Service 02 of 10 · Technology & Intelligence

Custom software for teams that have outgrown spreadsheets.

Your own customer records (CRM), dashboards and staff tools, built around how you work. Everything we build runs in your own cloud and is yours once it is paid for.

model / your-company / core-domain
Consentconsents
  • PKiduuid
  • FKcontact_id→ contact
  • purposeenum
  • statusenum
  • captured_attimestamptz
  • sourcetext
Eventevents
  • PKiduuid
  • FKcontact_id→ contact
  • typetext
  • channelenum
  • occurred_attimestamptz
Opportunityopportunities
  • PKiduuid
  • FKaccount_id→ account
  • stageenum
  • amountnumeric
Casecases
  • PKiduuid
  • FKcontact_id→ contact
  • FKorder_id→ order
  • priorityenum
  • sla_due_attimestamptz
Contactcontacts
  • PKiduuid
  • FKaccount_id→ account
  • emailcitext · PII
  • IDXemail_sha256text
  • phone_e164text
Accountaccounts
  • PKiduuid
  • nametext
  • tierenum
  • regiontext
Productproducts
  • PKskutext
  • nametext
  • pricenumeric
  • activeboolean
Orderorders
  • PKiduuid
  • FKaccount_id→ account
  • totalnumeric
  • channelenum
  • placed_attimestamptz
Typical first release
12–24 weeks
Scope
CRM · dashboards · staff tools
Ownership
Your cloud, your code
Illustrative model

An illustrative map of the business data for “Your company”, with 8 kinds of record. An account has contacts, sales deals and orders. A contact has events, consents and support cases. Each order has order lines, and each line names one product. A case may refer to one order. Beside the map, the same account appears as the form a person would use. It shows the tier, region, owner team, health score and latest order.

Over a colleague's shoulder: a laptop open on a dense spreadsheet of rows and columns at a shared table

An illustrative shared spreadsheet with CRM export, Billing and Support tabs. A range copied from the CRM export into Billing fails one VLOOKUP (#N/A) because the account name is spelled differently, and Support can only be matched by email.

Why now

Copy-paste and email approvals cost your team hours each week.

Custom software usually replaces a shared sheet, an export or an inbox rule your team relies on. These are the 4 signs.

  1. 01

    One customer, 4 IDs.

    Sales uses ACC-10442, billing C-88121 and support an email. Every report starts by matching them.

    9.5h / weekMatching records

  2. 02

    Approvals by email.

    Discounts, refunds and credit limits are agreed by reply-all. Nobody can tell who approved what.

    7h / weekChasing approvals

  3. 03

    Reports rebuilt every Monday.

    Someone pastes 5 exports together by hand. The numbers are a week old when read.

    6h / weekRebuilding reports

  4. 04

    No record of changes.

    Nothing records the old price or status, who changed it or why. Every audit starts from scratch.

    4.5h / weekTracing changes

Across all 4

27 h / week about 0.7 of a full-time job, before counting any mistakes

IllustrativeScreens marked Illustrative are examples, not client results. Figures assume a 40-person operations team.

Inside a tool

Staff tools with approvals, access rules and a record of every action.

Each tool is built for people who use it all day. It has keyboard shortcuts, instant updates and one log-in. The 6 notes explain why.

An illustrative staff tool for support cases at “Your company”. Saved views sit on the left. The list shows open P1 and P2 cases in the West region, all this role can see. It has 3 selected for a bulk action. A refund above 50,000 shows a second person’s approval and its history. Below is a linked record of every action. The 6 numbered notes explain the design.

What we make

6 kinds of software, built to work together.

We map your business data first, so a sale entered once shows in every tool and report. For each part we choose to build, buy or extend.

Two engineers talk through code on a monitor at a shared desk
01

Custom CRM

Every customer, enquiry and sale in one place, set up the way your team sells and serves.

