 – AI interpretation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether AI can help interpret, classify, extract, summarise, draft or recommend at this step.Format: Select 1-5 if this step is used.
Step 4 – automation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether this step can be executed or routed by automation once the logic and controls are clear.Format: Select 1-5 if this step is used.
Step 4 – risk if automated incorrectly
Select…
1 – Low consequence
2 – Manageable consequence
3 – Moderate operational consequence
4 – High financial/customer/compliance consequence
5 – Critical or regulated consequence
Guidance: Score the consequence if this step is automated badly or without the right human review.Format: Select 1-5 if this step is used.
Workflow step 5
Step 5 – activity / decision
Guidance: Enter one discrete workflow activity, decision, review, handoff or system action.Format: Text. Complete all fields in a row if this step is used.
Step 5 – owner / role
Guidance: Enter the role that performs or owns this step today.Format: Text role name.
Step 5 – system / tool
Guidance: Enter the main system, spreadsheet, inbox or tool used for this step.Format: Text system/tool name.
Step 5 – input
Guidance: Enter what information or artefact enters this step.Format: Text input description.
Step 5 – output
Guidance: Enter what this step produces or passes onward.Format: Text output description.
Step 5 – average minutes
Guidance: Estimate minutes spent on this step for one typical case. Use 0 only where instantaneous/system-driven.Format: Number in minutes.
Step 5 – pain level
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score the operational pain in this step: delay, rework, frustration, cost, dependency or inconsistency.Format: Select 1-5 if this step is used.
Step 5 – AI interpretation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether AI can help interpret, classify, extract, summarise, draft or recommend at this step.Format: Select 1-5 if this step is used.
Step 5 – automation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether this step can be executed or routed by automation once the logic and controls are clear.Format: Select 1-5 if this step is used.
Step 5 – risk if automated incorrectly
Select…
1 – Low consequence
2 – Manageable consequence
3 – Moderate operational consequence
4 – High financial/customer/compliance consequence
5 – Critical or regulated consequence
Guidance: Score the consequence if this step is automated badly or without the right human review.Format: Select 1-5 if this step is used.
Workflow step 6
Step 6 – activity / decision
Guidance: Enter one discrete workflow activity, decision, review, handoff or system action.Format: Text. Complete all fields in a row if this step is used.
Step 6 – owner / role
Guidance: Enter the role that performs or owns this step today.Format: Text role name.
Step 6 – system / tool
Guidance: Enter the main system, spreadsheet, inbox or tool used for this step.Format: Text system/tool name.
Step 6 – input
Guidance: Enter what information or artefact enters this step.Format: Text input description.
Step 6 – output
Guidance: Enter what this step produces or passes onward.Format: Text output description.
Step 6 – average minutes
Guidance: Estimate minutes spent on this step for one typical case. Use 0 only where instantaneous/system-driven.Format: Number in minutes.
Step 6 – pain level
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score the operational pain in this step: delay, rework, frustration, cost, dependency or inconsistency.Format: Select 1-5 if this step is used.
Step 6 – AI interpretation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether AI can help interpret, classify, extract, summarise, draft or recommend at this step.Format: Select 1-5 if this step is used.
Step 6 – automation fit
Select…
1 – Low
2 – Low/moderate
3 – Moderate
4 – Strong
5 – Very strong
Guidance: Score whether this step can be executed or routed by automation once the logic and controls are clear.Format: Select 1-5 if this step is used.
Step 6 – risk if automated incorrectly
Select…
1 – Low consequence
2 – Manageable consequence
3 – Moderate operational consequence
4 – High financial/customer/compliance consequence
5 – Critical or regulated consequence
Guidance: Score the consequence if this step is automated badly or without the right human review.Format: Select 1-5 if this step is used.
4. Target state, success metrics and evidence quality
Define the target operating model and evidence confidence. Weak or self-entered evidence produces a preliminary report. Only consultant-validated evidence can support a board-ready decision pack.
Target operating model *
Guidance: Describe how the workflow should operate after AI/automation is introduced.Format: Paragraph. Minimum 50 characters.
Human-in-the-loop review model *
Guidance: State exactly when humans review, approve, reject, override or escalate AI outputs.Format: Paragraph. Minimum 40 characters.
Go-live KPI *
Guidance: Define the KPI that must be achieved for pilot go-live success.Format: Metric with target and timeframe.
Scale KPI *
Guidance: Define the KPI that must be achieved before scaling beyond pilot.Format: Metric with target and timeframe.
Management reporting cadence *
Guidance: State how often benefits, issues, costs and risks will be reviewed.Format: Text. Example: weekly, monthly, steering committee cadence.
