Follow-up that keeps its own record.
Buyer enquiries sat across spreadsheets, inboxes and chat, so follow-up depended on memory and the CRM lagged behind the conversation.
SML · We start with your business
Business consulting. Your own AI staff. Built and engineered by AI.
SML is a Hong Kong business consulting firm that designs custom AI systems around how your company actually works — then hands them over to your staff. You hold your own accounts and vendor licences. Your data stays under your control.
Our mission
Turn AI into working staff your team runs and controls.
We exist to remove the routine work that slows good teams down. We do it by understanding the business first, building only what has a clear business case, and handing over a system your own people can run, approve and improve.
Proof
Our record so far.
3AI systems delivered. 5 more in discussion.
2023Founded
7Industries served
MultilingualThai, Japanese, Russian and more. We work with your team in English.
What we do
Six service lines. One system.
AI consulting sets the direction; five build services deliver it. Each has a written scope of work, so you know exactly what is delivered and who does what.
00
AI Consulting
Business-first diagnosis and roadmap that decides where AI should — and should not — be used.
Scope of work →01
System Integration
Connect the tools you already pay for so data moves once, correctly, without re-typing.
Scope of work →02
Database Build & Data Mining
Turn scattered lists, inboxes and spreadsheets into one clean, searchable database.
Scope of work →03
Batch Outreach (Email & WhatsApp)
Personalised email and WhatsApp outreach at volume — with your staff approving every send.
Scope of work →04
CRM Automation (HubSpot)
A HubSpot set-up your team actually uses: clean pipelines, automated tasks, reliable reports.
Scope of work →05
Social Media Seeding
A planned, approved system for placing your content where your audience already talks.
Scope of work →Problems we solve
Where work stalls, and what we build.
The problems we hear most often, grouped by service line, and what we do about each.
AI Consulting Scope of work →
AI trials that stay as demos
We start from a commercial process with an owner and an agreed before/after, not from a tool demo.
Too many tools, no single plan
A written roadmap ranks opportunities by effort, risk and value, and names which tools you licence yourself.
Staff unsure what AI may or may not do
We set clear rules: what is automated, what is drafted for approval, and what stays fully human.
System Integration Scope of work →
Staff copy data between systems by hand
Automated, logged hand-offs between your existing tools, with error alerts instead of silent failures.
Every tool has its own version of the truth
A defined system of record for each data type, and sync rules everyone can read.
Integrations nobody inside the company understands
Plain-language runbooks and diagrams handed over to your staff, plus a named internal owner.
Database Build & Data Mining Scope of work →
Duplicate and incomplete contact records
Rules-based and AI-assisted de-duplication, standard formats for names, phones and companies.
Valuable history trapped in email and chat
Structured extraction of contacts, companies and interactions into defined fields.
Nobody trusts the numbers in reports
A documented data model, field definitions and quality checks that run on a schedule.
Batch Outreach (Email & WhatsApp) Scope of work →
Follow-ups slip because staff are busy
Drafted, personalised messages queued daily for review, ready for your team to approve.
Mass messages feel generic and get ignored
Segment-based templates plus AI personalisation from the fields in your own database.
Risk of sending to the wrong person or without consent
Consent and opt-out fields checked before a message can be queued; human approval on every send.
CRM Automation (HubSpot) Scope of work →
CRM data is incomplete, so reports are unreliable
Required fields, automated enrichment and AI-assisted logging from email and chat.
Leads wait days for a first response
Routing rules and task automation that put every new lead in front of an owner quickly.
Pipelines do not match how you sell
Stages, properties and views redesigned around your process, documented in plain language.
Social Media Seeding Scope of work →
Posting is irregular and depends on one person
A calendar and drafting workflow that keeps output steady, with approval in one place.
Content does not reach the right communities
Mapped channels, groups and partners relevant to each audience segment.
Hard to tell what is working
Simple tracking of posts, placements and responses fed back into the plan.
How we work
Discover. Prove. Build. Hand over. Care.
Every engagement follows the same five phases, each with a deliverable and a decision point for you.
- 01Discover
We map the process, data and people involved, and agree what success looks like.
- 02PoC
A small working version on real samples, so you can judge fit before a full build.
- 03Build
We build, integrate and test the full system with your team in the loop.
- 04Handover
Training and runbooks go to your staff; accounts stay in your name. You run it.
- 05Care
An optional monthly Care Plan for new workflows, fixes and fine-tuning.
Our advantage
Why clients work with SML.
01
Consulting before code
We begin with the commercial problem and the people who own it. Tools are chosen to fit the process, never the reverse.
02
Built to be handed over
Every system ships with runbooks and training, on accounts you already hold. Your staff run it; you are not locked in to us.
03
Human approval where it matters
AI drafts, sorts and prepares. Your people approve anything that leaves the company or changes a key record.
04
Your licences, your data
You hold your own accounts and vendor licences. Your data stays under your control. We recommend and order the software for you, in your name.
05
One partner across the stack
Consulting, integration, data, outreach, CRM and social seeding, designed to work as one system.
06
We run on our own custom logic.
The decision rules are ours. They run on proven, mature software, with a person approving every external send.
Frequently asked questions
What we do and how we work.
What does SML do?
SML is a Hong Kong-based business consulting firm that designs and builds custom AI systems for companies — then hands them over to the client's own staff to run.
Which services are available?
AI consulting, system integration, database build and data mining, batch outreach by email and WhatsApp, CRM automation on HubSpot, and social media seeding. Each is delivered through the same five phases: Discover, PoC, Build, Handover and Care.
How is SML different?
We lead with business consulting rather than software sales, we design every system to be run by your staff, you keep your own licences and data, and a human approves every external send.
How do you handle our data and systems?
Credentials and data stay in accounts you own. We work with the minimum access needed, back up before bulk changes, document what we build, and design data handling with the PDPO (Cap. 486) in mind.
Testimonials
Real friction. Real systems.
Anonymised client stories. Each names the friction and what we built or are building to fix it.
AI staff the team runs and approves.
Daily briefings, client records, outreach drafts, lead routing, scheduling and monthly reports were still assembled by hand, across tools the team already used.
One client view for a broker team.
Client records sat in separate systems, so monthly reporting and follow-up meant finding the same people by hand.
AI for growth, social media and business development.
The firm is looking at AI to grow the business, to build its social media, and to set a business development strategy.
Office
Hong Kong.
We work with clients in person in Hong Kong and remotely across the region.
- General enquiries: info@smlhk.com
- Consultations: ask@smlhk.com
- WhatsApp: +852 9268 8195
- Office hours: 24/7
- Where we work: HKG | SIN | BKK | TYO
Next step
Let us map where AI fits in your process.
Start with a consultation. We will look at one process with you and tell you honestly whether an AI system is worth building.