Sathish Susai

More cash in. Less cash lost.
Through Operational Excellence and AI automation.

Lean Six Sigma Black Belt 15+ years Pune, India — remote · US hours, other zones by project sathish@satfinai.com
ImpactVerified totals

The process is fixed first, then automated. Cash, capacity and time follow from that order, not the reverse.

65+
AI assistants running in live production
$57M+
Total measurable business impact
$43M+
Cash recovered
14
Performance awards, including CEO MVP

Figures are verified and independently attributable, not estimated. Where an engagement is confidential, results appear in proportional terms rather than absolute values. Targets are labelled as targets throughout.

What I am engaged forOne outcome, three methods

The result a client pays for is cash: money arriving sooner, and money stopping where it should not leak away. I get there by fixing the process first, then automating what survives.

Method 01

Operational excellence and PMO

Operating model design, KPI frameworks built from zero, transition governance, and the change management that makes any of it hold after the consultant leaves.

Method 02

AI-enabled transformation

Governance for AI adoption in regulated environments — intake design, risk-tiered review, and assistants built into live finance workflow. Mapped to NIST AI RMF, GLBA, ISO 27001 and SOC 2.

Method 03

Order-to-Cash and cashflow

Aged portfolio recovery, dispute root-cause elimination, credit policy, and the orchestration layer that puts ERP, CRM and billing into a single view.

EngagementsFourteen projects · activity, then impact

Select an entry to expand.

Problem
Every AI and technology request entered a single sequential queue. Average end-to-end approval: 17 weeks. No SLAs, no routing logic, no visibility. A low-risk licence renewal followed the same path as a high-risk external AI deployment, and the organisation was slowing its own AI adoption while believing the reviewers were the constraint.
Solution
Analysed 4,583 historical submissions across 425 data fields alongside seven stakeholder shadow sessions. The data located the bottleneck upstream of the reviewers — requests idling while incomplete information was chased. Replaced one queue with three risk-tiered parallel paths, governed by a routing matrix carrying 12 tier rules and 25 conditions, seven team SOPs and four BPMN process maps, mapped to 36 regulatory and policy sources.
Impact
Controls strengthened rather than loosened — high-risk requests still receive all seven reviewing functions, now concurrently instead of in sequence.
97%
Cycle time removed, low-risk path — 17 weeks to 3–5 days
89%
Cycle time removed, mid-tier path — to 10–15 days
10
Wasteful process steps eliminated
4 wks
To full governance artefact set, before charter sign-off

Engagement in progress. The financial value case is under validation with the client and is deliberately not quoted here.

Problem
A large global receivables operation with high AI interest and no AI capability in the workflow. Analysis, dashboards, business requirements, SIPOCs and executive material were all built by hand, and what AI knowledge existed sat with individuals rather than in the process.
Solution
Designed and deployed 65+ custom assistants with structured prompt architecture directly into live finance workflow, then trained seven fellow Black Belts across cash application, onboarding, offboarding, billing, IT risk management and PMO on prompt engineering and LLM design.
Impact
Analysis and documentation cycles compressed across the function, and the capability stayed after the engagement ended rather than leaving with the consultant.
65+
Custom assistants live in production
7
Black Belts trained on prompt engineering and LLM design
6
Functions given standing AI capability
Problem
Three core systems held three disconnected views of the same customer. Collectors reconciled by hand before they could make a single call, and any figure quoted to a customer was already stale.
Solution
Architected an orchestration layer integrating Oracle Cloud, Salesforce and Billtrust into a single hourly-synced data pool spanning billing, pricing, collections, payments and disputes — with a parallel data-integrity programme inside the ERP so the unified output could actually be trusted.
Impact
Manual pre-call reconciliation removed from the collector's day. Automation was designed only after the underlying data was made reliable, not before.
3 → 1
Systems consolidated into one live customer view
Hourly
Synchronisation, replacing manual reconciliation
5
Data domains unified — billing, pricing, collections, payments, disputes
Problem
Customer outreach coverage was capped by collector capacity, leaving large portfolio segments untouched between cycles — the accounts most likely to age quietly into a write-off.
Solution
Steered a voice and conversational bot programme, refining coverage logic by account age and balance threshold rather than applying it uniformly across the book.
Impact
Coverage extended into previously untouched segments. Separately, a vendor code-transparency gap was escalated and closed, with documented security and compliance review made mandatory before production release.
70%
Outreach improvement
1
Vendor transparency gap escalated and closed pre-production

The 70% figure is the programme target, not a measured outcome. Stated as a target because the programme had not completed within the engagement.

