01 / AI & human judgment Try first
AI task-fit map For small operations teams
A task-by-task map of draft, review and human-only work, with explicit failure costs.
Explore the experiment The need Teams buy an AI tool before deciding which part of the work needs help. First small experiment Map one recurring process with its owner. Compare an assisted draft with the current approach on ten approved examples. What to look for Reviewers can identify a useful task without increasing correction time or accepting a critical error. Before building Approved examples, a named reviewer and a decision about where data may be processed. Start with an available tool Small prototype Proposed concept
02 / AI & human judgment Explore
Source-led research assistant For research and client-service teams
A prototype that drafts only from an approved document set and links each claim to a passage.
Explore the experiment The need A fluent answer is difficult to use when its supporting evidence is missing. First small experiment Use public documents to answer five narrow questions. Ask a reviewer to trace every substantive claim. What to look for Claims are traceable and unsupported questions produce an explicit gap rather than a fabricated answer. Before building Document permissions, retrieval evaluation, model cost limits and a human publication gate. Start with an available tool Medium prototype Proposed concept
03 / AI & human judgment Explore
Human review bench For teams evaluating generated outputs
A small evaluation workspace that compares versions against agreed examples and records reviewer disagreements.
Explore the experiment The need Prompt changes are judged by a few memorable examples. First small experiment Score two versions of a low-risk drafting task on twenty anonymized examples using the same rubric. What to look for The team can explain where a version improves, regresses or still needs a human decision. Before building Representative cases, an agreed rubric, private storage and reviewer time. Start with an available tool Medium prototype Proposed concept
04 / Operations & automation Try first
Handoff rehearsal studio For founders and distributed teams
A guided handoff builder covering inputs, exceptions, acceptance checks and the next owner.
Explore the experiment The need A process works only while its original owner is available. First small experiment Write one handoff and ask a colleague unfamiliar with the task to follow it without coaching. What to look for The colleague completes the bounded task and identifies missing steps before a real handover. Before building A process owner, safe test access and a rollback or escalation path. Start with an available tool Small prototype Proposed concept
05 / Operations & automation Explore
Exception-first intake desk For service operations teams
A proposed intake queue with visible exceptions, responsible people and time-bound follow-up.
Explore the experiment The need Automation handles the normal case while unusual requests disappear into chats. First small experiment Model one request type in a shared worksheet before connecting any live systems. What to look for Every exception has an owner and a next action; urgent cases remain visible. Before building Clear access roles, escalation rules and agreement on which requests must stay manual. Start with an available tool Medium prototype Proposed concept
06 / Operations & automation Explore
Workflow capacity sandbox For small service businesses
An interactive scenario model for arrivals, review capacity, rework and waiting time.
Explore the experiment The need Adding demand can create a new bottleneck rather than useful growth. First small experiment Use a week of aggregate counts to compare the current process with one staffing or handoff change. What to look for The owner understands which assumption drives the outcome and what observation to collect next. Before building Non-personal operational counts and an explicit statement that this is a model, not a forecast. Start with an available tool Medium prototype Proposed concept
07 / Software decisions Try first
Software ownership-cost lab For teams choosing business software
A comparison worksheet with adjustable user growth, migration effort and maintenance assumptions.
Explore the experiment The need A low subscription price hides onboarding, administration and exit costs. First small experiment Compare two real short-listed tools using published prices and a timed sample workflow. What to look for The team can state its preferred option and the assumption that would change the decision. Before building Current vendor terms, confirmed requirements and a buyer-owned budget. Start with an available tool Small prototype Proposed concept
08 / Software decisions Try first
Software exit drill For teams replacing a business-critical tool
A rehearsal kit for exporting sample records, attachments, relationships and permissions.
Explore the experiment The need Export capability is assumed until the team needs to leave. First small experiment Move a fictional project out of one trial product and reconstruct it in a neutral format. What to look for Someone other than the original tester can read the output and describe what was lost. Before building Trial permissions, representative sample data and access to a realistic destination. Start with an available tool Small prototype Proposed concept
09 / Software decisions Explore
Inclusive buying scorecard For teams with varied access and device needs
A task-based procurement scorecard combining access needs, support and everyday workflows.