Build or buy

Extend Salesforce or HubSpot when their process fits. Build when yours sets you apart or per-user fees keep rising.

store
PostgreSQL · RLS
flow
Temporal: renewals, onboarding
sync
Email, calendar, orders · two-way
02

One record per customer

One profile per customer, merged from web, app, store and more, shared by every team, with consent.

Build or buy

Buy a ready-made one for standard marketing. Build your own when you need control of matching, consent and cost.

match
Exact matches, then scored
consent
Checked per purpose
activate
Synced to CRM, email, ads
03

Staff tools and portals

Admin screens, partner portals and office tools that replace spreadsheets and email chains.

Build or buy

Use a drag-and-drop tool for simple screens. Build when you need access rules, bulk actions, approvals and change history.

auth
One log-in · SAML or OIDC
work
Bulk actions by keyboard
audit
Every action recorded
04

Approval and request flows

Onboarding, refunds, credit and change approvals run step by step, with deadlines, reminders and two-person sign-off.

Build or buy

Use your main system’s approvals when every step stays in it. Build when a task crosses tools or waits days.

engine
Temporal · Camunda (BPMN)
timers
Retries, time limits, undo steps
rules
Deadlines, reminders as code
05

Data warehouse and dashboards

All your data in one warehouse, with each number defined once, so every dashboard and AI tool agrees.

Build or buy

Buy ready-made connectors and a cloud warehouse. Build the data models and metrics, because they describe your business.

ingest
CDC · Debezium and Kafka
model
dbt, auto-tested
metrics
Defined once, used everywhere
06

Replacing old systems

Old systems replaced part by part, running side by side with a way back at each step.

Build or buy

Move to the cloud when the system works but its servers are ageing. Rebuild when it no longer fits.

route
Parts swapped one by one
sync
Synced, checked every day
undo
Undo any part with a switch

Customer data demo

Customer matching that merges 8 events into one profile.

One record per customer starts with knowing who is who. Here one customer does 8 things on web, app, store, support and email. Watch them merge, then change the rules or withdraw consent.

Interactive demonstration with fictional data. Use Next event, Play all or Reset to step through eight events. Switch the match rules between exact matches only and exact plus likely matches, and switch marketing consent to withdrawn. The status line below the controls announces each change; the profile, segments, destinations and steward queue update to match.

cdp / identity-resolution / Your company Fictional data
Match rules

Event 8 of 8 · Web add_to_cart · ck_e913 scored 0.72, below 0.85 · queued for data steward review

Event stream5 channels

Identity graph7 identifiers · 1 profile · 1 in review

Golden profileKnown

P-00016 identifiers linked
Email
9b2e…41 · verified at login
Phone
+91 ••••• 4471
Loyalty
L-20931 · Silver
Lifetime value
4,850
Last seen
21:03 · Email
Marketing consent
Granted · 09:14 · signup form
  • Web
  • App
  • Store
  • Support
  • Email

Segments5 of 5

  • Known contactsverified emailIn
  • Store buyers · 30 daysPOS purchaseIn
  • Open support caseticket status openIn
  • High intent · 24 h≥ 2 product viewsIn
  • Email engaged · 7 daysclick or replyIn

Activationconsent checked per purpose, every sync

  • SalesforceCRM · sales and service Purpose · Service Sendsemail, phone, open case Lawful basis: contract Synced Contact + open case · 16:22
  • HubSpotMarketing email Purpose · Marketing Sendsemail, list membership Consent granted · 09:14 Synced List: Email engaged · 21:03
  • Google AdsCustomer Match Purpose · Marketing Sendsemail_sha256 only Consent granted · 09:14 Synced Audience: High intent · 19:48
  • MetaCustom Audiences Purpose · Marketing Sendsemail_sha256 only Consent granted · 09:14 Synced Exclude: recent store buyers

Data steward queue1 waiting

ck_e913 → P-00010.72 / 0.85

Same /24 network · similar session path · different browser

Nothing waiting. Matches that score below 0.85 land here for a person to decide.