Workflow evidence confidence *
Select…
Measured – taken from system/report data
Sampled – based on reviewed cases
Estimated – provided by business owner
Assumed – consultant/rough assumption
Unknown – not validated yet
Guidance: Select how well the workflow steps and pain points are evidenced.Format: Select one confidence level.
Baseline metric confidence *
Select…
Measured – taken from system/report data
Sampled – based on reviewed cases
Estimated – provided by business owner
Assumed – consultant/rough assumption
Unknown – not validated yet
Guidance: Select whether volume, time, error and cost baselines are measured, sampled, estimated or assumed.Format: Select one confidence level.
Implementation cost confidence *
Select…
Measured – taken from system/report data
Sampled – based on reviewed cases
Estimated – provided by business owner
Assumed – consultant/rough assumption
Unknown – not validated yet
Guidance: Select how reliable the build/configuration/integration cost estimate is.Format: Select one confidence level.
Benefit confidence *
Select…
Measured – taken from system/report data
Sampled – based on reviewed cases
Estimated – provided by business owner
Assumed – consultant/rough assumption
Unknown – not validated yet
Guidance: Select how reliable the expected benefit case is.Format: Select one confidence level.
Pricing confidence *
Select…
Measured – taken from system/report data
Sampled – based on reviewed cases
Estimated – provided by business owner
Assumed – consultant/rough assumption
Unknown – not validated yet
Guidance: Select how reliable the model, API, tool, licence and automation pricing assumptions are.Format: Select one confidence level.
Pricing registry status *
Select…
Official pricing source checked
Vendor quote available
Internal estimate only
Sales-required / unpublished pricing
Not checked yet
Guidance: State whether pricing came from an official source, vendor quote or manual estimate. Reports snapshot this assumption.Format: Select one status.
Volume source *
Select…
System report / export
Time-and-motion sample
Finance report
CRM / ERP / ticketing report
Vendor quote / official pricing
Business-owner estimate
Consultant assumption
Unknown / not yet validated
Guidance: State where the volume figure came from. Use Unknown only if no source exists yet.Format: Select one source.
Cycle-time source *
Select…
System report / export
Time-and-motion sample
Finance report
CRM / ERP / ticketing report
Vendor quote / official pricing
Business-owner estimate
Consultant assumption
Unknown / not yet validated
Guidance: State how the current handling-time figure was obtained.Format: Select one source.
Error / rework-rate source *
Select…
System report / export
Time-and-motion sample
Finance report
CRM / ERP / ticketing report
Vendor quote / official pricing
Business-owner estimate
Consultant assumption
Unknown / not yet validated
Guidance: State where the error, query, exception or rework rate came from.Format: Select one source.
Implementation-cost source *
Select…
System report / export
Time-and-motion sample
Finance report
CRM / ERP / ticketing report
Vendor quote / official pricing
Business-owner estimate
Consultant assumption
Unknown / not yet validated
Guidance: State whether implementation cost is from a quote, delivery estimate, prior project or assumption.Format: Select one source.
Annual run-cost source *
Select…
System report / export
Time-and-motion sample
Finance report
CRM / ERP / ticketing report
Vendor quote / official pricing
Business-owner estimate
Consultant assumption
Unknown / not yet validated
Guidance: State where licence, model, automation, support and run-cost assumptions came from.Format: Select one source.
Benefit-source basis *
Select…
System report / export
Time-and-motion sample
Finance report
CRM / ERP / ticketing report
Vendor quote / official pricing
Business-owner estimate
Consultant assumption
Unknown / not yet validated
Guidance: State the source of the productivity, revenue, quality, risk or cost-saving assumptions.Format: Select one source.
Evidence validation actions required *
Guidance: List the validation actions required before TFx should treat this as consultant-validated or investment-ready.Format: Paragraph/list. Minimum 40 characters.
Source evidence note *
Guidance: Summarise where the key numbers came from. Do not upload or paste confidential raw data into the public form.Format: Paragraph. Minimum 50 characters.
6. Economics, pricing assumptions and scenario inputs
These inputs are required because the report is intended to support an investment decision. Use measured data where available. If a cost or benefit category is genuinely not applicable, enter 0. Do not leave vital economics blank.
Current cycle time per case / workflow, in minutes *
Guidance: Enter the current average handling time for one item/case/run of the workflow, in minutes. Use a measured average where possible.Format: Number in minutes. Must be greater than 0. Example: 22.5
Target cycle time after AI support, in minutes *
Guidance: Enter the expected average handling time after the AI-assisted or automated workflow is working. Use the controlled-pilot target, not a theoretical perfect state.Format: Number in minutes. Can be 0 for fully automated straight-through processing, but only where realistic.
