“AI-First” – Yes but AI isn’t a strategy document.
Approach matters…
Matured organisations are building one specific use case at a time and scale up gradually
1. Start Where It Hurts. Not Where comfort lies
i. CBD (UAE) tackled retail banking delinquencies with an AI risk model
–> ~20% drop in NPLs, faster collections
ii. KreditBee (India) used behavioral ML to underwrite young, credit-invisible borrowers
–> NPLs down 15%, disbursals in 10 minutes
iii. Tokopedia (Indonesia) built AI squads for cart drop-offs and returns
–> Repurchase rates up 18%, complaints down 25%
Don’t “launch AI.” Solve something real. Fast
2. AI is Not a System. It’s a Set of Tiny Superpowers
i. Gulf International Bank’s meem (Bahrain) went modular – AI for onboarding, fraud, and CX
–> Onboarding time down 40%, CSAT up 25%
ii. Razorpay (India) used AI to automate fraud checks and compliance
–> $3M saved in ops cost, onboarding TAT cut by 60%
iii. Lemonade (US) (not a Fortune 500 giant, but a disruption cannon)
–> 30% claims processed in under 3 minutes. AI bots ran ops like clockwork
Start stacking these AI micro-skills, like Legos. One at a time. You don’t need the whole mansion today
3. Get AI accepted by your people organically
i. Emirates NBD created an internal AI Academy — not just to train devs, but to de-jargon AI for managers
–> 150+ AI pilots, cross-department adoption
ii. Udaan (India) embedded AI pricing suggestions into mainstream sales workflows
–> AOV up 12%, faster closures
iii. TymeBank (South Africa) rolled out conversational AI in local languages
–> 6M+ customers onboarded, 60% via
AI-only channels AI succeeds when it feels less like a tool, and more like a trusted teammate
4. Learn from other’s success stories – big or small
i. Duolingo (US) used GenAI to make learners feel like they had a private tutor
–> Retention up 30%, outcomes improved 20%
ii. Xiaohongshu (China) used GenAI to tag user-generated product reviews and drive intent-based shopping
–> 40% spike in click-to-buy, moderation costs cut by 60%
iii. Notion (US) used AI to auto-summarise docs and emails for SMBs.
–> Massive adoption surge without chasing the enterprise market
This isn’t about building the next gen GPT. It’s about building the smartest workflow your competitors can’t copy in time
Here’s What You Can Do Tomorrow:
a. Identify a messy, repetitive process with existing data (fraud, churn, collections, service tickets)
b. Assemble a small, cross-functional squad — business + ops + tech
c. Use lean, open-source, or local AI partners — no moonshots
d. Track outcomes, train humans, then scale
e. Bake AI into the day-to-day, not just dashboards
You already have the use cases, the data, and the ambition.
All that’s left is to stop treating AI like a keynote…
#AIFirst #RealAI #AsiaDigital #GCCInnovation #AfricaTech #AIThatWorks #NoDecksJustResults #BuildNotBuzz #OpsLedTransformation
