Maple Spark Labs
Founder-led AI for operations · Toronto

Find where AI earns its place in your operation.

We learn how the work actually runs, test where AI helps with the people who do it, and help put what proves useful into everyday use.

Talk through a workflowSee how it works
Where the time goes

Friction like this costs hours. It also caps how much business you can take on.

  • Work that waits

    on one person, one inbox or one approval.

  • Work done twice

    re-typed between email, spreadsheets and systems.

  • Customers who wait

    for quotes, replies and updates.

Not all of it needs AI. Sometimes the fix is a form or a rule, and we'll tell you.

How we work

We start with the work, not the tool.

Walter brings years of process and delivery practice to a three-week test: your people try a working prototype, not just a slide deck.

  1. Map the workWeek 1
    Find the handoffs and where time is lost.You getA journey and workflow map
  2. Find the optionsWeek 1 to 2
    Compare a process change, plain automation and AI.You getAn opportunity shortlist
  3. Choose what to testWeek 2
    Pick what is valuable, feasible and responsible with the resources you have.You getOne prioritized candidate
  4. Test with your peopleWeek 3
    Try a working prototype, then decide go or no-go.You getTest observations and a go or no-go decision

Want to see the work products? The sample Sprint is fictional.

See the sample Sprint's work products →
Where to start

Find what to fix, or take it live.

AI Opportunity Sprint

We know work could be better. Where should we start?

3 weeks · fixed fee
Find, choose and test one promising improvement with your people. You end with a go or no-go decision and a prototype they tried.
How the Sprint works →
Pilot to Production

We have a promising pilot. How do we make it work every day?

4 to 8 weeks · fixed fee per phase
Make it reliable, integrated with your systems, used by your people and governed. No Sprint needed first.
How to take it live →

Selected experience

  • Walter's prior experience · Slalom · Shopify

    Defining the customer before a CRM migration

    For Shopify Plus's move from HubSpot to Salesforce, Walter led a Slalom team through cross-functional discovery. A key question was how merchants, divisions and storefronts should be represented as customers in the new system. The work produced a roadmap, and Slalom was later brought back for the first migration.

    Read Walter's Slalom experience →
  • Walter's prior experience · Bitbuy

    Building required checks into the customer journey

    At Bitbuy, Walter worked with compliance and engineering to build identity and transfer requirements into the customer's journey. Behaviour data exposed an unexpected sign-up obstacle, while saved recipient details made later transfers simpler.

    Read Walter's Bitbuy experience →

Our own AI product, not client work:Droplet Learn, in the Lab

Walter Varma-Chang leads every engagement and does the core work himself, with 20+ years modernizing how work gets done, from IBM and Accenture to a 20-person delivery practice at Slalom.

See all selected workAbout Maple Spark Labs
Privacy and client data

By default, client information stays in the client's own systems and licensed tools. MSL's own AI tools work on de-identified or synthetic material. Anything else is agreed in writing first.

How we approach systems and data

Have a workflow that's slowing you down?

Send a short note. Walter will reply with times for a 30-minute call. No preparation needed.

Talk through a workflow