Systems, data, and AI that hold up in the real world.
We build the software, information systems, and data platforms organizations run on — and put AI to work inside them, taking the repetitive work off your team while people keep the judgment. For research programs, government, and growing businesses, from one automated workflow to a national-scale program.
Who we work with
Four kinds of organization, one way of working. AI runs through all of it — always on top of data you can trust, and always with a person reviewing the calls that matter.
Research, survey, and development programs
Survey firms, NGOs, research institutions, and the M&E teams of donor-funded programs.
- What we do
- Data collection systems, data processing and cleaning, MEL dashboards, and statistical tables your analysts can defend.
- Where AI helps
- Automated data checks, tabulation, and progress reporting straight from field data.
- Proof
- A national Universal Health Care monitoring survey across 17 regions.
Government agencies
National agencies, LGUs, and state universities — directly, or as the named specialist inside a prime contractor’s team.
- What we do
- Information systems, data management, independent QA, and Privacy Impact Assessments for the E-Governance Act and Data Privacy Act.
- Where AI helps
- Document processing and routine report assembly, with an audit trail for every automated step.
- Proof
- Systems behind DSWD, DAR, DOST, and PCA programs, built inside implementing teams.
Growing businesses and service firms
SMEs, agencies, and operators who have outgrown spreadsheets and inboxes.
- What we do
- Workflow automation, cloud and software cost audits, custom software, dashboards, websites, and digital marketing with tracking underneath.
- Where AI helps
- Intake, follow-ups, reconciliation, and reporting taken off your team’s plate.
- Proof
- Product and QA delivery for an Australian software client across three years.
Teams building their own capability
Organizations that want their own people able to run the system — and the next one.
- What we do
- Career pathways, technical courses, team workshops, and training of trainers, built on your own data.
- Where AI helps
- Using AI safely in your own workflows — including what must never be pasted into it.
- Proof
- 15+ workshops delivered, including national training rounds for DSWD.
Services
Ten service areas, one delivery method — starting with the one most teams ask about first. Every engagement is led by our founder and scales with vetted specialists brought in as the work requires.
Ways to work with us
Four shapes of engagement. Every one is scoped from a discovery call into a written proposal, so you know exactly what you are buying before anything starts.
Assessment sprint
Two to three weeksFind out what to fix, what to automate, and what to leave alone. You get a ranked, costed roadmap — and the fee is credited against a build if you go ahead within 90 days.
Scoped project
Fixed scope, paid by milestoneA defined system or piece of work. You know the scope before we start, the milestones are the payment schedule, and the handover includes documentation and a trained team.
Embedded partner
MonthlyA senior engineer on your program for as long as it runs — maintenance, enhancements, field support, and a named owner for every live system.
Specialist inside your bid
For primes and consortiaWe join proposals as the named data, systems, or AI specialist, with CVs and references supplied for your submission, and deliver our component under your prime contract.
Case studies
We prove capability the only way that counts: systems running in production.
All work shown was led or delivered by Analytiflow’s founder, Carl Patrick Reyes. ▲ marks programs where he served inside the implementing organization’s delivery team: the technical work described is his own, and the programs were team efforts.
Automation in practice
Three workflows our founder built and runs on a live national survey — the same kind of work we take off your team.
AI-assisted ticket triage, with a person approving every fix
- The problem
- Issue reports from field teams arrived through a public register, chat channels, and a no-account form — at all hours, in every format, during live fieldwork where a bad fix reaches every tablet.
- What we built
- A watch that checks every channel three times a day, screens each ticket for personal data, tags it by severity on a written scale, and drafts the triage: what it is, the field impact, and the fix it would propose. Nothing is changed until a person approves that ticket.
- Result
- 309 instrument issues resolved through testing, every ticket raised before fieldwork closed, and no fix shipped during the survey without a human decision.
Field reports reconciled against the data that actually arrived
- The problem
- Supervisors tracked progress in their own spreadsheets, from field reports rather than from the server — so the numbers in meetings and the numbers in the data rarely matched, and nobody knew why.
- What we built
- A monitor that watches the trackers three times a day, snapshots every change, normalizes the sheets, and reconciles reported completions against synced cases region by region.
- Result
- One meeting-ready brief, reported against synced, with a reason for every gap — built and running within a day of the request.
Statistical tables that rebuild themselves every hour
- The problem
- A survey commits to a long list of statistical tables up front. Compiling previews by hand means analysts only see the shape of the results weeks after the data arrives.
- What we built
- A tested generator that maps the committed table plan to the live dataset and rebuilds preview tables every hour, each stamped with its status so nobody mistakes a preview for an official figure.
- Result
- 135 preview tables regenerated hourly against a 197-table plan — results visible while fieldwork is still running.
“…successfully launched and is now being populated with project data.”
Dr. Maria Elizabeth B. Naredo — Project Leader, PCA Albay Research Center, reporting the germplasm information system to central office, 2026
“Thank you very much… for your excellent services! Super napadali po ang ating data gathering because of the software.”
Kristia Lei Reyes — Batangas State University, on the DOST-CEST field data collection system, 2021
Have a system that needs to exist — or one that needs rescuing?
Book a discovery callWho you’ll work with
Carl has built data and software systems since 2017 — more than 25 projects for government agencies, research institutions, and international organizations. That includes twelve-plus survey and data-collection systems, the information systems behind national programs, and the quality assurance that took them to production.
He co-founded a software company that grew out of a delivered government project, and has trained field teams, analysts, and implementers across more than fifteen workshops. Analytiflow is the practice he leads now — founder-led, with vetted specialists brought in as each engagement needs them, and AI built into how the work gets done.
