DotheMath
Log in
DotheMath
Log in

§ 00 — Intro

Skip intro

None of them talk.

This doesn't add up.

  • Web
  • WhatsApp
  • Booking
  • Payments
  • Email
  • CRM
  • ERP
  • Accounting
  • Operations
  • Sheets
  • Data

Do the Math.

We figure out, build, connect and automate the technology behind modern businesses.

Let's figure it out See how we work

§ 01 — What we do

One studio. Five verbs.

Every engagement moves through the same five states — sometimes all of them, sometimes one.

Figure it out

Not everything needs an app. Not everything needs AI. We do the math first.

  • Strategy
  • Architecture
  • Platform evaluation
  • SaaS selection
  • Vendor evaluation
  • Technical due diligence
  • AI readiness

Build it

If buying it doesn't make sense, we build it.

  • Web applications
  • Mobile applications
  • Platforms
  • Portals
  • Dashboards
  • Internal systems
  • APIs
  • Business tools

Connect it

Your tools should talk to each other.

  • CRM
  • ERP
  • Websites
  • Payments
  • Accounting
  • WhatsApp
  • Booking
  • Operations
  • APIs
  • Webhooks
  • Data

Make it smarter

AI where it helps. Not where it looks impressive.

  • Workflow automation
  • AI agents
  • Knowledge systems
  • RAG
  • Document intelligence
  • AI search
  • Decision support
  • Communication automation

Run it

Consider us your tech team.

  • Technical management
  • Vendor management
  • Integrations
  • Incidents
  • SaaS
  • Automation
  • Upgrades
  • Continuous improvements

§ 02 — How we think

The equation

Every system that works has the same four terms. Remove one and see what you get.

System that works

§ 03 — On AI

Don't add AI. Add a reason for AI.

AI earns its place the same way a new hire does: a clear job, access to the right data, enough context to decide, and an output someone can measure.

Job description
Job
Read every inbound inquiry and classify it.
Data
The message, plus the customer's history in the CRM.
Context
Your services, routing rules and tone of voice.
Output
A category, an owner and a draft reply — measured by how often a human agrees.

Example workflow: an inquiry, end to end

  1. Inquiry

    “Do you have space for a group of 12 on Friday?” arrives on WhatsApp.

  2. AI classification

    Type: group booking. Urgency: high. Language: English.

  3. CRM

    Contact matched to an existing record. Conversation attached.

  4. Assigned employee

    Routed to the events lead, not the general inbox.

  5. Follow-up

    Reply drafted for review. Reminder set if nobody responds.

  6. Report

    Weekly view: inquiries by type, time to first reply, conversions.

A person approves anything that leaves the building.

§ 04 — Show your working

The Working

Anyone can show you the answer. We show the working — the problem, the mess, the math, the answer and the result.

Entries marked illustrative show how we write up an engagement. They are not client records.

  1. Project 001

    Hospitality Platform

    Figure + Connect + AutomateIllustrative example

    1. The problem

      Group bookings arrived through four channels and were confirmed by whoever saw them first.

    2. The mess

      WhatsApp, a web form, email and phone notes. A shared spreadsheet as the calendar. Deposits chased manually.

    3. The math

      Replacing the booking system would cost a season. Connecting the channels to it would cost a month.

    4. The answer

      One intake layer feeding the existing booking platform, with payment links and automatic confirmations.

    5. The result

      Every request lands in one queue with an owner. The spreadsheet is retired.

  2. Project 002

    Retail Operations

    Connect + BuildIllustrative example

    1. The problem

      Stock levels in the shop, the website and the warehouse never matched.

    2. The mess

      Point of sale, e-commerce and ERP each held a different truth. End-of-day reconciliation took a person half a shift.

    3. The math

      All three systems were fine on their own. The missing piece was the conversation between them.

    4. The answer

      A small sync service with a single inventory record and an exceptions dashboard for the cases that need a human.

    5. The result

      Reconciliation becomes review, not re-typing. Nobody sells what isn't there.

  3. Project 003

    Professional Services

    Figure + AI + RunIllustrative example

    1. The problem

      Senior staff spent evenings finding the right precedent document for new matters.

    2. The mess

      Fifteen years of files across shared drives, inboxes and a document system nobody trusted.

    3. The math

      A chatbot would look impressive. Retrieval with citations would actually be used.

    4. The answer

      A private search tool over approved sources, citing every answer, with access rules matching the firm's own.

    5. The result

      Answers point to the source document. Juniors can check their own work before asking.

  4. Project 004

    Marketplace Infrastructure

    Build + RunIllustrative example

    1. The problem

      A growing marketplace had outgrown the no-code tools it launched on.

    2. The mess

      Automations chained across six services. When one failed, orders silently stopped.

    3. The math

      Keep the tools that still earn their cost. Rebuild only the transaction core.

    4. The answer

      A purpose-built order and payout core, with the surrounding tools kept and properly monitored.

    5. The result

      Failures are visible and owned. The team adds features instead of patching chains.

§ 05 — Start

Got a technology problem?

Good.

Start with the problem

Tell us what's broken, slow or confusing. No brief, no deck, no jargon required.