Problem
Collectors spent more of the day preparing reports than talking to customers. Aging and statement-of-account preparation alone consumed 60+ minutes daily per collector, and the function had begun planning twelve incremental hires to absorb the load.
Solution
Lean non-value-added analysis first — eliminating work rather than accelerating it — then generative AI, VBA and HighRadius applied only to what survived the elimination pass.
Impact
Capacity created from within the existing team. The planned recruitment was cancelled.
75%
Reporting time removed — 60+ min to ~15 daily
Collector throughput — 25–30 to 50–70 accounts per day
12
Planned incremental hires avoided
25–40%
Overall productivity gain
Problem
A global credit control function operating across 48 divisions with no measurement framework of any kind. No DSO, no CEI, no collection effectiveness view — and an SME portfolio nobody owned end to end.
Solution
Introduced API, DSO and CEI measurement across all divisions, installing measurement where none existed rather than tuning an existing system. Designed a Regional Coordinator operating model separating customer-facing collections from back-office divisional follow-up, carried by ADKAR change management across a 26-member global team.
Impact
Roughly a third of the SME portfolio recovered. Two team members on performance improvement plans were turned around through process fixes rather than discipline; both exceeded doubled targets and were retained.
48
Divisions brought under KPI measurement for the first time
~⅓
Of SME portfolio recovered
26
Global team members supervised
3 yrs
Operating model still in use after handover
Problem
A global 3PL Order-to-Cash operation moving from the US to India with no documented process to transfer — the knowledge existed only in the heads of the outgoing team.
Solution
Took the role of primary subject-matter expert and first collector simultaneously, writing the process playbook from scratch while running the live portfolio, and led ADKAR change management across both geographies.
Impact
A complete playbook where none had existed, and an offshore operation that scaled on that foundation rather than rebuilding it.
70+
Professionals through the change programme, US and India
60+ hrs
Structured training delivered
0 → 1
Process playbooks — built from nothing
Problem
A large global AR operation under pressure to absorb more volume without a corresponding budget increase, and no current view of where its capacity was actually going.
Solution
Defined the operating model, governance and performance frameworks, then built a capacity strategy on portfolio segmentation and the systematic removal of non-value-added work. Standardised a global call framework — opening, handling and closing structure — so automated agents and the human team followed one method.
Impact
A costed, segmented capacity strategy with a defined implementation path, plus a call framework embedded into the operating layer.
50%
Capacity creation
1
Global call framework standardised across agents and team

The capacity figure is the strategy's design target, not a realised outcome. Implementation continued beyond the engagement.

