Skip to content
Maysa Strategy Partners
ESEN
STRATEGY · OPERATIONS · APPLIED AI

Less repetitive work. More judgment where it matters.

We redesign how teams operate and automate what shouldn't consume human hours. What we build doesn't get handed over and forgotten: it stays live, monitored, and measured.

Run the free audit →
THE DIAGNOSTIC

The cost of the repetitive

No leadership team makes poor decisions for lack of intelligence. They make them because the week is spent consolidating reports, copying data between systems, and chasing approvals.

That work doesn't show up on the P&L, but it's paid in full: in attention, in decision quality, and in the month that's gone.

WHERE THE WEEK GOES
Consolidate information
Rework and validation
Follow-up and approvals
Decide and execute

A pattern we see in most operations we review. The work is to invert the ratio.

WHAT THAT COSTS, IN NUMBERS

An illustration, not a client result.

A program team with three recurring processes:

Monthly reporting to a funder or board
3 people, 6 hours each per cycle
18 h / month
Chasing approvals and manual follow-up
4 hours a week across the team
17 h / month
Preparing proposals for funding calls
2 per month, 10 hours each
20 h / month
___________
55 h / month

Not all of it can be automated, and not all of it should be: writing the argument in a proposal is judgment, not data entry. Assuming 70% of the reporting, 60% of the follow-up, and 40% of the proposal assembly can be removed, that is about 31 hours a month returned to the team.

USD 1,240 a month.
USD 14,900 a year, and it repeats every year the process keeps running.

Our assumptions, deliberately conservative. Run the numbers on your own operation with the free audit.

Run the audit →

OUR POSITION

AI is the how, not the what.

Business first: which decision, with what information, in how much time, and who owns it. Technology comes after, and only if it survives that question.

That's why we don't sell tools or pilots nobody uses. We deliver redesigned processes, with an owner, measurement, and a working version inside your operation.

THE MARKET

Everyone bought AI. Almost nobody changed how they work.

88%

of companies use AI

6%

see significant EBIT impact

21%

have redesigned a workflow

Buying the tool was the easy part. What's hard is redesigning the process and getting a team to change how it works. That is why the biggest driver of financial impact isn't the model you choose: it is whether the workflow gets redesigned, and it remains the step most organizations skip.

That is the work we do.

Source: McKinsey, The State of AI (survey of ~2,000 organizations across 105 countries). Industry benchmarks, not Maysa results.

HOW WE DE-RISK IT

We can't ask you to take our word for it.

  • We measure before we build. Frequency, minutes, people involved, and rework rate, timed as the process runs today. You sign off on the baseline before we touch anything.
  • We agree the payback window before starting. It goes into the contract with the success criteria, not into a conversation at the end.
  • Success criteria are written into the contract, not agreed at the end. We define what the ninety-day number has to look like.
  • The diagnostic is credited against the build if you continue within sixty days.
  • Managed operation is month to month. Cancel any time; your data, your accounts, and your documentation are yours from day one.

HOW WE WORK

One method, four moves

We work in short cycles with your team inside the work. Nothing ships as a document nobody opens: it ships working, with someone who owns it.

01

Map of the operation

Interviews and direct observation. We leave with the real map of processes, bottlenecks, and hours spent. We also map where automation would collide with the people who do the work, and say so. Some processes are better left alone.

02

Prioritize by value

We choose a few interventions with measurable impact and low risk. Everything else is documented for later, with explicit criteria.

03

Redesign and build

We redesign the process and build what's needed: automations, dashboards, agents, decision formats. A usable version in weeks.

04

Handover with an owner

Every process we redesign ships with a named owner on your side, the documentation of how it actually runs, and a working session with the people who use it. We keep the infrastructure running: monitoring, alerts, and adjustment as APIs and formats change. A system handed over without an owner degrades within months.

Roughly 70% of AI's transformative value depends on people, organization, and process. The technology is the other 30%. (BCG)

SERVICES

How we can help

From the first diagnostic to a process running and measured. The cost is one time; the hours come back every month.

  1. 01

    Operating hours audit

    See where the week goes, with an estimate in hours and money. No call required.

