Operating record
LOKAL has operated since 2017 and has served more than 65 clients. That establishes organisational history and multi-client delivery; it is not rewritten as a count of AI transformations.
AI delivery evidence
The evidence is smaller than the story. That is the point.
For buyers who want to know which AI claims can survive procurement, not just which ones fit in a pitch deck.
A striking percentage becomes useless once nobody can find the denominator, measurement window or permission. This page keeps the current public record and the evidence boundary in the same place.
The short answer
What does LOKAL’s public AI evidence prove?
LOKAL’s public record proves an operating company established in 2017, work across more than 65 clients, Claude architecture capability and founder participation in the OpenAI Champions Network. Older enterprise adoption percentages are not included because the available evidence does not yet reconcile their dates, denominators and public-use permission.
A striking percentage becomes useless once nobody can find the denominator, measurement window or permission. This page keeps the current public record and the evidence boundary in the same place.
What the work contains
Useful artefacts, not an AI theatre programme
Technical capability
Claude Certified Architect — Foundations capability and OpenAI community participation support technical and ecosystem familiarity. Neither credential is described as a vendor endorsement of a client engagement.
Delivery artefacts
A scoped AI engagement can expose the readiness score, workflow map, architecture decision, risk control, implementation record, training material and measurement method produced for that work.
Claims on hold
LOKAL’s older enterprise programme figures are excluded from new copy until source files resolve the dates, denominator, measurement-window conflict and exact public permission. Familiarity is not clearance.
How it moves
A decision sequence the operating team can follow
- 01
Register the claim
Record the exact wording, number, client status, channel and page family where it may be used.
- 02
Attach the method
Keep the baseline, intervention, denominator, date range, measurement method and source artefact beside the result.
- 03
Confirm permission
Separate named, anonymised and private-reference use. Permission for one proposal does not become permission for the public site.
- 04
Publish the limit
State what the evidence demonstrates and what it cannot establish. Remove a result when its method or permission cannot be reconstructed.
What you can verify
Capability, evidence and limits kept together
Since 2017
An established operating company
LOKAL’s history predates the current AI-services category. The date is useful context, not a claim that all those years were spent delivering the same programme.
65+ clients
Multi-client operating experience
The public count covers LOKAL’s company-wide work. It is not presented as 65 AI clients or 65 adoption programmes.
CCA-F
Claude architecture capability
The credential is named at its actual level. LOKAL does not expand it into an Anthropic partnership or endorsement.
Founder member
OpenAI Champions Network
The membership belongs to LOKAL’s founder. LOKAL is not described as an OpenAI partner.
Before the scope
The useful no
- No enterprise participant-count figure appears in new copy until the source-review gate clears.
- No weekly, daily or satisfaction percentage appears without a reconciled denominator and measurement window.
- A credential proves a credential; it does not prove a client outcome.
- A workshop photo, client logo or private reference is used only within its specific permission.
Questions to settle before anyone buys a tool
Why are some older LOKAL AI figures missing?
The surviving evidence conflicts on the method behind the older enterprise metrics. LOKAL is holding those figures out of public copy until the dates, denominator, measurement method and permission are reconciled.
Can LOKAL provide a private AI reference?
Possibly, where the client has permitted it and the proposed work makes the request relevant. A private reference is arranged deliberately and does not imply permission to publish the relationship.
What evidence comes with an AI readiness assessment?
The buyer receives the inspected inputs, scoring logic, ranked workflow map, open questions, gap register and proposed sequence. The evidence shows why a workflow was ranked, delayed or rejected.
Do credentials guarantee implementation quality?
No. Credentials show defined training or assessment capability. Scope quality, architecture decisions, QA, user adoption and operating outcomes still need evidence from the engagement itself.
When will outcome case studies be added?
When the claim register contains a defensible baseline, intervention, period, denominator, source artefact and public-use permission. The page count does not set that timetable.
A useful first brief
Bring one workflow that is slow, expensive or hard to control
We’ll look at the work, systems, data and risk around it. If an assessment is premature—or AI is the wrong answer—we’ll say so before proposing a programme.