Resolution logappend-only

  1. merge.queued ck_e913 → P-0001 · 0.72 < 0.85
  2. event.attached campaign_click → P-0001
  3. identity.linked ck_d201 → P-0001 · 0.91 · likely match
  4. profile.merged P-0002 → P-0001 · email + phone
  5. profile.created P-0002 · loyalty + phone

Exact matches first

Records join on exact keys: a verified login, email, loyalty ID or phone number. These merges can run automatically.

Likely matches, with a cut-off

Device, network and behaviour data are scored. Above an agreed cut-off, records link as likely matches. A named person decides the rest.

Consent checked on every send

Consent is stored per purpose and checked on every send. A withdrawal stops marketing within minutes and is logged.

  • GDPR — General Data Protection Regulation (EU) 2016/679
  • DPDP Act 2023 — Digital Personal Data Protection Act, 2023

Behind your reports

Data warehouses with numbers finance can trace.

Every table has an owner, agreed rules and an update deadline. Data that breaks the rules waits until it is fixed.

The inside of an opened hard drive: a mirrored platter, the read arm and its spindle
Storage is the cheap partDisks cost little. Data nobody trusts costs a great deal, so we build for trust.Photo: William Warby / Unsplash
data-platform / your-company / prod
Rows today
18,406,212
Contract checks passed
1,284
Batches quarantined
1
Freshness p95
6 min
Lineage coverage
100% of gold
Illustrative

An illustrative data warehouse for “Your company”. Data arrives from product apps, online tools, files and the tills in 40 stores. Some comes live, the rest in batches. Bronze keeps raw data as received. Data rules are checked on the way into Silver. Data that fails is held for its owner. Silver holds checked tables without duplicates. Gold holds tested tables ready for reports. Net revenue, active customers and how long orders take are each defined once there. Dashboards, the CRM, AI agents and month-end finance reports all use them.

Freshness SLOs per table, monitored

TableSLONowState
silver.orders≤ 15 min6 minMet
silver.customers≤ 60 min22 minMet
silver.inventory≤ 5 min3 minMet
gold.fct_revenue≤ 24 h4 h 10 minMet
gold.dim_product≤ 24 h21 h 40 minAt risk

Breaches page the table owner, not a shared inbox.

Data contract checked on every batch

# contracts/silver/orders.yaml
dataset: silver.orders
owner:   order-platform@your-company
schema:  # all fields required
  order_id:   { type: uuid, unique: true }
  account_id: { type: uuid, ref: accounts }
  total:      { type: decimal(12,2), min: 0 }
  placed_at:  { type: timestamptz }
sla:
  freshness:    15m
  completeness: 99.9%
on_violation: quarantine + page owner

Data tools we work with

  • Apache Kafka
  • Airbyte
  • Snowflake
  • Databricks
  • Google BigQuery
  • Apache Airflow
  • Apache Spark
  • DuckDB
  • Looker
  • PostgreSQL
  • Debezium
  • Apache Iceberg
  • dbt
  • Power BI

Debezium streams database changes to Kafka. Apache Iceberg stores tables in an open format any tool can read.

Tools we use

7 layers of technology, each choice written down with its reason.

Everything we build has the same 7 layers. We pick each layer’s tools to suit what your team can run.

Layers top to bottom · select one

7 layers · hover or tab in to spread them out

Layer 01 · Web, mobile and admin screens

Screens

Admin screens start from the data map, then are refined by hand. Tools used all day must feel fast on a 5-year-old laptop.

Our default
React with Next.js and TypeScript
Something else when
Flutter or native when the field app must work offline

Tools we work with here

  • React
  • Next.js
  • TypeScript
  • Storybook
  • Tailwind CSS
  • Flutter

ADR-004Fast customer pages from the server; admin screens from the data mapAccepted

Layer 02 · Rules and connections

Server code

Code is split the way your business works, in one language your team can hire for.

