Client testimonials

What Clients Say

Candid feedback from organisations we've worked with

These are genuine accounts from clients across different industries and service types. We've kept them honest β€” including the observations about what took longer than expected.

← Back to Home
47

Client Organisations

Since 2019

4.8

Average Rating

Out of 5.0

94%

Would Recommend

Post-engagement survey

68%

Return Clients

Multiple engagements

Client Reviews

In their own words

"We came to Thoughtform with a genuinely messy problem β€” years of client feedback in mixed Cantonese and English, unstructured, no consistent format. I expected them to either oversell what AI could do or struggle with the bilingual aspect. They did neither. The discovery session was thorough, they were upfront about what was tractable, and the tool they delivered has genuinely changed how our client services team operates."

MC

Margaret Chan

Head of Client Services, Asset Management Firm

Wan Chai, HK β€” January 2026

"The AI Literacy Program was well-pitched for our team's level. We had a mix of people β€” some quite tech-comfortable, others who'd been quietly anxious about AI tools appearing in their workflow. Ada managed the room well, kept things grounded, and the materials actually reflected how our industry works rather than generic examples. The fourth session on workflow integration was particularly practical. My only note is that we could have used more time on the ethics session."

RK

Raymond Kwok

Operations Director, Healthcare Group

Kowloon Tong β€” December 2025

"We engaged Thoughtform for the decision support dashboard after a colleague recommended them. What stood out was that they were willing to tell us at the scoping stage that one of the things we'd asked for wasn't realistic with our current data. That kind of honesty is rare. We adjusted the scope, and the delivered tool has been used in our weekly planning meetings without fail for the past two months."

LT

Lillian Tam

General Manager, Logistics Company

Kwai Chung β€” January 2026

"Our legal team processes a significant volume of contract language and we'd been exploring ways to automate some of the categorisation work. Thoughtform spent time understanding the specific clauses we cared about, built a prototype we could test with real cases, and refined it through two rounds of feedback. The documentation they provided means our IT team can actually maintain it going forward. We've already engaged them for a second project."

DL

David Lau

Senior Counsel, Financial Services Firm

Central, HK β€” November 2025

"We did the Literacy Program as preparation before rolling out an AI tool company-wide. It helped our staff approach the new tool with context rather than anxiety. I'd say the program worked as intended β€” people were more willing to engage with the tool because they had a sense of what it was doing and where its limits were. The calibration between business-relevant examples and technical concepts was about right for a non-technical audience."

SP

Shirley Poon

HR Director, Media Company

Tsim Sha Tsui β€” December 2025

"The thing I appreciated most was that Rachel spent our very first call asking questions rather than presenting slides. By the time we agreed on scope, I felt like they understood our situation. The NLP solution they built categorises our incoming patient feedback in a way that would have taken our team several hours a week. That time is now spent acting on the feedback rather than sorting it. It was worth every dollar."

AT

Albert Tsui

CEO, Private Healthcare Clinic Group

Mid-Levels β€” January 2026

Case Studies

How the work played out in practice

Three client stories in more detail β€” including what didn't go exactly as planned and how we navigated it.

Case Study 01 β€” NLP Solution

Automating contract clause categorisation for a legal team in Central

Challenge

A small legal team was spending approximately three hours per day manually reviewing and tagging incoming contracts for clause type, jurisdiction, and risk category. The volume was growing faster than headcount.

Approach

We spent two weeks auditing their existing contract corpus and building a taxonomy. A prototype classifier was reviewed against a sample of 200 contracts by their team before we refined the model and built the final interface. Total engagement: six weeks.

Outcome

Categorisation time reduced by approximately 80%. The tool now handles initial tagging for roughly 90% of incoming contracts, with manual review retained for edge cases. The team uses the saved time for substantive review work.

Case Study 02 β€” Decision Support

Inventory decision support for a retail distribution operation in Kwai Chung

Challenge

Procurement decisions were being made based on gut feel and spreadsheet pattern-matching. The team suspected they were over-ordering certain categories and under-ordering others, but couldn't see the patterns clearly across 400+ SKUs.

Approach

We discovered in week one that their data had inconsistent date formatting and three years of partially duplicated records. We spent an extra week on data cleaning β€” communicated openly β€” before proceeding. The model was built and validated against their last six months of known outcomes.

Outcome

The dashboard now flags likely over-order candidates weekly. In the two months since handover, the team has adjusted purchasing on 34 lines based on dashboard recommendations β€” with an estimated 12% reduction in overstock carrying costs.

Case Study 03 β€” Literacy Program

Pre-rollout AI training for a 40-person communications team in Tsim Sha Tsui

Challenge

The organisation was introducing AI writing tools into their content workflow. Internally, there was a range of reactions β€” from enthusiasm to significant concern about role security. Leadership wanted to address the concern directly before the tools launched.

Approach

We ran the four sessions over three weeks, dedicating significant time in session three to the honest limitations and risks of AI-generated content β€” something the team needed to hear directly. Materials were built around content and communications examples, not generic software scenarios.

Outcome

Post-program survey showed 78% of participants felt confident evaluating AI-generated content critically. Tool adoption in the first month was higher than the team had seen with previous software rollouts. Three team members subsequently requested the NLP solutions service.

Get in Touch

Ready to have a similar conversation?

We're happy to discuss your specific situation β€” with no expectation that it will turn into a project. If the right fit is there, we'll know by the end of the first call.

Start a Conversation

Address

88 Queensway, Admiralty, Hong Kong

Hours

Mon–Fri 9:00 am – 6:00 pm HKT
Sat 10:00 am – 1:00 pm HKT

Professional Standing

HKPC AI Practitioner Certified

All technical practitioners hold current certification from the Hong Kong Productivity Council.

HKCS Member Organisation

Active member of the Hong Kong Computer Society since 2020.

HKCS Excellence Finalist 2023, 2024

Recognised for responsible AI delivery in professional services contexts.