Decision owner
Process Excellence, Transformation, Operations, or Compliance owns the business outcome.
Buyer tool / Vendor evaluation
Evaluate process intelligence platforms on one real process.
For Process Excellence, Enterprise Architecture, Compliance, IT, and Procurement teams with an active shortlist. Set the rules before the demos, score every supplier on the same evidence, and keep the decision trail.
Use it at the decision stage
A working evaluation file for the shortlist.
Use this after you have a shortlist and a credible pilot process. Earlier in the buying journey, start with the buyer guide. Existing Geist users should use product documentation and support.
Technical reviewers
Enterprise Architecture, IT, Data, and Security test integration, access, operation, and exit.
Commercial reviewers
Procurement and Legal validate evidence, contract assumptions, risk, and total cost.
Evaluation support
Consultants may facilitate the process, but the buyer owns criteria, weights, and the final record.
The evaluation sequence
Freeze the rules before a vendor enters the room.
A defensible comparison starts with your use case and ends with recorded evidence. Do not change criteria or weights because one demo looked impressive.
- 01
Scope one pilot
Name the process, owner, source material, operating constraints, and decision date.
- 02
Set gates and weights
Separate non-negotiable controls from preferences that can be scored and traded off.
- 03
Run the same proof
Give every supplier the same scenario, inputs, questions, timebox, and scoring scale.
- 04
Record the decision
Keep scores, evidence, failed gates, assumptions, and the reason for the final choice.
Interactive scorecard
Build the evaluation in your browser.
Your entries stay in this browser. Compare up to three suppliers, export the full record as CSV, or print a decision copy.
Name the one process you will test.
Every supplier gets the same scenario. One bounded process keeps the comparison honest.
Who is on the shortlist?
Compare up to three suppliers side by side. If Geist is on your list, score it under the same rules.
Names appear in the report exactly as entered.
Decide what matters most, before the demos.
Set how much each area counts. The weights are calculated for you and always total 100%.
Process fitData fitModeling and portabilityAnalysis and evidenceGovernance and AI controlRollout and commercial fit
Process fitCan the platform represent the work you actually need to govern?
Data fitCan your real source material enter the platform with its meaning and provenance intact?
Modeling and portabilityCan teams edit, exchange, and leave with usable process models?
Analysis and evidenceDoes the platform explain findings under conditions close to your deployment?
Governance and AI controlAre authority, human review, and the audit trail explicit?
Rollout and commercial fitCan a bounded pilot become an operable service without trapping the buyer?
Check the non-negotiables.
A failed gate rules a supplier out, whatever its score. Mark each gate once you have evidence.
Only a gate marked Fail rules a supplier out. Unchecked gates keep the result open.
Security and data handling cleared
Supplier A
Supplier B
Supplier C
Required import, export, and offboarding proven
Supplier A
Supplier B
Supplier C
Access, approval, and audit controls proven
Supplier A
Supplier B
Supplier C
Pilot scope, responsibilities, and exit agreed
Supplier A
Supplier B
Supplier C
Score what each supplier proved.
Rate the evidence, not the pitch. Add a short note where a reviewer would ask why.
0Not shown1Claim only2Partial demo3Representative proof4Proven on our process
Can the platform represent the work you actually need to govern?
Scope and variants
Model the pilot process, its critical variants, exceptions, handoffs, and escalation paths.
Supplier A
Supplier B
Supplier C
Ownership and controls
Show owners, approvers, controls, risks, and review cadence in the process context.
Supplier A
Supplier B
Supplier C
Change workflow
Change the model, route it for review, and show the retained version and decision history.
Supplier A
Supplier B
Supplier C
Can your real source material enter the platform with its meaning and provenance intact?
Required sources
Use the pilot documents, files, system exports, or event data rather than a prepared vendor dataset.
Supplier A
Supplier B
Supplier C
Incomplete data
Show what the platform rejects, flags, or asks a human to resolve when input is incomplete.
Supplier A
Supplier B
Supplier C
Provenance and access
Trace an output to its source and show how access follows roles, organisations, and data boundaries.
Supplier A
Supplier B
Supplier C
Can teams edit, exchange, and leave with usable process models?
Editable process model
Edit the discovered or imported model without rebuilding it in a separate tool.
Supplier A
Supplier B
Supplier C
BPMN 2.0 fidelity
Import and export representative BPMN 2.0, then compare semantics, layout, labels, and extensions.
Supplier A
Supplier B
Supplier C
Operational handoff
Export a model and decision pack that remain useful without a continuing vendor session.
Supplier A
Supplier B
Supplier C
Does the platform explain findings under conditions close to your deployment?
Observed versus intended
Compare actual variants with the intended process and identify the evidence behind a finding.
Supplier A
Supplier B
Supplier C
Limits and uncertainty
Show a weak or ambiguous result and explain confidence, missing evidence, and safe failure.
Supplier A
Supplier B
Supplier C
Ongoing monitoring
Show how new data, model changes, incidents, and drift alter findings after the pilot.
Supplier A
Supplier B
Supplier C
Are authority, human review, and the audit trail explicit?
Action boundaries
Distinguish read, draft, propose, approve, execute, block, and roll back for the pilot use case.
Supplier A
Supplier B
Supplier C
Human oversight
Show who can intervene, reject, correct, or stop an AI-assisted action and what context they receive.
Supplier A
Supplier B
Supplier C
Decision trail
Trace source, model version, actor, recommendation, approval, action, outcome, and correction.
Supplier A
Supplier B
Supplier C
Can a bounded pilot become an operable service without trapping the buyer?
Pilot plan
Name the first process, owner, source, success measure, timeline, responsibilities, and exit criteria.
Supplier A
Supplier B
Supplier C
Operating model
Show administration, support, change management, training, service levels, and incident handling.
Supplier A
Supplier B
Supplier C
Cost and exit
Record implementation and operating costs, contract assumptions, export, deletion, and offboarding.
Supplier A
Supplier B
Supplier C
Your comparison and decision record.
Ranked by weighted score, with failed gates and open items called out.
Score by area
Share of the maximum evidence score in each area. Further right is stronger.
Decision record
An A4 record with the ranking, both charts, the gates, every score and these next steps.
Your browser’s print window opens. Pick “Save as PDF” as the printer.
Clears this evaluation from this browser.
Method
Why the scorecard works this way.
The structure follows public procurement, software assurance, AI risk, and BPMN guidance. Adapt it to your organisation and applicable law; it is not legal or procurement advice.
UK Cabinet Office
Open source ↗Consistent weighted scoring
Set distinct criteria and weights with the people who will use and govern the service.
CISA
Open source ↗Control questions with evidence
Treat non-negotiable supplier and software controls separately from tradeable quality scores.
NIST AI RMF
Open source ↗Testing in deployment context
Evaluate AI behavior, limits, oversight, and monitoring under conditions close to the intended use.
Object Management Group
Open source ↗Portable process semantics
Test BPMN exchange against the published specification, including machine-readable structures.
Optional commercial help
Need help setting up the evaluation?
The Geist team can prepare the evidence request and facilitate a workshop on one of your processes. If Geist is on the shortlist, it is scored under the same rules. You keep the evaluation file and the decision.