Casebook · Education & Training · China

Study-abroad consultancy pilot: 18 scattered student files indexed into 28 searchable records

Pilot · DemoRecords as of 9 Oct 2026BEE Sigma Delivery Team

The client is a study-abroad consultancy in China with planners, growth coaches and academic coaches working on each student. We connected AI agents to three tools, Lark Docs, Lark Drive and Lark Base, to turn one student's service records into a searchable profile. From 2026-08-10, 18 source files plus 10 team notes for the first pilot student were indexed into 28 structured records. This is still a pilot and demo, in the co-design phase with the firm.

IndustryEducation & training (study-abroad consulting)
MarketChina
Delivery statusPilot · Demo (co-design phase)
TimingFirst pilot index built 2026-08-10; records to 2026-09-10
Systems connectedLark Docs, Lark Drive, Lark Base
Decisions kept by peopleProgramme design, SOP content, educational approach and professional judgement
01

How it worked before

A student's service history sat in PDFs and Word files: assessment reports, notes from each growth-coaching session, academic mentor feedback and a three-year plan. There was no single record, so each hand-off between planner, growth coach and academic coach meant digging through folders to see what had been done and what came next.

Before and after
Before and after
02

What AI does now, and what people do

We used one student's records from close to a full school year as the pilot, structuring them first before asking whether the approach could scale.

  • AI does: Reads each source file and writes a summary; builds a Lark Base index where every record holds file name, category, service stage, summary and intended use; pulls up the relevant files by stage for coaches.
  • People do: The firm proposes and confirms programme design, SOP content and professional judgement. Anything involving educational approach, business decisions or staffing is confirmed by the firm; AI only structures and drafts.
Process: what the AI does and what people do (gold = people)
Process: what the AI does and what people do (gold = people)
03

Timeline

All dates come from project records.

  • 2026-08-10The firm shared 18 source files for the first pilot student; the team built the index table
  • 2026-08-17 to 08-21Cross-organisation file access resolved by switching to email-based sharing
  • 2026-08-31Kick-off meeting defined four roles: planner, growth coach, academic coach, AI coach
  • 2026-09-03Service period set as 2026-07-01 to 12-31, with the first stage review on 2026-09-30
  • 2026-09-07 to 09-09Standard course plan completed
Timeline: every point comes from project records
Timeline: every point comes from project records
04

Current status and limits

This is a pilot and demo. There is one pilot student and 28 index records (18 source files plus 10 team notes, as of 2026-08-10). Whether the profile can be extended to more students, and how it fits into daily service, is still being co-designed; as of the records to 2026-09-10 there is no conclusion. No student assessment results or personal details appear here.

05

What we learned

Our first index held only file names and summaries, with no link to the full text: it looked structured, but coaches could not reach the original. We now link every record to its source file, and we set up cross-organisation file access before collecting documents rather than fixing permissions file by file afterwards.

Scope note: an anonymised case is a summary of project records with client names, people and amounts removed. This page draws on those records; client names, people and amounts are withheld. Dates and figures come from project records and some have not been confirmed in writing by the client. The delivery status is as labelled at the top; results vary with each business’s systems, data and involvement and are not a promise to other businesses.

FAQ · Quick answers

Does the AI make planning decisions for the growth coaches?

No. Programme design, SOP content and professional judgement are proposed and confirmed by the firm. The AI summarises each file, builds the Lark Base index and pulls up files by stage for coaches.

Is the student profile in use for all students?

No. It is a pilot and demo with one student; 28 index records were built on 2026-08-10. Extending it to more students is still being co-designed, as of records to 2026-09-10.

How is student privacy handled?

Files are organised inside the firm's own Lark workspace, with cross-organisation access granted by email-based sharing. Public write-ups omit student names, assessment results, grades and family details and describe only the process.
More anonymised cases
← All anonymised cases

Hand your first workflow to the agents

One free scan shows how visible you are in the AI era; the AIM assessment finds your best angle of adoption.