Our default
Node.js with NestJS, or Python with FastAPI
Something else when
Go for heavy traffic; Java with Spring or .NET where your team already runs them

Tools we work with here

  • Node.js
  • NestJS
  • FastAPI
  • Python
  • Go
  • Spring Boot
  • .NET
  • GraphQL
  • OpenAPI Initiative
  • Apache Kafka
  • RabbitMQ

Kafka and RabbitMQ pass messages between services

ADR-002One code language per team; a second needs a measured reasonAccepted

Layer 03 · Tasks that take days

Processes

Onboarding, approvals and renewals can take days across several tools. Temporal keeps their place through updates and outages.

Our default
Temporal
Something else when
Camunda when your team wants to own the process diagrams; a simple queue for small jobs

Tools we work with here

  • Temporal
  • Camunda

ADR-007Long tasks run in Temporal, so they survive restartsAccepted

Layer 04 · Where records live

Databases

PostgreSQL is our default. It can limit each person to their own rows. A second database needs a written reason.

Our default
PostgreSQL
Something else when
Redis for caching; Elasticsearch or OpenSearch for search; MongoDB only with a clear reason

Tools we work with here

  • PostgreSQL
  • Redis
  • Elasticsearch
  • OpenSearch
  • MongoDB
  • Supabase

ADR-001PostgreSQL holds the records; any second database needs a reasonAccepted

Layer 05 · Data warehouse and metrics

Reporting

Data models are tested automatically and documented. Each metric is defined once, and warehouses pause when idle.

Our default
dbt on your cloud warehouse
Something else when
Databricks or Spark when data scientists use it too; DuckDB for local runs and tests

Tools we work with here

  • dbt
  • Snowflake
  • Databricks
  • Google BigQuery
  • Apache Airflow
  • Airbyte
  • Apache Spark
  • DuckDB
  • Looker
  • Power BI

Also Apache Iceberg tables

ADR-009Each metric is defined once; dashboards may not change its meaningAccepted

Layer 06 · Who can do what

Access

Staff use your company log-in (SAML or OIDC). Accounts open and close automatically. System passwords sit in a secure store.

Our default
Your company log-in through OIDC
Something else when
Keycloak or Auth0 when you have no company log-in yet, or for customer portals

Tools we work with here

  • OpenID
  • Okta
  • Auth0
  • JWT
  • Vault

Also Keycloak · Microsoft Entra ID

ADR-003Access is checked in the code, then again in the databaseAccepted

Layer 07 · Where it runs

Hosting

Everything runs in your cloud, set up in code. Each release is reviewed and monitored, and can be undone.

Our default
Kubernetes or a managed container service, Terraform, GitHub Actions
Something else when
Serverless for bursts of traffic; one small server for a tool with 20 users

Tools we work with here

  • Kubernetes
  • Terraform
  • Docker
  • Helm
  • Argo
  • GitHub Actions
  • AWS
  • Microsoft Azure
  • Google Cloud
  • OpenTelemetry
  • Grafana

ADR-005No manual changes to the live system; every change is checked firstAccepted

  • Well-tested tools first

    We use well-tested tools underneath, so our effort goes where it pays: your data, rules and processes.

  • Fewer databases

    Each database needs a measured reason. Fewer parts mean less to patch, back up and explain to an auditor.

  • Everything as code

    Server set-up, data steps, access rules and documents live in code you own, reviewed like any change.

How we use AI

AI does routine data work. A named person signs off.

We use AI agents, which do tasks such as labelling data and drafting documents. A named person approves each change.

catalogue / silver / customers classifier v4 · 2 proposals · 2 approved by policy

Interactive demonstration with fictional data. The columns list shows agent-proposed classifications with confidence. The steward queue has approve and reject buttons that apply or discard a proposal. Run classifier repeats the scan. The status line announces changes. Column-level lineage traces email from crm.contacts.email through bronze and silver to a hashed column in gold used by the Active customers dashboard, and the catalogue answers lineage questions in plain language. A deletion request, DSR-1042, runs across the CRM, the warehouse, the email platform and the ad audiences in order, with a receipt from each sealed into an evidence bundle.