Problem
A pricing and billing dispute backlog that grew regardless of how much collections effort was applied. The organisation was treating it as a collections capacity problem and preparing to resource it accordingly.
Solution
Reframed it as a data defect. Fishbone and Pareto analysis traced cause from tariff type, through specific lanes, down to zip-code changes silently breaking the rating logic. A Bill of Lading correction policy and a Zip Code Error SOP were written and deployed globally.
Impact
Disputes fell because the defect generating them stopped, not because anyone chased harder. Out-of-period pricing leakage was identified and corrected in the same pass.
60%
Dispute volume reduction
2
Corrective SOPs written and deployed globally
Zip code
Resolution depth of root-cause analysis
Problem
A global third-party logistics Order-to-Cash operation newly moved from the US to India, with no process playbook, an unexplained portfolio, and stakeholders who did not believe an offshore team could recover it.
Solution
Full DMAIC with portfolio segmentation and structured root-cause work. A small sub-portfolio was piloted first to prove the method and earn permission, then scaled across the book, with a global SOP set deployed on the back of it.
Impact
Roughly half the assigned portfolio recovered within twelve months. Recognised with MVP appreciation from the CEO and an Exceeded Expectations rating in both years.
40%
DSO reduction
50%
Portfolio aging improvement
~half
Of assigned portfolio recovered in 12 months
12 mo
From pilot to full-book deployment
Problem
Three separate problems — bad debt, unallocated cash, and credit balances aged beyond 540 days — each individually deprioritised for years because none was large enough on its own to win resourcing.
Solution
Ran all three concurrently as Agile programmes rather than sequencing them: targeted write-off and dispute analysis, reconciliation redesign, and a digital payment migration.
Impact
Three long-standing ledger problems closed inside a single delivery window.
95%
Bad-debt reduction
100%
Unallocated payments cleared
98%
Credit cleanup on 540+ day aging
1,200+
Customers migrated to digital payment
Problem
Pricing and billing disputes more than two years old across the cruise key-accounts book, against elevated post-pandemic customer bankruptcy risk — meaning every month of delay carried a real chance the balance became uncollectable entirely.
Solution
Voice-of-Customer and DMAIC, with individual review of every disputed invoice and a cross-functional escalation route for the ones that had stalled between departments rather than inside them.
Impact
A backlog older than two years cleared inside a single quarter, ahead of the credit risk crystallising.
90 days
To clear a 2+ year dispute backlog
2,500+
Invoices individually reviewed and resolved
Problem
An aged North America receivables portfolio, much of it past the point where standard collections activity produced any return, and previously written down as effectively unrecoverable.
Solution
Risk segmentation and structured settlement strategy, with settlement economics evaluated against third-party collection agency commission rates rather than against face value — so every concession was defensible against the alternative.
Impact
Over half of the aged portfolio recovered, sustained rather than spiked. Advanced from Analyst to Senior Analyst on the strength of it.
8
Consecutive quarters meeting or exceeding target
>50%
Of the aged portfolio recovered
Problem
The Canada air and ocean receivables portfolio transferred between two organisations — the precise point at which portfolio knowledge is normally lost and collections performance drops for two or three quarters.
Solution
Retained continuous ownership of the portfolio straight through the transition, then applied structured dunning, credit analysis and proactive customer engagement on the other side of it.
Impact
No service disruption through the handover, and performance improved rather than recovered. Earned two performance-based salary increases within a single year.
45%
Collections increase
40%
Unapplied cash reduction
0
Days of service disruption through transition
MethodWhat travels between clients

Measure the waiting, not the working

Before improving any review process, measure how much of the total time is spent waiting for information rather than doing the work. That single number usually changes what the project is about.

Sort by risk, then run in parallel

One queue for everything is the problem. Sort the work by risk: low-risk items skip human review, high-risk items get reviewed at the same time rather than one step after another. Speed and control stop competing.

Remove waste before you automate

Automating a broken process only makes the errors happen faster and more often. Remove the steps that add no value first. Automate what is left.

Prove it small, then ask for more

Test the method on a small part of the portfolio first. A working result on a small scale is what earns permission to do the rest.

Write down what is not in scope

Recording what each team does not check prevents more rework than recording what it does check. Most duplicated effort sits in the gap nobody wrote down.

Build the handover from day one

A tracker, a list of open actions, and a named owner for each one. The person who takes over should receive clear decisions, not have to dig for them.

Numbers before opinions

No recommendation until the data supports it. An opinion costs nothing and is often wrong.

Build the team, do not become the dependency

A consultant who makes himself impossible to replace has worked in his own interest, not the client’s. Train the team and leave.

Fixed-fee engagementsDefined scope, defined end date

AI Governance & Intake Readiness

Current-state intake assessment, risk-tiering design, routing logic, and a control framework mapped to your regulatory obligations.

3–4 weeks
$12k–20k

Cash & AR Recovery Diagnostic

Quantified baseline, dispute root-cause analysis, ranked recovery list, and the process fixes behind them.

3–4 weeks
$12k–20k

OpEx Opportunity Assessment

Waste and gap analysis across a defined function, with a sized and sequenced improvement backlog.

2–3 weeks
$8k–15k
CredentialsVerified

Certification

  • Lean Six Sigma Black Belt — Six Sigma Academy Amsterdam, 2026
  • AI-Powered LSS Implementation Expert
  • Lean Management & Manufacturing Expert · Risk Management Expert
  • Generative AI for Executives — IBM
  • Google Prompting Essentials · Automation Anywhere
  • In progress — UiPath Agentic Automation · Celonis Process Mining

Education and research

  • Master of Science — Bharathidasan University, 2007
  • Bachelor of Science — Madras University, 2005
  • Peer-reviewed publication, 2008 — 143 citations
  • Ranked 62nd of 500, Cameron University international research selection

Systems

  • Microsoft Copilot Studio · Power Automate · Power BI · Power Apps
  • UiPath · Automation Anywhere · Celonis process mining
  • Oracle Cloud · SAP · Salesforce · HighRadius · Billtrust