Frequently asked questions
Pricing, data security, risk boundaries, how to start — straight answers, no hedging. If something is missing, add me on WeChat (caishide) and ask.
The 12 questions we get most
What is an FDE, and how is it different from a consultant or an outsourcing team?
A Forward Deployed Engineer works inside your operations to move AI from demo to production. The difference is the deliverable: consultants deliver recommendations and slides, outsourcing teams deliver features by the hour, an FDE delivers verifiable running results in stages — you can stop at the end of any stage, and the next one starts only after acceptance criteria are met.
What kind of clients do you work with?
Mostly small and mid-sized manufacturers and export teams: a concrete pain point (inquiry triage, product data retrieval, quoting, cross-system sync), at least one hour a day spent on it, and a willingness to validate small before scaling.
How does billing work?
Fees for discovery and pilots are fixed in writing before signing, itemized in a quotation covering a process map, interview notes, a data/system/permission inventory, pain-point priorities, risks and dependencies, one minimal pilot proposal and a draft acceptance spec. The first call is 30 minutes free, and discovery fees can be partially credited toward a later pilot contract.
How long until we see results?
A minimal pilot usually runs in weeks: we set a baseline and acceptance metric first (for example, inquiry triage from 3 hours/day down to 30 minutes), run it with human review, and only then talk about production. We do not promise overnight results — we promise verifiable evidence at every stage.
How is data security handled? Will my customer data be used for training?
Assessed and configured per project: sensitive-field masking, least privilege, local or private deployment options, and human approval kept for key outbound actions (sending emails, quotes, orders). Specific measures go into the SOW, and we verify your data licensing before touching anything.
Do you need to be on-site? How does remote collaboration work?
The discovery phase is best done on-site, seeing the real workflow (same-day reachable within Dongguan and surrounding areas). Pilot and integration phases can run remotely with periodic visits; the rhythm is written into the contract.
What if the AI solution does not work? Who carries the risk?
Proving whether it is worth doing is the entire point of the pilot. If targets are missed, we do not move to implementation — you keep the results-and-limitations report, and exposure stays within the discovery and pilot scope. We say so plainly when a scenario is not a fit.
We have no IT team. Can we still work with you?
Yes — most clients have no dedicated IT staff. What we need from you: one person who knows the business as the contact, and access to the necessary systems and data. The engineering side is on me, and deliverables include documentation, training and handover.
What tools do you use? Will we be locked into one SaaS vendor?
Chosen per scenario: LLM APIs, automation/RPA platforms, and the interfaces of your existing CRM, email and knowledge base. The principle is to use what you already have and not introduce new systems just to sell tools. We do not promise every system can be integrated — API availability, vendor limits and data licensing are verified first.
How do we start? What is the first step?
Add me on WeChat (caishide) or leave a message on the contact page with three things: your industry, your most time-consuming workflow, and roughly how much time it costs daily. After a free 30-minute call, you get my first read on whether it is worth doing and where to start.
What is the essential difference between an FDE and an on-site outsourcing team billed by person-day?
Three differences: (1) billing — on-site teams bill by the hour; an FDE delivers in stages and is accepted against results, and you can stop at the end of any stage; (2) foundation — on-site teams write everything from scratch; an FDE brings an engineering method, an evaluation system and delivery discipline; (3) ending — on-site staffing drags on and stops when the people leave; an FDE leaves when the job is done, with capability staying in the system and your team. Put plainly: on-site teams are accountable for attendance; an FDE is accountable for whether your business actually got better.
After the project ends and you leave, who maintains the system?
The finish line is your team running it independently, not our exit. Concretely: documentation, training and handover completed before acceptance, checked item by item against a handover checklist; your key people trained into internal instructors and first-line support; the evaluation set and test bank stay with you, so you can catch quality slippage yourself. Renewal and ongoing support are voluntary options, not lock-in — not needing us anymore is exactly what success looks like.
Have a different question?
A free 30-minute call: first a read on whether it is worth doing, and where to start.