Five capabilities, mapped to the arc of an engagement.
Good strategy is the foundation, not the finish line. The second half starts after the deck is approved, where the interesting question stops being “what should we do?” and becomes “who owns it on Monday, how will we know it worked, and what happens when it’s wrong?”
Decide
Strategy & sequencing
A sprawling backlog turned into a sequenced, costed plan a CFO will fund, including a straight answer on which problems genuinely need AI and which are better solved without it.
Build
AI where it earns its place
Retrieval over your own institutional knowledge and agents embedded where the work already happens: scoped narrowly, evaluated honestly, and kept under human review where the stakes require it.
Prove
Decision science & measurement
Segmentation, offer and pricing models, identity resolution and real-time activation, tying a change to a number an executive will defend in a board meeting.
Sustain
Operating at scale
Once a model works in one place, this phase takes it everywhere it should run: rolled out across the business, with value tracked, not assumed, so it keeps compounding after the engagement ends.
Data & AI Advisory
Senior judgement on retainer, for organisations that need it before they need the headcount.
Jason Lam
Fifteen years as a strategy consultant to large organisations across aviation, quick-service restaurants, financial services, telecom, media and the public sector, in roles at Publicis Sapient, Accenture Song, RSA Canada, Rogers, iProspect/Aegis in Hong Kong, and Microsoft, with a career spent unusually close to delivery: the roadmap and the thing that actually shipped.
That has meant designing a Canadian airline’s experimentation practice from a standing start; leading personalisation and pricing science for a global QSR brand; layering agentic AI onto a global beverage company’s experimentation centre of excellence, with retrieval across thousands of internal reports and design documents so campaign teams get recommendations inside their own workflow; and leading an industry-first audience segmentation and attribution platform for a U.S. broadcast media company.
A larger firm sells senior judgement in the pitch, then hands delivery to a team still learning the technology alongside you, and prices for the coordination that requires. AI moves fast enough that this gap matters: the discipline of fifteen years has to be paired with the agility to keep pace with where the technology actually is, not filtered through a staffing pyramid several rungs removed from the tools. JL Digital applies that judgement directly and continuously, by the person who has it.
What follows is Jason’s own track record, earned leading these programmes personally inside client organisations, not a result JL Digital claims as a firm.
JL Digital ships its own software.
Two products built, operated and supported by the company. Client advisory sharpens the thinking; these are where the consequences are ours.
A platform of AI agents built for how a real estate practice actually runs: reading incoming rental applications, checking legal history, flagging missing documents, and drafting follow-ups, so the operational grind runs itself.
It accelerates the parts of a deal that don’t need a licensed decision, and stays out of the way of the parts that do: the calls that carry liability stay with the agent.
Most apps in this category teach recognition. Bamboo Academy teaches a child to actually write: guided stroke tracing built on real Hong Kong EDB stroke order, checked stroke by stroke.
An AI model tracks which characters a child keeps confusing, resurfaces them in spaced review, and turns the pattern into a plain-language progress report for a parent who may not read Chinese themselves.
The company behind the practice.
JL Digital Ltd. is the operating entity for the consulting practice and the products it ships.
Let’s get to the second half.
Consulting enquiries, partnerships and product support all reach Jason directly.