Columns6 of 8 classified

  • customer_id uuid — primary key
  • email text PII.contact 0.99 pattern + column name
  • phone_e164 text PII.contact 0.98 E.164 pattern
  • pan_last4 char(4) Payment.truncated 0.97 name + source: payments
  • date_of_birth date PII.sensitive 0.96 name + value distribution
  • region text — reference data
  • lifetime_value numeric Business.metric 0.91 derived in gold · auto-approved by policy
  • consent_marketing boolean Consent.purpose 0.99 source: consents · auto-approved by policy

Approval policy policies/classification.yaml

auto_approve:
  when: tag in [Business.*, Consent.*]
    and confidence ≥ 0.90
otherwise: queue for data_steward
never_auto: [PII.*, Payment.*, Health.*]

Steward queue4 waiting

  • Classification Tag email as PII.contact confidence 0.99 · masks in non-prod · 30-day access log Approved · data_steward · 02:14
  • Classification Tag pan_last4 as Payment.truncated confidence 0.97 · last four digits only · finance role Approved · data_steward · 02:14
  • Classification Tag phone_e164 as PII.contact confidence 0.98 · masks in non-prod
  • Classification Tag date_of_birth as PII.sensitive confidence 0.96 · restricts to support lead role
  • Draft Data contract silver.customers v4 drafted from the schema and last 30 days of batches
  • Draft Column descriptions · 8 of 8 drafted from names, sources and sample values
Decided this week
38
Median time to decision
4 h 10 min
Applied without a person
0 regulated tags

Column-level lineageemail · source to use

›Where does the email in Active customers come from, and is it masked?

dim_customer.email_hash ← silver.customers.email ← bronze.contacts_raw.email ← crm.contacts.email. SHA-256 hashed in gold, masked in non-prod since 02:14. Owner: growth. 1 consumer, 4 sources cited.

Deletion requestDSR-1042 · subject 9b2e…41 · GDPR Art. 17 / DPDP s.12

  1. CRMreceipt · a91c…
  2. Warehousereceipt · 4e07…
  3. Email platformreceipt · c2f8…
  4. Ads audiencesreceipt · 77d1…

Evidence bundle e7c1…3a sealed · 4 of 4 systems · 38 s

Steward approved pan_last4 → Payment.truncated · restricted to finance role · logged 02:14:07

Labelling sensitive data

An agent suggests labels from column names and sample values. Low-risk labels apply by rule. Personal data waits for a named person.

Data documents kept up to date

Agents draft data rules, column notes and each number’s source. They redraft them as data changes. People sign off.

Deletion with proof

A deletion request clears a person’s data from every system. Each returns a receipt for regulators under Indian or EU law.

Cost and carbon

Lower warehouse bills and less energy.

What cuts your warehouse bill also cuts its energy use. We report cost and carbon for each job.

finops / your-company / warehouse
Scanned per day
41.8 TB6.2 TB−85%
Warehouse credits / day
15038−75%
Storage bill / month
$841$102−88%
SCI per 1,000 queries
182 g47 gCO₂e
Illustrative

Illustrative figures for “Your company”, before and after the policies; the buttons switch the window between the two states. Storage: 38.2 TB before, of which 34.6 TB sat in the hot tier; after tiering and retention, 0.9 TB hot, 2.4 TB warm, 12 TB cold and 3.6 TB archive, with 19.3 TB deleted. Compute: 150 warehouse credits a day before and 38 after, from incremental models, partition pruning and clustering, change data capture merges, auto-suspend and changed-rows-only syncs. Data scanned falls from 41.8 TB to 6.2 TB a day, the storage bill from about $841 to $102 a month, and software carbon intensity from 182 to 47 grams CO₂e per 1,000 queries. The warehouse now runs about 10 of 24 hours a day instead of all 24.