Estimated weekly volume *
Guidance: Enter how many items/cases/workflow runs occur per week. Convert monthly volume by multiplying by 12 and dividing by 52.Format: Number of items per week. Must be greater than 0.
Loaded hourly cost, £ *
Guidance: Enter the fully loaded hourly cost of the staff time affected, including salary, employer costs, overhead or charge-out proxy.Format: GBP number per hour. Must be greater than 0. Example: 40
Additional annual revenue or margin benefit, £ *
Guidance: Enter expected annual revenue or margin upside directly attributable to this use case. Enter 0 if not applicable.Format: GBP number per year. Use 0 only if genuinely not applicable.
Annual risk, quality or compliance saving, £ *
Guidance: Enter annual savings from reduced errors, rework, penalties, claims, quality issues, compliance work or audit exposure. Enter 0 if not applicable.Format: GBP number per year. Use 0 only if genuinely not applicable.
Current error / rework rate, % *
Guidance: Enter the percentage of cases that currently require correction, rework, dispute handling or exception resolution.Format: Percentage number from 0 to 100. Enter 0 only if genuinely none.
Target error / rework rate, % *
Guidance: Enter the expected error/rework rate after the solution is piloted. Use a realistic controlled-pilot target.Format: Percentage number from 0 to 100.
Average rework time per error, minutes *
Guidance: Enter average minutes spent resolving one error, query, failed case or exception.Format: Number in minutes. Use 0 only if not applicable.
Annual delay, SLA, penalty or leakage cost, £ *
Guidance: Enter annual cost of delays, SLA breaches, penalties, missed revenue, churn or service credits. Use 0 if not applicable.Format: GBP per year. Use whole pounds.
Productive hours per FTE per year *
Guidance: Enter productive working hours per full-time equivalent after leave, meetings and non-productive time. 1,500-1,700 is common for UK modelling.Format: Number of hours. Must be greater than 0.
Target automation / AI-assisted percentage, % *
Guidance: Enter the percentage of total workflow volume expected to be handled by the AI-assisted or automated target process.Format: Percentage number from 0 to 100.
Target user adoption percentage, % *
Guidance: Enter the percentage of target users expected to use the redesigned workflow consistently after launch.Format: Percentage number from 0 to 100.
One-off discovery / workflow design cost, £ *
Guidance: Estimate one-off discovery, process mapping, requirements, stakeholder workshops and solution design cost. Enter 0 only if already completed and funded elsewhere.Format: GBP one-off cost. Use whole pounds.
One-off build / configuration cost, £ *
Guidance: Estimate configuration, automation, AI workflow build, prompt/model setup, application work or vendor implementation cost.Format: GBP one-off cost. Use whole pounds.
One-off data / integration cost, £ *
Guidance: Estimate data preparation, system integration, API, connector, migration, permissions or environment setup cost.Format: GBP one-off cost. Use whole pounds.
One-off testing / governance cost, £ *
Guidance: Estimate UAT, security review, compliance review, DPIA, controls testing, model/output validation and go/no-go governance cost.Format: GBP one-off cost. Use whole pounds.
One-off training / change cost, £ *
Guidance: Estimate training, communication, SOP updates, adoption support, manager enablement and change-management cost.Format: GBP one-off cost. Use whole pounds.
One-off procurement / vendor setup cost, £ *
Guidance: Estimate legal, procurement, vendor onboarding, contract review, tenant setup or paid proof-of-concept setup cost. Enter 0 if not applicable.Format: GBP one-off cost. Use whole pounds.
Other one-off implementation cost, £ *
Guidance: Enter any additional one-off cost not captured above, such as hardware, external audit, data clean-up reserve or contingency. Enter 0 if none.Format: GBP one-off cost. Use 0 if none.
LLM / AI requests per month *
Guidance: Estimate monthly AI/model calls. Use 0 if the use case does not use an LLM or model API.Format: Whole number per month. Use 0 if not applicable.
Average input tokens per request *
Guidance: Estimate average prompt/context tokens per request, including retrieval context if applicable. Use 0 if unknown/not applicable.Format: Whole number. Use 0 if not applicable.
Average output tokens per request *
Guidance: Estimate average generated output tokens per request. Use 0 if unknown/not applicable.Format: Whole number. Use 0 if not applicable.
Documents processed per month *
Guidance: Estimate monthly document volume for document AI, OCR, extraction or review use cases. Use 0 if not applicable.Format: Whole number per month. Use 0 if not applicable.
Average pages per document *
Guidance: Estimate pages per document where document AI/OCR costs are page-based. Use 0 if not applicable.Format: Number. Use 0 if not applicable.
Automation runs per month *
Guidance: Estimate workflow automation runs per month. Use 0 if no automation platform usage is expected.Format: Whole number per month. Use 0 if not applicable.