- building data and software systems
- Since 2017building data and software systems
- projects for government, research, and international organizations
- 25+projects for government, research, and international organizations
- survey and data-collection systems
- 12+survey and data-collection systems
- training workshops delivered
- 15+training workshops delivered
How we work
The Analytiflow Framework
AI is the top rung, and it only holds if the rungs below it do. Most organizations try to jump from spreadsheets straight to AI, and it falls over. Four phases move you up a four-rung ladder: the phases are how we work, the ladder is what changes.
Download the framework diagram (SVG)
How we work · four phases
- 1Find
What is true today?
A current-state review of your systems, your data, and the workarounds people have built between them. We establish which rung of the ladder you are actually standing on, and what it would cost to move.
→ Costed roadmap and a maturity assessment
- 2Frame
What exactly are we building?
Specification, data model, and acceptance criteria agreed before anything is coded. Validation rules, edge cases, and the awkward questions get resolved on paper, where they are cheap to fix.
→ Signed specification and acceptance criteria
- 3Flow
Build it in working increments.
Two-week cycles, each ending in something that actually runs and a demo you attend. Priorities are reset every cycle, so a change of mind costs a conversation rather than a change order.
→ Working software at the end of every cycle
- 4Fortify
Make it survive without us.
Documentation, runbooks, training, and a real handover. A system nobody in the building can operate has not been delivered — it has been abandoned in place.
→ Handover pack, trained team, and a support plan
What changes · four rungs
- 4Automated
The system acts on the data. Automation and AI do the repetitive work, under human review where being wrong is expensive.
AI Workflows & Automation - 3Visible
People can see the numbers and decide from them. One metric, one definition, one source — and every figure traceable back to where it came from.
Data Analytics & BI - 2Trusted
The data is clean, validated, defined, and governed. You can defend the numbers when somebody challenges them.
Data Management & Processing - 1Captured
The data exists digitally, recorded at the moment it happens. Not on paper. Not re-typed from a form three weeks later.
Information Systems · Survey Systems - 0Scatteredwhere most start
Spreadsheets, paper, and people’s memory. Every question needs a person to answer it, and two people may answer it differently.
Start with an assessment
Every engagement moves you at least one rung — and leaves you able to hold it without us.
How we hold ourselves to it
One accountable owner
Our founder leads from first call to production, with vetted specialists brought in as scope requires. No hand-off gaps, and never a stranger arriving halfway through.
Written before built
Specifications before code, generators over hand-edits. Anything that will need revising is produced by something we can re-run, not retyped by hand in a GUI.
AI-assisted, engineer-accountable
We run AI agents through our own delivery pipeline — planning, code generation, testing, documentation — under engineering review. It compresses the timeline. It does not move the accountability.
FAQ
About us
Is Analytiflow new or established?
Analytiflow is new. The practice isn’t: it’s built on work since 2017, and the platforms in our case studies are in production today. We’d rather show you running systems than a long client list.
Who will I actually work with?
Our founder, Carl Patrick Reyes, leads every engagement from the first call to production. Vetted specialists — developers, QA engineers, data programmers, statisticians — join as the scope requires. You get senior attention without paying for a bench.
Who do you work with?
Research and development programs, government agencies, growing businesses, and teams that want to build their own capability. Our track record is deepest in large-scale data programs, and the same engineering carries to an automated workflow or an operations dashboard.
Do you work outside the Philippines?
Yes. We’re based in Manila and work with local and international organizations. Delivery is remote-first; field deployment support is available for programs in the Philippines.
Working together
How does an engagement start?
A 30-minute discovery call, or a project brief by email. Then a written scope with fixed milestones. No retainer is required to find out what you need.
How long does a project take?
An assessment sprint takes two to three weeks. A scoped build is planned in two-week cycles, with the end date set in the written scope. Embedded and support arrangements run monthly for as long as the system is live.
Why don’t you publish prices?
Because the number depends on your scope — how many systems, how much data, which integrations, whether there is field work, and how long it has to be supported. Describe those on the first call and we’ll give you a range. For agency and donor work, the budget is set by your procurement or funding envelope.
Can you work as a subcontractor or join our bid?
Yes, and often that’s the cleanest arrangement. We deliver components inside a larger program under a prime contract, and we join proposals as the named data, systems, or AI specialist with CVs and references supplied for your submission.
AI and data
Where does AI actually help — and where doesn’t it?
It helps with high-volume, repeatable work: reading and sorting documents, chasing follow-ups, checking data, assembling routine reports. It doesn’t replace judgment calls, and it fails on data nobody trusts — so we fix the data first and keep a person reviewing wherever a mistake is expensive.
What does “AI-assisted delivery” mean?
We use AI agents inside our own delivery — drafting code, running tests, generating documentation — under engineering review. It compresses timelines; it doesn’t move the accountability. What we install for clients is the same discipline we run on ourselves.
What happens to our data?
We work to the Data Privacy Act. Your data stays in systems you control, AI tools see only what the agreed scope allows, and every automation is documented with what it can access.
Who owns what you build?
You do. The handover includes source, documentation, credentials, and a trained team — the Fortify phase of our framework exists so you can run it without us.
Contact
Thirty minutes, no pitch. Tell us what you’re trying to ship — we’ll tell you honestly whether we’re the right fit, and what it would take either way.
- Email ·
- carlpatricklreyes@gmail.com
- Timezone ·
- Manila (UTC+8)
- Founder ·
- Carl Patrick Reyes
Or send a project brief
Five questions, a few sentences, and it lands in our inbox.