1. Qualification call
In 15–30 minutes we test urgency, sponsorship, budget willingness, scale, technical context and whether measurable value is likely.
Outcome: go / nurture / no-goFresh Kiwi · AI consultancy
For organisations that want a decision-ready roadmap rather than another disconnected AI pilot.
Perspective
The audit produces an AI Opportunity Portfolio: relevant tasks across the organisation scored for value, feasibility, risk and implementation effort, plus priority business cases, a pilot recommendation and management roadmap.
Opportunity audit
A typical outcome could look like this. These are example figures, not client claims.
Method
From first qualification to a management decision on the first pilot.
Capabilities
In 15–30 minutes we test urgency, sponsorship, budget willingness, scale, technical context and whether measurable value is likely.
Outcome: go / nurture / no-goWith the sponsor, operations and IT we map business goals, growth pressure, bottlenecks, headcount pressure, systems, current AI use and success criteria.
60–90 minutesWe identify relevant departments, processes and roles and where repetitive or administrative work is concentrated.
Typically 5–8 departmentsEmployees describe recurring tasks, frequency, time spent, systems, inputs, outputs and frustrations.
Often hundreds of observations in an 80-person companyWe validate promising tasks with process owners and employees and explore exceptions, quality problems and dependencies.
Typically 8–12 interviewsWe inventory data, APIs, Microsoft 365/Entra, ERP/CRM, confidentiality, personal data, contractual constraints and approval requirements.
No architecture decision before the problem is clearCapabilities
We estimate the time, cost, revenue, quality or risk improvement a better process can create.
We quantify how much employee time the task consumes each week, month and year.
We assess how broadly the task occurs and how many people benefit from improvement.
We test whether AI is appropriate or whether conventional automation is simpler and more reliable.
We assess whether the required data is available, accessible and of sufficient quality.
We map integrations, exceptions, process change and technical dependencies.
We assess confidentiality, access control, logging and required technical safeguards.
We determine whether personal data is involved and which safeguards are required.
We consider relevant regulation, contractual obligations and internal policy.
We determine where people must remain responsible for review, approval or escalation.
We compare implementation and operating cost with annual value and expected payback.
Example output
| Opportunity | Annual value | Complexity | Risk | Recommendation |
|---|---|---|---|---|
| Weekly reporting | €32k | Low | Low | Start |
| Lead research | €28k | Low | Low | Start |
| Invoice processing | €41k | Medium | Medium | Next |
| Proposal assistant | €35k | Medium | Medium | Next |
| Customer email triage | €29k | Medium | Medium | Explore |
Capabilities
Relevant opportunities clustered and prioritised into Do Now, Next, Explore and Don’t Automate.
For the best opportunities: baseline, expected value, implementation effort and payback.
Security, privacy, compliance and human oversight requirements for each priority.
A concrete recommendation for the first capability that can prove value quickly at manageable risk.
A sensible sequence for pilots, expansion, platformisation and reassessment.
A decision document for management and process owners, not a long technical report.
Next step
Understand where AI can produce measurable return before selecting tooling or building agents.