  1. Only new rows processed

    Each run handles only new and changed rows, instead of rebuilding the whole table.

    materialized='incremental'
  2. Tables split by date

    Tables are split by date, so a dashboard reads only the days it needs.

    PARTITION BY DATE(placed_at)
  3. Warehouses pause when idle

    Processing pauses after 1 idle minute and restarts on the next query.

    AUTO_SUSPEND = 60
  4. Keep data for a set time

    Each dataset has a time limit. Old data moves to cheaper storage, then is deleted unless held for legal reasons.

    retention: 13 months
  5. Delete what nobody reads

    Usage logs find tables nobody has opened in 6 months. We delete them once their owners agree.

    0 reads · 180 days

Switching over

Replace old systems one part at a time.

Old and new run side by side behind a switch, as work moves across in stages. Each old part goes off only when the numbers match.

migration / your-company / legacy-erp → platform
  1. Customer accounts retired · data archived
    accounts-service live · legacy switched off archive checksum verified
    Customer accounts: 100% of live traffic on accounts-service. Legacy retired · data archived.
  2. Product catalogue read-only · rollback ready
    catalogue-service live · 100% reconciled daily · 0 diffs
    Product catalogue: 100% of live traffic on catalogue-service. Legacy read-only · rollback ready.
  3. Orders serving 50%
    orders-service 50% · by region 4.8 M compared · 0 diffs
    Orders: 50% of live traffic on orders-service. Legacy serving 50%.
  4. Pricing & quotes serving 90%
    pricing-service canary · 10% of tenants 96,540 compared · 0 diffs
    Pricing & quotes: 10% of live traffic on pricing-service. Legacy serving 90%.
  5. Billing & invoicing serving 100%
    billing-service shadow · reads mirrored 1.9 M compared · 3 diffs open
    Billing & invoicing: shadow run, no live traffic moved. Legacy serving 100%.
  6. Reporting serving 100%
    data platform · gold marts not started no parallel run yet
    Reporting: 0% of live traffic on data platform · gold marts. Legacy serving 100%.
Modules on the platform
2 of 6
Live traffic moved
43%
Open diffs
3
Rollbacks
1
Unplanned downtime
0 min

Cutover log append-only

  1. Accounts legacy module retired · 412,908 records archived · billing shadowed through month-end
  2. Orders at 10%: diff rate 0.3% over the 0.1% budget · flag back to 0% in 40 s · fixed and resumed
  3. Accounts to 50% · catalogue 10% · orders in shadow on 4.8 M records

Illustrative plan6 parts over 40 weeks for “Your company”. Your own stages follow your risks, release dates and month-end close.

How it works

Map, build, switch over, run.

First we map how your business runs. Then one working flow goes live early. New parts follow every 2 weeks.

A whiteboard covered in sticky notes in rows, with boxes and arrows drawn between them

Key to the sticky notes

  • Eventorder.placed
  • Actionplace order
  • Rulewhen paid, ship
  • Questionwho approves?
  1. Map how you work

    Wk 01–03

    We run workshops with the people who do the work. Together we map each step, rule and record.

    AI agents help
    Turn the workshop wall into a draft data map overnight. Flag words your teams use in two ways.
    People decide
    The parts, their names and which goes live first.

    What you keepWorkshop map · Data map · Decision records · Quality targets

  2. Launch one working flow

    Wk 04–08

    One working flow goes live behind an on-off switch. Log-in, live data, testing and alerts work from day one.

    AI agents help
    Draft admin screens, connection code and tests from the data map.
    People decide
    A named engineer approves each change. Users accept the flow.

    What you keepFirst live flow · Automatic releases · Server set-up in code · Alerts

    release 1 live · new onboarding on for 12 users · 95% under 180 ms

  3. Build part by part

    Wk 09+ · every 2 weeks

    We work in 2-week cycles, in priority order. Each ends with a demo and a release you can switch on.