    Run the audit →
    FREE · 2 MINUTES · SELF-SERVE
  2. 02

    Operations diagnostic

    Interviews and direct observation. Deliverable: the real map, priorities, and a plan. Credited against the build if you continue within 60 days.

    2 WEEKS · FROM USD 5,500
  3. 03

    Redesign and build

    One critical process rebuilt end to end and running inside your operation. Includes the first three months of managed operation.

    SCOPED AFTER THE DIAGNOSTIC
  4. 04

    Managed operation

    What we built stays live: monitoring, alerts, fixes, and a monthly report of hours reclaimed. Add flows as you grow. Response within one business day. Cancel any time; your data is yours. The cost of running it does not grow with volume, because nothing is billed per execution.

    SCOPED AFTER THE DIAGNOSTIC
  5. 05

    Decision partnership

    A partner for the month's calls: indicators, adjustments, and judgment on the decisions that matter.

    SCOPED AFTER THE DIAGNOSTIC · LIMITED

A 15-minute conversation to see if there is a fit. No commitment.

Let's talk

RESULTS

We ran the method on ourselves first

We applied the method to our own operation before offering it. The figures are approximate, measured on real work, and each one states what it was measured on.

THE PROCESS · RECEIVABLES
60 hrsa month10 hrs

from 3 hours a day to 30 minutes

Reconciling receivables stopped meaning rebuilding them

The day used to go into keying in invoices, working out three separate withholding taxes by hand with a different rule for each counterparty, and matching payments against invoices. Now the electronic invoice is read on its own, the withholding is calculated with that counterparty's rule, and the payment is logged from the dashboard. What is left is validating.

Measured in our own operation, at 150 to 300 invoices a month across purchases and sales.

THE VISIBILITY · RECEIVABLES
7 a.m.every day

Management stopped asking for the spreadsheet

Every morning an email goes out with what is due today and what is 5, 10 or 15 days overdue. The receivable and payable balance is there at any moment, without asking anyone for it. The automation runs no AI model at execution time, so the result is the same every run and it can be audited.

Same operation, same period.

THE CONTROL · RECONCILIATION
34%of one month's spending

The classifier was working fine. The money still vanished from the report.

Eleven automatic checks run over the ledger. The strictest one requires that total spending, minus everything the report accounts for, come to exactly zero. That is the check that found 34% of one month's spending booked to categories with no budget line. Nothing was misclassified, it simply never reached the report. A faster reconciliation does not find that gap. A control that demands it does.

Measured on our own financial control system: 4,560 records, three years of history, two months closed against the statements of five accounts with a difference under one US cent.

WHERE IT SHOWS
  • Decisions made on today's number, not on the one from close
  • No collection slips through
  • The result is the same every run and it can be audited
HOW WE MEASURE

Before we touch anything, we time the process as it runs today: frequency, minutes, people involved, and rework rate. We measure again at ninety days. That difference is the number, and it is the client's, not ours.

CAPABILITIES

What we build

Document capture and reconciliation

A source document (an invoice, a payment receipt, a bank statement) is read and turned into a validated record against your chart of accounts, ready to enter the system where you keep the books, with human approval before anything posts. At close, reconciliation matches what was recorded against the bank and leaves a short list of what is missing, what does not add up, and what fell outside the budget.

Decision dashboards

A few indicators (the ones that move the needle) updated without anyone building them on Monday.

Assisted agents

Assistants that draft, classify, or review against your documents, always with human approval before they act.

Monitoring and scoring agents

Agents that watch external sources, prioritize what comes in against criteria you define, and alert the team only when something needs a human decision.

Custom web applications

Complete platforms built from scratch when a process needs its own interface: role-based access, dashboards, and reporting. Built to run in production, not as a prototype.

Operating documentation

The process written as it is actually run, with an owner and criteria, so the operation doesn’t live in someone’s head.

HOW THE CONTROL WORKS
  1. 01

    Source document

    Invoice, receipt or statement

  2. 02

    Validated record

    Against the chart of accounts

  3. 03

    Eleven checks

    If the difference is not zero, there is spending the budget cannot see

  4. 04

    Close

    A short list, not a review

WHAT SURFACES
  • Spending booked outside the budget
  • Logged payments the bank does not show
  • Transfers with no matching entry
Diagram of the reconciliation control. Not a screenshot.