Explore the experiment The need A polished demo hides barriers for keyboard users, mobile workers or low-bandwidth connections. First small experiment Ask consenting users with different needs to complete three relevant tasks in a trial. What to look for The selection records observed barriers and workarounds alongside commercial criteria. Before building Participant access needs, appropriate compensation and specialist review when required. Start with an available tool Small prototype Proposed concept
10 / Product experiments Try first
One-question offer lab For founders testing a new service
A prototype brief that connects one audience, a concrete problem and a testable offer.
Explore the experiment The need A large product build begins before the customer problem is understood. First small experiment Show two clearly labeled concept descriptions to five relevant people and discuss their current workaround. What to look for Participants describe a recurring problem and agree to a concrete next research step. Before building Relevant participants and an honest statement that the product is not available yet. Start with an available tool Small prototype Proposed concept
11 / Product experiments Explore
First-value journey lab For early product teams
A storyboard and clickable prototype focused on the first useful result rather than more signup steps.
Explore the experiment The need People complete setup but do not reach the outcome they came for. First small experiment Observe five people attempt one goal without explaining the interface. What to look for The team identifies a repeatable point of confusion and tests one specific repair. Before building A defined user goal, consent for observation and access to a testable prototype. Start with an available tool Medium prototype Proposed concept
12 / Product experiments Explore
Experiment evidence ledger For teams running several product tests
A record linking each hypothesis to observations, limitations, a decision and a review date.
Explore the experiment The need Old assumptions become accepted facts after the original experiment is forgotten. First small experiment Document three past decisions and run one new test using the same lightweight format. What to look for A colleague can distinguish an observation from an interpretation and find the next unresolved question. Before building A decision owner, transparent evidence standards and a manageable review cadence. Start with an available tool Small prototype Proposed concept
13 / Data & research Try first
Shared metric dictionary For teams combining reports from several tools
A proposed dictionary with definitions, examples, source fields, owners and version history.
Explore the experiment The need People use the same metric name for different calculations. First small experiment Resolve the definitions of five frequently disputed measures using a small synthetic dataset. What to look for Two colleagues calculate the same result from the same inputs and can explain exclusions. Before building Source-system knowledge and agreement about who approves definition changes. Start with an available tool Small prototype Proposed concept
14 / Data & research Explore
Research evidence board For small product and service research teams
A source-aware board that labels observations, interpretations, gaps and consent boundaries.
Explore the experiment The need A quote, an assumption and a verified fact look identical in a presentation. First small experiment Reorganize one existing research summary and ask a new colleague to trace its main conclusion. What to look for The conclusion can be checked without exposing participant identity or overstating the sample. Before building Research permissions, redaction rules and a retention decision. Start with an available tool Medium prototype Proposed concept
15 / Data & research Explore
Data quality watch cards For teams maintaining recurring csv or json reports
Local-first checks for required fields, ranges, duplicates and schema changes with readable explanations.
Explore the experiment The need Missing fields and changed formats are noticed after a report is shared. First small experiment Check a synthetic file with deliberate errors against a human-reviewed reference. What to look for Known errors are found and normal variation does not create a flood of false alarms. Before building Agreed rules, representative fixtures and a decision about whether files may leave the browser. Start with an available tool Medium prototype Proposed concept
16 / APIs & connected systems Try first
API contract planner For teams planning their first integration
A visual contract covering inputs, permissions, responses, timeouts and manual recovery.
Explore the experiment The need The happy-path demo hides ownership, field mapping and failure behavior. First small experiment Describe one integration using fictional payloads and walk through a failure with both owners. What to look for Each team can explain what happens when a request fails or its outcome is unknown. Before building Provider documentation, technical owners and a test environment. Start with an available tool Small prototype Proposed concept
17 / APIs & connected systems Later
Integration replay room For operators of scheduled data workflows
A sandbox for reviewing failed events, deduplication rules and explicitly approved replay.
