LOKAL Book a consultation
Sydney and Manila, since 2017

Responsible use of AI

How LOKAL uses AI

We use AI in most of our work. It makes the work better. Ultimately, a person is still responsible and accountable for every output and every decision.

Run on the AI we built for ourselves.
9 AI tools, every account on terms that exclude training
~80% Of our own work runs on LOKI, our AI system
1 Named person responsible for every decision
30 days Ely guest-chat logs kept, or longer where the law requires

Owned by Joshua Pielago, Founder · Last updated

What we use AI for

Research, audits, drafts, reporting and quality checks.

We use AI for research, audits, first drafts, reporting and quality checks, with LOKI as the system under every account.

LOKI is LOKAL’s in-house AI system. About 80% of our own work runs on it: 433 tools and 19 skills in the July 2026 inventory. Read the internal case ↗

A senior strategist owns every plan, and a person is responsible for every decision.

What we don’t use it for

  • Final judgement on strategy.The strategist decides what a business needs, what comes first and what to stop.
  • Client-facing decisions without a person.An AI never settles a question with a client on its own.
  • Anything that publishes without approval.Nothing goes live until a named person has signed it off.
  • High-stakes steps unattended.Budget changes, account access and anything that touches money or personal information have a person at the controls.

Tools and account types

The tools we use, and the accounts we use them under.

We use business, team and enterprise accounts only, under terms that stop the AI companies training on client data.

AI tools in use, October 2026
ToolAccount typeWhat we use it forTrains on data?
Claude (Anthropic)Claude TeamResearch, drafting, audits, QANo
ChatGPT (OpenAI)ChatGPT TeamDrafting, analysisNo
Gemini (Google)Gemini EnterpriseResearch, summarisingNo
Grok (xAI)Grok for BusinessResearch, analysisNo
LOKI (LOKAL’s own system)Blended enterprise API accessThe system under every accountNo
CanvaCanva TeamsDesign and concept visualsNo
ElevenLabsEnterpriseVoice and audioNo
RunwayEnterpriseVideoNo
Black Forest Labs (FLUX)Business accountImage generationNo

Account terms reviewed on 1 October 2026. Each account sits on a tier whose terms exclude training on our content.

Client data rules

What goes into an AI tool, and what never does.

Client material goes into AI tools only when the work needs it, and it stays inside that client’s account.

What goes in

  • Brand guidelines and tone of voice.
  • Approved content: pages, posts, ads and emails the client has already signed off.
  • Research we have gathered for the client’s plan.
  • Campaign data the client has agreed we can work with.

What never goes in

  • Payment details of any kind.
  • Passwords, access keys and login details.
  • Guest or customer personal information beyond what the agreed work needs.

Each client’s material stays inside that client’s account. We never reuse it to prospect, never carry it into another client’s work, and never use it to train a model. The same rules are written into our partner protection ↗ page, our terms ↗ and our privacy policy ↗.

Human review

Where a person checks and signs off.

A person checks and signs off at four points before anything reaches a client or the public.

  1. The strategist owns the plan.

    A senior strategist sets the priorities and decides what AI-assisted work is for.

  2. A specialist reviews every AI-assisted output.

    Drafts, audits and reports are read and corrected by the person whose name is on the work.

  3. The client approves before anything publishes.

    Nothing goes live on a client’s channels without their sign-off.

  4. A named person on every decision.

    If something goes wrong, there is one person who answers for it, not a tool.

Ely

How our guest-chat product behaves.

Ely is LOKAL’s AI guest-chat product for hotels. It is live in client hotels, and the hotel waitlist is open.

How Ely behaves

  • Ely answers only from information the hotel has approved. If the approved information does not cover a question, Ely says so.
  • Ely tells guests it is an AI assistant at the start of the conversation.
  • Guest conversations are not used to train any model.
  • Ely hands the conversation to hotel staff when it does not know the answer or is unsure.
  • Chat logs are kept for 30 days, or longer where the law requires it.
  • Corrections: the hotel edits its approved information itself, and Ely answers from the updated information.

Where data is processed

Sydney, Manila and the regions each tool runs in.

Our teams work from Sydney and Manila, and the AI tools we use process data in each tool company’s own regions.

The AI tools process data in the regions run by the AI companies, mainly the United States and the European Union.

Our clients are in Australia, Canada, the United States and the Philippines, and our team works from Sydney and Manila. Client data therefore moves from the client’s own country to Australia and the Philippines, and through the AI tools above; no other country is involved. Our privacy policy ↗ covers cross-border transfers in full.

Frameworks we align with

Four public frameworks we align with.

We align our practice with four public frameworks. This is alignment, not a badge or an audit.

Australia’s Guidance for AI Adoption.

Published by the National AI Centre on 21 October 2025, it superseded the Voluntary AI Safety Standard. We align our work with its six essential practices: decide who is accountable, understand impacts and plan accordingly, measure and manage risks, share essential information, test and monitor, and maintain human control.

Australia’s Privacy Act and OAIC guidance.

The Privacy Act’s automated decision-making transparency obligation commences on 10 December 2026, and the OAIC’s October 2024 guidance on commercially available AI products says public-facing AI tools such as chatbots should be clearly identified as AI.

Canada’s PIPEDA and Quebec’s Law 25.

Law 25 requires notice when a decision is based exclusively on automated processing and gives people a right to have it reviewed by a person, which is why a person is responsible for every decision in our work.

The EU AI Act, Article 50.

Its transparency duties, under which people must be told they are talking to an AI, apply from 2 August 2026, and every AI that faces a client or guest on our side says it is AI.

Partner networks

The networks we belong to.

LOKAL is a member of the Claude Partner Network and the OpenAI Partner Network.

Our founder, Joshua Pielago, is a member of the OpenAI Champions Network.

Meet the team behind the practice ↗

What we will never do

The lines we hold on every account.

Five commitments that hold on every account, whatever the tool.

  1. Pass off AI content as human-written where disclosure is expected.

  2. Train models on client data.

  3. Run high-stakes steps unattended.

  4. Let an AI face a client or guest without saying it is AI.

  5. Mix one client’s account material with another’s.

Questions about AI

Ask the person who owns this page.

Joshua Pielago, our founder, owns this page and answers questions about how we use AI.

Email [email protected] · Contact us ↗

Common questions

How we use AI, answered.

Does LOKAL use AI on my account?

Yes. About 80% of our own work runs on LOKI, our in-house AI system, for research, audits, first drafts, reporting and quality checks. A senior strategist owns every plan, and a person is responsible for every decision. Read the internal case ↗

Will my data be used to train an AI model?

No. We use business, team and enterprise accounts only, under terms that stop the AI companies training on client data. We never use client material to train a model ourselves, and it never leaves the client’s own account. See partner protection ↗.

Who is responsible when AI gets something wrong?

A person. Every AI-assisted output is reviewed by the specialist whose name is on the work, every plan is owned by a senior strategist, and nothing publishes without client approval. If something is wrong, one named person answers for it and fixes it.

Will I know when I am talking to an AI?

Yes. Ely tells hotel guests it is an AI assistant at the start of the conversation, and no AI faces a client or guest on our side without saying it is AI. Meet Ely ↗

Which partner programmes is LOKAL part of?

LOKAL is a member of the Claude Partner Network and the OpenAI Partner Network, and our founder, Joshua Pielago, is a member of the OpenAI Champions Network.

Where can I read the full policies?

The privacy policy ↗ covers personal information and cross-border transfers, the terms ↗ cover the engagement, and partner protection ↗ covers how we keep each client’s account separate.