ABOUT THE FIRM

A firm built on systems, not hours

Maysa exists because of a pattern we keep seeing from the inside: teams with real expertise spending their week consolidating reports instead of using it. The tools were never the problem.

We work mostly with organizations that manage portfolios and report to someone else: funds, accelerators, and development programs. We take work outside that when the fit is obvious.

Maysa Strategy Partners is a boutique strategy and AI transformation firm, led by Rafael García. We draw on over a decade of experience in operations, sales, and management at companies like 3M and Samsonite, and on current work designing and building AI and automation systems inside a global consultancy, across acceleration and financial inclusion programs in Latin America and Africa.

What we build does not depend on our availability. Automations, dashboards, and agents run on infrastructure we operate and monitor, so adding a process never means waiting in line. Judgment is the part that stays scarce: advisory retainers are limited on purpose, and we bring in the specialist each project needs. You work directly with the founder, with the closeness of a boutique firm and none of the layers.

STRATEGYOPERATIONSFINANCEAPPLIED AI

WHY US

Why work with us

01

We measure, before and after

We time the process as it runs today and again at ninety days. The difference is the number, and it is yours. Most AI work never gets measured at all.

02

We deliver working

Every phase ends in something people use: a live process, a dashboard, a running flow. Not a report.

03

Nothing we build gets abandoned

We transfer the criteria and the documentation. We keep the infrastructure running. You can leave whenever you want and your data goes with you, but nothing we build is left to rot.

CONFIDENTIALITY

Your information is yours. We sign an NDA on request. Your data is never used to train third-party models. Secure handling and controlled access on every project. We don't publish client names, identifiable cases, or screenshots of your operation.

QUESTIONS

Frequently asked questions

If yours isn’t here, write us. We reply the same day.

How does an engagement start?

Start with the free audit: two minutes, no call, no email required to see the result. If it makes sense to go further, we propose a two-week diagnostic with clear scope and price. The diagnostic is credited against the build if you continue within sixty days.

How do I know this pays for itself?

We scope the work so the hours recovered cover the investment within a window we agree before starting, and that window goes into the contract next to the success criteria. The cost is one time; the saving repeats every month the process keeps running. If the numbers don't support it, we tell you at the diagnostic, and you have spent two weeks instead of a build.

What happens when the project ends?

It doesn't end that way. What we build stays running under managed operation, priced monthly per flow and agreed in the original contract, not renegotiated at the end. You can cancel any time. Your data, your accounts, and your documentation are yours from day one.

What does managed operation actually include?

Monitoring with automated alerts, fixes when something breaks, and adjustment when APIs or formats change. Response within one business day, during business hours (GMT-5). A monthly report of runs, failures caught, and hours reclaimed. It does not include new features or new processes: those are scoped separately. If you need round-the-clock coverage, we'll tell you upfront that we're not the right fit.

What if my team resists the change?

That's the normal case, not the exception, and it's why we map it in the diagnostic. Where automation would collide with the people who do the work, we say so and often recommend leaving that process alone. What we ship comes with a named owner, documentation, and a working session with the people who use it. We don't sell change management, and we won't pretend a tool fixes a trust problem.

Do you replace people with automation?

We remove tasks, not people. What we automate is the work a machine does better: consolidating, copying, chasing. We don't automate where human judgment is what produces the value. The hours recovered are measured and documented; what your organization does with that capacity is your decision.

What happens to our information?

We sign an NDA on request. Your data doesn't train third-party models. Minimum necessary access, and data stays in your systems when possible. We never publish identifiable cases.

Does automating a process fix the errors already in it?

Not on its own. The bottleneck is rarely the time the process takes to run, it is the quality of what goes into it. Take a monthly close: automating reconciliation does not correct a payment that was never logged or a category that does not exist in the budget; what it does is make them visible in days instead of months, provided the controls that demand it get built. The same holds for a report, an approvals pipeline or a set of tracked indicators. Which controls are needed is decided at the start of a project, not at the end.

NEXT STEP

Let's take the first step

A short 15-minute conversation, no commitment, to understand your business and see if we can help.

No cost, no commitment · reply within 24 hours

info@maysasp.com