Explore the experiment The need Rerunning a failed job can duplicate records or repeat an external action. First small experiment Simulate a dropped response and a duplicate event in a disposable test system. What to look for The team recovers once without creating a second business action and can trace the outcome. Before building Provider-supported retry behavior, isolation from production and an accountable operator. Start with an available tool Large prototype Proposed concept
18 / APIs & connected systems Explore
Integration dependency atlas For growing teams with overlapping saas tools
An ownership map connecting systems, data categories, credentials and recovery contacts without storing secrets.
Explore the experiment The need No one knows which workflow will break when a token, field or vendor changes. First small experiment Map one critical customer journey and rehearse a single provider outage on paper. What to look for Every dependency has a named owner, a fallback and a location for operational instructions. Before building System inventory, security review and a process for keeping ownership current. Start with an available tool Medium prototype Proposed concept
19 / Learning & collaboration Explore
Scenario practice builder For teams teaching recurring work decisions
Branching practice scenarios with choices, explanations and a facilitator-controlled debrief.
Explore the experiment The need People can repeat the instructions but struggle with realistic exceptions. First small experiment Turn one common exception into a three-choice story and test it with a small consenting group. What to look for Participants explain the decision and apply the reasoning to a different example. Before building Subject expertise, accessible presentation and review of any sensitive scenario. Start with an available tool Medium prototype Proposed concept
20 / Learning & collaboration Explore
Teach-back learning paths For small teams onboarding colleagues
A sequence of short practice tasks followed by a learner-created explanation or artifact.
Explore the experiment The need A completed video does not show that a colleague can do the work. First small experiment Replace one passive onboarding step with a safe task and an optional peer conversation. What to look for The learner can complete the task and identify where they still need help. Before building Supportive review, accessible alternatives and no use as an undisclosed employee ranking system. Start with an available tool Small prototype Proposed concept
21 / Learning & collaboration Try first
Team decision debrief kit For remote and cross-functional project teams
A facilitated board connecting expectations, observed outcomes and one owned adjustment.
Explore the experiment The need Retrospectives collect opinions without changing the next decision. First small experiment Run a thirty-minute debrief of a small completed project with anonymous input available. What to look for The team chooses one practical adjustment and checks it during the next comparable task. Before building A facilitator, psychological safety and a clear boundary around personnel information. Start with an available tool Small prototype Proposed concept
22 / Specialist projects Try first
Consent-aware specialist brief For teams preparing to speak with external specialists
A brief builder separating a public problem summary from information shared only after agreement.
Explore the experiment The need Sensitive documents are shared before scope or permission is clear. First small experiment Rewrite a fictional client request into public, restricted and unnecessary information. What to look for A potential specialist can assess the scope without receiving confidential material. Before building Owner approval, sharing rules and appropriate agreements before real information is disclosed. Start with an available tool Small prototype Proposed concept
23 / Specialist projects Explore
Scoped collaboration trial For teams considering a new delivery partner
A paid-trial planning template with a bounded deliverable, review points and an exit decision.
Explore the experiment The need A long engagement starts without evidence of how the parties work together. First small experiment Define a fictional one-week collaboration and ask both sides to identify ambiguous responsibilities. What to look for Both sides agree on the output, access boundaries and what the trial can and cannot prove. Before building Mutual agreement, fair compensation and relevant professional advice where needed. Start with an available tool Small prototype Proposed concept
24 / Specialist projects Later
Specialist availability and fit board For a future opt-in specialist community
A consent-based directory of self-described skills, working preferences and availability windows.
Explore the experiment The need An introduction fails when availability, constraints or expectations are unknown. First small experiment Interview a small invited group using a private prototype without publishing profiles or making matches. What to look for Participants can assess relevance, correct their details and withdraw without friction. Before building An actual participant network, profile consent, moderation and verified operating capacity. Start with an available tool Large prototype Proposed concept