Average automation actions per run *
Guidance: Estimate actions/tasks/steps consumed per automation run for Make, Zapier, Power Automate, n8n or similar tools.Format: Number. Use 0 if not applicable.
SaaS / AI tool seats *
Guidance: Estimate number of paid user seats needed for the target workflow. Use 0 if no per-seat licensing is expected.Format: Whole number. Use 0 if not applicable.
RPA bots required *
Guidance: Estimate attended/unattended bot count if RPA is needed. Use 0 if not applicable.Format: Whole number. Use 0 if not applicable.
Annual licences / SaaS cost, £ *
Guidance: Estimate annual licences, subscriptions, user seats, automation plans, RPA bots or SaaS platform costs.Format: GBP per year. Use 0 only if no recurring licence cost exists.
Annual model / API / compute cost, £ *
Guidance: Estimate annual LLM/API/token, document-processing, cloud compute, storage, vector database or automation-usage costs.Format: GBP per year. Use usage assumptions or vendor pricing where possible. Enter 0 only if not applicable.
Annual support / maintenance cost, £ *
Guidance: Estimate annual support, bug fixing, prompt/model tuning, platform maintenance, vendor support or minor enhancement cost.Format: GBP per year. Enter 0 only if no support cost will exist.
Annual internal ownership / admin cost, £ *
Guidance: Estimate internal business-owner/admin time required to own, review and keep the workflow operating during the year.Format: GBP per year. Convert expected internal hours to cost using loaded hourly cost. Enter 0 only if genuinely none.
Annual monitoring / governance cost, £ *
Guidance: Estimate annual cost for audit checks, access reviews, output QA, compliance reviews, policy updates, incident review or KPI reporting.Format: GBP per year. Enter 0 only if not applicable.
Other annual run cost, £ *
Guidance: Enter any other recurring annual cost not captured above, such as data labelling, external assurance, extra storage, premium support or contingency. Enter 0 if none.Format: GBP per year. Use 0 if none.
Formula used: Annual hours saved = max(0, current minutes – target minutes) x weekly volume x 52 / 60. The v2 model also calculates baseline labour cost, rework cost, FTE equivalent, current cost of problem, year-one cost, three-year scenarios, ROI and payback. Enter 0 only where a category is genuinely not applicable.
7. Buy vs build assessment
This section prevents expensive over-build. It compares whether the use case should be bought, configured using the existing stack, built bespoke, or handled as a hybrid.
Market / vendor solution fit *
Select…
No known vendor option
Generic tools could cover part of it
One or more credible SaaS/vendor options exist
Existing Microsoft/Google/CRM stack may cover it
Unsure
Guidance: Assess whether credible existing tools could solve most of the problem. This prevents unnecessary bespoke build.Format: Select one option. Choose Unsure if no market scan has been done.
Strategic differentiation *
Select…
Commodity / non-differentiating workflow
Important but not competitively unique
Strategically differentiating capability
Core IP / proprietary operating advantage
Unsure
Guidance: State whether the workflow is commodity, important, differentiating, or core IP. This influences buy/configure/build decisions.Format: Select one option.
Integration / data complexity *
Select…
Low – standalone or manual upload
Moderate – connects to one or two systems
High – multiple systems/data sources
Very high – regulated, real-time or complex permissions
Unsure
Guidance: Estimate the number and complexity of systems, data sources, permissions and integration points involved.Format: Select one option. Pick the higher complexity level if unsure.
Internal build / configuration capability *
Select…
No internal delivery capability
Some configuration capability
Strong automation/low-code capability
Strong software/AI engineering capability
Unsure
Guidance: Assess whether the organisation can configure/build/support the solution internally or will need external delivery support.Format: Select one option.
Urgency to value *
Select…
Need value within 30 days
Need value within 60-90 days
Can wait for tailored build
Unsure
Guidance: Select the time window in which useful evidence or operational value is needed. This heavily affects MVP route and sourcing choice.Format: Select one option.
Existing stack to leverage *
Guidance: List platforms, licences, enterprise agreements, approved tools or internal capabilities that should be considered before new procurement.Format: Paragraph or comma-separated list. Minimum 20 characters. Use TBC if not yet known.
Vendor, procurement or security constraints *
Guidance: State any preferred vendors, prohibited vendors, procurement rules, security restrictions, data residency constraints or contract limitations.Format: Paragraph or bullet-style text. Minimum 20 characters. Use None known if genuinely none.
Preferred MVP boundary *
Guidance: Define what should be included and excluded from a safe first pilot. This prevents over-scoping and helps the report generate a credible delivery plan.Format: Paragraph. Minimum 30 characters.
