Client Experiences
What teams say
after the work.
These accounts come from organisations that brought us text data challenges and left with working systems — and teams better equipped to maintain and extend them.
Back to HomeEngagements
Avg. Rating
Client Retention
Years Practice
Client Reviews
From the teams we've worked with.
"We had been manually triaging support tickets for three years. The classification system reduced that load by around 70%. What stood out most was how clearly they'd documented the model's weaker areas — it helped us set realistic expectations internally."
Tan Li Wei
Head of Operations · Fintech, SG
February 2025
"I'd recommend starting with the feasibility review. Their written output answered questions we hadn't thought to ask — including ones that would have caused problems had we gone straight to development. The cost was a fraction of what we'd have spent going in blind."
Aisha Mohamad
Digital Transformation Lead · Healthcare, SG
January 2025
"Our policy documents span English and Mandarin. The summaries needed to hold up with senior management. The team's explanation of where the system struggles was actually more reassuring than a claim of perfection would have been."
Chen Kai Qiang
Policy Analyst · Public Sector, SG
February 2025
"The handover period made the real difference. My team isn't composed of NLP specialists, but after two weeks working through the system with them, we understood it well enough to retrain it ourselves when we expanded the taxonomy three months later."
Priya Ramasamy
Engineering Manager · Logistics Platform, SG
December 2024
"We had a tight timeline and modest budget. The feasibility review gave us a clear picture of what was achievable within those constraints — direct, no padding. We've since budgeted for the full classification system based on what we learned."
Jason Lim
CTO · Legal-tech Startup, SG
January 2025
"Our transcripts mix English and Malay. Most tools I'd evaluated handled this poorly. The ambiguity flagging was particularly valuable — it surfaced exactly the passages where a human reader needed to be in the loop."
Nurul Ain Binte Yusof
Research Lead · Regional Consultancy, SG
February 2025
Project Accounts
Three engagements, in detail.
Contract clause tagging for a Singapore commercial law firm
Challenge
The firm processed 300–500 contracts monthly. Junior associates spent significant time on manual clause categorisation — obligation, liability, termination — rather than substantive legal review.
Solution
A text classification system trained on the firm's annotated contract corpus. Low-confidence outputs were flagged for human review rather than passed through silently.
Results
65% reduction in time spent on initial clause tagging. Associates redirected that time to substantive review. Internal team maintained the system with one retraining cycle in six months.
"The confidence thresholds were set conservatively — more items flagged for review than we'd hoped initially. But the items we accepted directly were ones we could actually rely on. That's the right trade-off."
— Senior Associate, Commercial Practice
Meeting transcript summarisation for a government agency
Challenge
A department managing multi-stakeholder consultations accumulated transcripts in English and Mandarin. Officers spent hours extracting action items and key positions before each session could be circulated.
Solution
A multilingual summarisation workflow producing structured summaries with flagged ambiguities. English-Mandarin code-switching handled explicitly rather than treated as noise.
Results
Draft summaries circulated within 30 minutes of transcript receipt. Officers validated summaries rather than producing them from scratch — estimated 3 hours saved per session cycle.
"The ambiguity flags helped us catch the passages where the model was uncertain. We could review those specifically, rather than reading every summary end to end. That's a much more efficient use of our time."
— Senior Policy Officer
NLP feasibility review for a regional logistics operator
Challenge
A regional logistics firm wanted to automate routing decisions from free-text shipping instructions. Internal estimates suggested significant development investment. No one had assessed whether their data could actually support the task.
Solution
An NLP Feasibility Review reviewing 2,000 shipping instruction samples. Two viable approaches identified; a third ruled out due to data inconsistency. Written output included realistic performance estimates and risks.
Results
The firm chose the lower-cost approach outlined in the review, reducing their development budget by 40%. The written assessment was used directly in an internal business case presentation.
"We came in with one idea and left with a clearer one. The review wasn't a sales pitch for more work — it was a genuine technical assessment. That's what made it useful."
— Head of Technology, Regional Logistics Operator
Get in Touch
Ready to discuss
your project?
We're happy to have a direct conversation about your data and what you're hoping to do with it. No obligation — just a technical discussion to see whether there's a fit.
Phone
+65 6751 3982Address
3 Fusionopolis Way, #09-21
Symbiosis, Singapore 138633
Working Hours
Mon – Fri: 9:00 AM – 6:00 PM SGT
Weekends: By appointment
Next Step
Your project could be
the next one we discuss.
Whether you have a defined brief or an open question, a conversation costs nothing and usually produces useful clarity.
Get in Touch