    AI agents help
    Draft tests, documents and release notes. Flag code that no longer matches the data map.
    People decide
    Design, priorities and code review. No agent approves code.

    What you keepNew releases · OpenAPI 3.1 reference · Automatic tests · Release notes

    cycle 6 · 3 parts live · 1,184 tests · 0 failed

  4. Move over in stages

    Per part

    Rehearse, then run old and new side by side with nightly checks. Each team is trained before its move.

    AI agents help
    Compare sample records in old and new, group the differences and summarise them for the owner.
    People decide
    Sign-off per stage. Finance approves parts that move money.

    What you keepData move scripts · Check reports · Undo plans · Training

    stage 3 · 4.8 M records match · signed off 14:20

  5. Run and improve

    Ongoing

    We agree speed and reliability targets and review use and data quality weekly. Your team owns the to-do list.

    AI agents help
    Sort alerts, draft problem reports and answer “where does this number come from?”
    People decide
    Urgent fixes, the plan and what to switch off.

    What you keepTargets · How-to guides · Usage dashboard · Design reviews

What you get

Code and documents in your own accounts as we go.

The code, data map and how-to guides sit in your accounts. Your team can search them on one site.

docs.your-company.internal / docs/domain/model.md

Domain model

v12 · migrated

The entities, relations and business terms your platform is built on, versioned with the schema so the diagram can never drift from the database.

Account
A legal entity that buys from you. Not a login: people are Contacts.
Order
A confirmed purchase. Quotes are not orders until accepted.

Model history

  1. v12orders.channel added as an enum: web, app, storemigration 0042 · 2 reviews
  2. v11consents split by purpose, one row per purposemigration 0039 · 2 reviews
  3. v10cases may refer to one order (0..1)migration 0035 · 1 review

API reference

OpenAPI 3.1 · linted

Generated from the spec the services are tested against, with typed clients for the web app and your integrations.

  • GET/v1/accounts/{id}Read an account
  • POST/v1/ordersPlace an order · idempotency key
  • PATCH/v1/cases/{id}Update a case · If-Match
  • GET/v1/customers/{id}/profileGolden profile · consent-aware
/v1/orders:
  post:
    operationId: placeOrder
    security: [{ oidc: [orders.write] }]
    parameters: [{ $ref: '#/components/parameters/IdempotencyKey' }]
    responses:
      '201': { $ref: '#/components/responses/Order' }
      '409': { description: Request with this key in progress }
      '422': { description: Same key, different body }

Data dictionary silver.orders

5 of 5 described

Every column with its type, meaning, classification and owner. Descriptions are drafted by agents from the schema and approved by the owner.

ColumnTypeDescriptionClassOwner
order_iduuidPrimary key, generated at checkout—orders
account_iduuidBuying account; FK accounts.id—orders
totaldecimal(12,2)Order value incl. tax, in account currencyBusinessfinance
placed_attimestamptzWhen the customer confirmed, UTC—orders
ship_emailtextDelivery notification addressPII.contactorders

Data models & tests

CI · 412 tests passing

Every transformation is SQL in version control, tested on each pull request and documented from its own schema. Gold models carry an enforced contract, so a breaking change fails the build instead of a dashboard.

ModelMaterialisedTestsContractOwner
gold.fct_revenueincremental18enforced · v3finance
gold.dim_customerincremental22enforced · v4growth
silver.ordersincremental14enforced · v3orders
silver.refundstable9—finance

Runbook orders CDC lag

Reviewed 12 days ago

One page per alert: what it means, how bad it is, and the exact steps, written for the person on call at night.

Alertcdc_lag_seconds{table="orders"} > 300 for 10m

ImpactDashboards and the CRM show orders late. Nothing is lost; Kafka retains 7 days.

