Consumer lending and pawn / Head of Infrastructure / Before / after

We were told to use AI. We asked one thing: can this be verified?

How F88’s infrastructure team earned trust in automation, one audited check at a time.

Published July 13, 2026/01 / 06

F88 runs one of Vietnam’s largest consumer-finance businesses: millions of customers, 800+ branches, a hybrid estate across AWS and on-premises. Its infrastructure team was told to use AI. It asked one question first: can this be verified?

So it did not hand production to an agent. It started with a read-only pilot on the dullest daily work: inventory, drift, waste, health checks. Nothing changed automatically. Every finding said what was checked, what was found and why it mattered, and the team compared each one against what it already knew. The results held. The team trusted the checks, then the workflow, then widened the scope.

A year on, F88 runs its hybrid cloud through one governed operating layer. 80% of daily manual operations are automated, with human approval where risk demands it. AWS spend is down 30% through continuous optimization. Incidents that took days to root-cause take minutes. Compliance evidence is written to the audit log on a schedule, so posture is provable any day, not only in audit week.

The number we ask you to check yourself: zero changes to production in the first 30 days. One scoped read key, one report, your team’s own judgment on whether the findings hold.

Do not trust the AI. Verify it. Then let it earn the next step.

#AISRE #FinServ #FrontierInvestigationAgents #CloudThinker

02 / 06 The problem

One team. AWS plus on-premises. Millions of customers who expect the branch to open.

Every day

Inventory and health checks

What is actually running, across many accounts and environments, versus what the documents say.

Every month

Cost drift found late

Waste shows up on the AWS invoice, after the money is spent.

Every quarter

Audit season

Compliance evidence assembled by hand for the cycle, then forgotten until the next one.

Days

to root-cause a production incident by hand: dashboards, alerts, logs and infrastructure context, one tab at a time.

Audit week

the only time the team could prove its compliance posture. The rest of the quarter, nobody knew for sure.

The constraint

Any system touching a regulated hybrid estate had to be scoped, auditable and safe by default. It had to produce evidence, not recommendations.

The team did not need more automation. It needed automation it could check.

03 / 06 Why the current answers fail

Two ways to fail: keep doing it by hand, or hand production to an agent on day one

Option A: the status quo

  • One-off scripts

    one per account, none the same

  • Tribal knowledge

    the standard lives in someone’s head

  • Manual follow-up

    tickets chased, dashboards watched

  • Quarterly evidence

    compliant in audit week, unknown after

Resultsafe, does not scale

Option B: AI with full access

  • Trust on a promise

    a demo is not a control

  • Unscoped credentials

    one key for cloud and on-prem alike

  • Recommendations only

    still a human acts, still no evidence

  • No audit trail

    the regulator asks; the chat log answers

Resultscales, nobody signs off

What F88 chose

Neither. A read-only pilot, scoped to the dullest daily work, where every finding could be checked against what the team already knew.

Trust is not granted to a system. It is earned, one audited check at a time.

04 / 06 Frontier Investigation Agents

Four steps from read-only to autonomous. The team decides when to take each one.

Step 1

Verify

Read only

Live inventory, drift, waste and health checks. Nothing changes. Every finding explains what was checked, what was found, why it matters.

Step 2

Encode

Standards as Skills

Naming, tagging, guardrails, cost controls and runbooks written once and run the same way in every account. Tribal knowledge becomes a file.

Step 3

The gate

Act with approval

Agents investigate and propose. A person approves where risk, policy or production impact demands it. Every action is written to the audit trail.

Step 4

Auto mode

Autonomous under guardrails

Per environment, the team sets Notify, Act-with-Approval or Autonomous. Low-risk remediations run on their own. Production keeps the gate.

The infra lead’s question

“Why not just give the team ChatGPT?”

A chat window has no scoped access, no approval step and no audit log. The model is the small part. The connection, the gate and the record are what a regulator can check.

Frontier Investigation Agents for your cloud. They scale the team’s judgment, they do not replace it.

05 / 06 Proof

What F88 measured after the pilot became the operating model

Daily ops by hand

80%

of daily manual operations tasks automated, human approval where risk demands it.

Waste found on the invoice

30%

reduction in AWS spend through continuous optimization, no performance trade-off.

Days to root-cause

Minutes

Resolve correlates signals across the observability stack and hands the team a structured summary.

Quarterly audit scramble

Any day

Controls checked on a schedule, evidence written to the audit log. Posture provable on demand.

“We honestly did not believe this level of automation was possible at first. We started with a small read-only pilot, verified the results, and expanded from there. Now tasks that used to take days or weeks can finish in minutes or run automatically.”
Infrastructure ManagerF88, 800+ branches, AWS and on-premises

The deepest sign of trust came after: F88’s team now helps shape the product.

06 / 06 One number, and the ask

The only number we ask you to check yourself

Read-only, first 30 days

Zero changes to production

Start the way our customers did. One scoped, read-only key. The engine inventories your estate and reports drift, waste and health. Your team checks the findings against what it already knows. If they hold, you decide the next step.

What you get

1 report

Full inventory, drift and waste findings, each with what was checked and why it matters.

What you risk

1 read key

Scoped and revocable. Every check it runs is in the audit log from day one.

Do not trust the AI. Verify it. Then let it earn the next step.

Pilot shape as run at F88.

Sources

The evidence

F88

80% of daily ops automated, AWS spend down 30%, continuous compliance.

Read the full evidence

The module

Platform

The part of CloudThinker this story runs on, in product terms.

Explore Platform
Read-only, first 30 days

Start a read-only pilot

One scoped, revocable read key. One report on inventory, drift, waste and health. Zero changes to production. Your team judges whether the findings hold.