Frequently Asked Questions

A: Patrick scopes the engagement, makes the methodological judgement calls, and delivers the final work. Software and AI-assisted tools may help organise material, write or test code, compare results, and prepare documentation. A separate checking step tests the output against the source data, paper, cited evidence, or agreed acceptance criteria before Patrick signs off.

For a managed consulting engagement, you brief Patrick rather than managing prompts, models, or internal tools yourself. You still review the scientific meaning of the result and remain responsible for your research decisions.

A: The one-time setup starts from US$2,000 (approximately HK$16,000). It includes a fit review, provisioning on a server you provide, configuration for the agreed use case, a practical operating guide, handover, and one workflow pilot.

The included pilot covers one recurring workflow with agreed inputs, acceptance criteria, and one correction round. Additional workflows, integrations, substantial changes of scope, or further correction rounds are quoted separately.

You pay the server and any third-party AI or model usage costs directly. Ongoing support is optional and is not included automatically.

AI research team setup is a separate fixed-fee engagement and is not covered by the trial analytics assurance. The fee covers provisioning and isolation to the agreed scope, one agreed workflow pilot, one correction round, documentation and handover.

An owned setup is suitable when your team has a repeatable workflow and someone can remain responsible for reviewing its outputs. It is not a replacement for staff where the deliverable itself depends on continuing human judgement, supervision, participant care, or institutional authority. See Own an AI Research Team.

A: Before restricted data is transferred, Patrick and the client agree what data is necessary, whether the proposed use is covered by consent, ethics approval, and institutional rules, where processing may occur, who may have access, and when working copies must be returned or deleted.

The arrangement may use your lab's environment, a client-controlled server, or another approved setup. Whether data is sent to an external software or model provider depends on the chosen architecture and that provider's terms; this must be identified before transfer. If those conditions are not acceptable, the work must be redesigned or not proceed.

Patrick does not use client data to train his own models or reuse one client's confidential material for another client. That statement does not replace review of any third-party provider's retention and training terms. For public-data trials, the public deposit can be used without transferring an unpublished dataset.

A: A public demo shows an example method on public data. A trial analytics deliverable applies one agreed workflow to a scoped dataset or paper, from US$1,000 (approximately HK$8,000). Scope and acceptance criteria are agreed before work begins; if the completed trial analytics deliverable does not meet them, the fee is waived. Small deliverables start from US$2,600 (approximately HK$20,000). Larger, grant-scale projects start from US$5,100 (approximately HK$40,000) and are normally divided into milestones.

The scope states the inputs, analyses or workflow, acceptance criteria, outputs, timetable, included review or correction rounds, and handover materials. Depending on the engagement, those materials may include scripts, reports, documentation, and generated output files. Questions needed to understand the delivered result are included; new analyses or changed requirements are scoped separately before work begins.

Grant or budget eligibility depends on your funder's and institution's rules. Patrick can provide a scope of work and fixed-fee invoice that describe the service and deliverables, but cannot guarantee that a particular office will approve the expense.

A: Five visible steps: find the public deposit behind the paper, load and sanity-check it, re-express the paper's core analyses independently, verify item-by-item against the published statistics, then report what reproduced and what didn't. The whole flow is on the Process & Pricing page.

A: A mismatch is reported rather than hidden. The report identifies what differed, what could be checked, and what remains unresolved. For the qualifying trial analytics deliverable, if the completed deliverable does not meet the agreed acceptance criteria, the fee is waived.

A: Scope and acceptance criteria are agreed before work begins. If the completed trial analytics deliverable does not meet those agreed acceptance criteria, the fee is waived. This assurance applies only to the trial analytics deliverable; it does not apply to AI research team setup or other services.

A: Yes. Research retainers start from US$1,300 per month (approximately HK$10,000). The monthly scope, capacity, reporting, term, and notice arrangements are agreed in writing. A retainer is optional and is not created automatically by buying a trial or deliverable.

A: No. Trials and fixed-fee deliverables are one-off engagements. Ongoing support is a separate written agreement; nothing automatically converts into a retainer or subscription.

A: No. You may begin by email with a short description of the work. Patrick reviews the information before confirming scope or price. A call may be suggested where the data, method, privacy requirements, or acceptance criteria need discussion.

A: It can, if the written scope defines each project as a separate workstream and the reporting method is suitable for your lab, research group, department, or institution. Whether costs may be divided across grants is a question for the relevant research or finance office.

A: No — fixed fees are agreed before work starts. Internal effort estimates are used for scoping only and are never converted into hourly client billing; retainers are monthly service fees with scope-based work.

A: Both, at different levels. Every deliverable comes with a technical report — what the numbers show, significance, effect sizes, and the caveats. That is my job, and it is written for you to judge. The scientific interpretation — what the findings mean for your theory and your next paper — is yours; that's your expertise. If you'd like drafting help for a results or discussion section, that is scoped as a separate piece.

A: The tools do not determine whether a paper is publishable. Editors, reviewers, and institutions may care about methodological validity, reporting, reproducibility, disclosure, data governance, and compliance with their current policies.

The deliverable identifies the relevant software, versions, assumptions, and checks where these are in scope. You remain responsible for the target journal's rules, authorship and disclosure decisions, and the scientific claims made in a submission.

A: You remain responsible for the research question, ethics and consent, lawful access to the data, scientific interpretation, authorship, institutional approvals, and final submission decisions. Patrick is responsible for the services, checks, and handover materials expressly agreed in scope. Tools do not take academic or institutional responsibility.

A: No. For a managed engagement, Patrick operates the agreed tools and explains the delivered result. If your team will operate an owned setup or rerun a pipeline, the required access, software, training, and ongoing responsibilities are stated in the scope.

A: Yes, where the task can be clearly scoped and the client retains approval authority. Admin support may include drafting, document preparation, meeting summaries, or inbox triage; no message, submission, or institutional action is sent without the agreed human approval step. Admin retainers start from US$640 per month (approximately HK$5,000).

Teaching work may include reviewing and restructuring a lecture deck, updating references, revising explanations, and preparing slides or speaker notes. It is quoted by deck or course scope.

Budget eligibility depends on the relevant department, institution, and funding rules. Admin and teaching support are not covered by the trial analytics assurance. Any separately quoted qualifying trial analytics deliverable must be expressly identified in writing.

A: The site provides public-data demonstrations and per-project materials that can be inspected directly. Patrick holds a PhD in psycholinguistics from UNSW and a BA in linguistics from CUHK, and has worked in university research settings including EdUHK.

Patrick works remotely on Hong Kong hours. Any required campus visit is agreed and priced separately. Current demonstration counts, response times, and availability should be taken from the live site or written proposal rather than treated as permanent contractual terms in this FAQ.