Discover: mapping a system across a wall of diagramsDiagnose: reading dashboards to locate the problemDesign: wireframes and interface work for the solutionDeliver: shipping the finished productEvolve: continuous improvement after launchA customer record moving between connected systems
Connect · Retry · Reconcile

Integration & automation

Integrations that behave when the other side misbehaves. Every system you connect to will be slow, down, or wrong at some point, and the only question that matters is whether your business notices it as an alert or as a customer complaint.

  • Retry
    With backoff, always
  • Once
    Effect, not delivery
  • Queue
    For what cannot run
  • Match
    Both sides, daily

Sectors that trust us

What is integration work?

Integration is making two systems agree about something while neither is under your control. The mapping is the easy half; the work is what happens on a timeout, how a retry avoids charging someone twice, and how you learn that the two sides stopped agreeing.

Integration or automation?
An integration moves data between systems; an automation decides and then acts. The second needs an audit trail of what it decided and why, because a wrong decision must be explainable.

Why integrations rot

Where integrations quietly rot, and how ours are built to hold.

How they break

  • A timeout counts as failure, so it runs twice.

  • Errors are logged where nobody is looking at them.

  • The two sides drift and nothing is comparing them.

  • A schema change on their side breaks yours quietly.

  • Credentials live in one engineer's laptop, not a vault.

  • The spreadsheet between the systems is the real source.

How we build

  • Every call carries a key, so a retry is harmless.

  • Failures page whoever is on call, with context.

  • A daily reconciliation reports any difference found.

  • Their payload is validated before we act on it.

  • Secrets sit in your vault, rotated on a schedule.

  • The audit trail replaces the spreadsheet, row for row.

What we connect

Integration built to fail safely, and to say so out loud.

  • Payments and banking rails

    Idempotent charges and payouts, reconciled against the processor every single morning.

  • ERP, CRM and accounting

    Posting with a full trail, and a queue for the hours the ledger or API is closed.

  • Government and partner APIs

    Submission, polling and the retry rules their uptime demands.

  • Workflow automation

    Rules that log what they decided, and why, so a wrong one reverses.

Platforms we connect

Salesforce, Dynamics or Odoo: connected, not resold.

We hold no reseller agreement with any CRM, so the design follows your sales process and your data rules.

  • Salesforce
  • Microsoft Dynamics 365
  • Odoo

How the integration lands

From a shared spreadsheet to an audited trail.

Each step aims at a connection that fails loudly and heals.

Discover: mapping a system across a wall of diagrams
01Map

We trace what moves between the systems today, including the manual step.

Why SiaSol

Integrations your own team can run, read and reason about.

  • Safe retries

    Keyed calls, so running the same one twice changes nothing anywhere.

  • Daily reconciliation

    Both sides compared on a schedule, with a report that names every gap it finds.

  • Loud failures

    An alert with context, not a line in a log.

  • Yours to operate

    Runbooks, credentials and alert routes, handed over.

How it works · our method

Our five phases, applied to integration work.

  1. 01

    Discover

    Most technology projects fail long before the first line of code — they fail at the question.

  2. 02

    Diagnose

    Every slow process has an owner and a cost.

  3. 03

    Design

    Architecture is a set of decisions you live with.

  4. 04

    Deliver

    Software not in production is just cost.

  5. 05

    Evolve

    The handover is the point, not the end.

Where we work

Close where it counts, bilingual everywhere

  • Egypt: Cairo is home, and most of this work runs from there, because the systems we are asked to connect usually already sit in one of those buildings or in the clouds they rent next door.

  • Saudi Arabia: Riyadh, with teams on client time zones

  • Gulf and beyond: Arabic and English, one delivery standard

Cost & engagement

Integration: cost and engagement

Cost follows the number of systems and how well each one is documented, rarely their size. The first integration is quoted on its own, and what it teaches us calibrates every estimate after it.

  • Fixed · two to four weeks

    First connection

    One integration live, with its queue, its retries, its reconciliation job and its runbook.

  • Per system · quoted after the first

    More systems

    Each added on the same pattern and the same alert routes, so the tenth is cheaper than the first.

  • Monthly · dedicated

    Embedded team

    A standing team for an estate of integrations that keeps growing and changing.

FAQ

What to ask about Integration & automation.

01What if their API has no documentation?

That is common and it is priced in. We build against the behaviour we observe, record what we found as the documentation that did not exist, and add contract tests so that when their side changes we learn from a failing test, not from a customer.

02Can you use a no-code tool instead?

Yes, where it fits, and we have recommended that on paid work. It stops fitting when a retry must not repeat a financial effect, or when you need an audit trail of what was decided, because that is where the tool hides exactly what you need to see.

03How do you prevent a double charge?

Every outbound call carries a key generated from the business event, not the attempt, so the same event retried ten times produces one effect. The processor is reconciled every morning, because a key protects you only if both sides agree it was used.

04What happens when the other side is down?

Whatever your operations team decided it should, which is a question we ask in the first week. The three usual answers are queue and catch up, fall back to a manual path, or refuse the transaction, and each is a different build rather than a setting.

05How do we know the two sides still agree?

A reconciliation job compares them on a schedule and reports every difference it finds, including the ones that turn out to be fine. Most estates run it nightly; the ones moving money run it hourly, because a day of drift is a day of refunds.

06Do you build the automations too?

Yes, with a stricter rule: every automated decision writes down what it decided, which inputs it saw and which rule applied. Without that record a wrong decision cannot be explained to the customer it affected, which makes it unusable.

07Who gets the alert at three in the morning?

Whoever your rota says, and the routes are configured in your own account rather than ours. During the first weeks we take the pages alongside your team, then hand the rota over with runbooks that say what each alert means and what to do about it.

08What if a partner changes their API?

Contract tests run against their endpoints on a schedule, so a breaking change shows up as a failing test instead of as a silent gap in your data. The queue holds whatever could not be sent meanwhile, which turns an outage into a delay.

Related services

Often paired with this

Learn more

From the journal

Ready for integrations that fail safely?

Tell us which systems have to agree, and what happens today when they do not. We come back with a design and a number.

Book a call