  1. Check connector status in the Kafka Connect dashboard.
  2. If the task failed, restart it through the Kafka Connect REST API: curl -X POST "$CONNECT_URL/connectors/orders-cdc/restart?includeTasks=true&onlyFailed=true"
  3. If the replication slot is growing, page the database owner.
  4. Confirm lag falls below 60 s, then resolve with a note.

Access matrix

Mapped to controls

Who can do what, enforced in the application and in the database, and reviewed each quarter against your identity provider’s groups.

Permissionops_agentteam_leadfinanceadmin
View casesown regionAllowedAllowedAllowed
Bulk assignNot allowedAllowedNot allowedAllowed
Request refundown regionAllowedNot allowedNot allowed
Approve refund > 50kNot allowedNot allowedAllowedNot allowed
Export customer dataNot allowedNot allowedNot allowedAllowed
Edit price bookNot allowedNot allowedAllowedNot allowed

Source code

main · checks passing

In your organisation’s repositories from the first commit, with branch protection, required reviews and a changelog. No code sits in ours.

  • apps/web · admin
  • services/accounts · orders · cases
  • workflows/Temporal definitions
  • data/dbt models · contracts
  • infra/Terraform · Helm
  • docs/this site
  1. a41f2c9orders: idempotent place order2 reviews
  2. 9f07e11cases: bulk assign preview2 reviews
  3. 71be3d0data: contract v3 for silver.orders1 review

Infrastructure as code

No drift

The same Terraform builds staging and production in your cloud account. Every change is planned in CI, reviewed and applied by the pipeline, never by hand.

$ terraform plan -out=release-42
  ~ module.orders.aws_ecs_service.api
      desired_count: 3 → 4
  ~ module.data.snowflake_warehouse.bi
      auto_suspend:  3600 → 60
Plan: 0 to add, 2 to change, 0 to destroy.

What changes

Measures we agree before we build.

We track 4 measures in order: use, speed, data quality and cost. We measure your starting point, agree targets and report monthly.

  1. People use it

    Do people choose to use it?

    • People using it weekly ÷ people it is for%
    • Work done beside the tool: exports and side sheetsper week
    Measured from
    Log-ins · usage data
    First check
    Weeks 1–2, before anything goes live
    Example target
    ≥ 80% using it weekly, 8 weeks after launch
  2. Speed of work

    Does the work move faster?

    • Quote to payment, typical and slowest 10%days
    • Case closing time, typical and slowest 10%hours
    Measured from
    Task timestamps, start to finish
    First check
    Weeks 1–2, before anything goes live
    Example target
    −30% typical quote to payment in 6 months
  3. Data quality

    Do the numbers agree?

    • Customer match rate%
    • Data updated on time% of hours
    • Bad data caught before loadingcount
    Measured from
    Data rules · dbt tests · monitoring
    First check
    Weeks 1–2, before anything goes live
    Example target
    ≥ 99% on time for key tables
  4. Cost

    Is it cheaper to run?

    • Licences and tools droppedcount
    • Warehouse spend per 1,000 queriescredits
    • Hours spent matching by handper week
    Measured from
    Invoices · warehouse meter · time checks
    First check
    Weeks 1–2, before anything goes live
    Example target
    3 tools dropped · −40% warehouse spend

Offers & plans

Custom software and data services, from CRMs to replacing old systems.

We build CRMs, staff tools, approval flows and the data warehouses behind your reports. Each one follows how you work.

Categories
04
Offers
16
Plans
05
Not sure what you need? Tell us the problem

How to buy

  1. 01Pick offers. Ask about one, or add several to one enquiry.
  2. 02Choose a plan. A small job, a fixed project or ongoing help.
  3. 03Send your enquiry. We reply within one working day.

Browse by category

Timelines are typical. We write the scope before we quote.

01Business systems

5 offers
Typical timeline: 12–20 weeks

Custom CRM

A CRM built around your own sales and service steps.

What’s included

  • Enquiries, customers, deals and service cases
  • All calls, emails and WhatsApp together
  • Access by role and change history
  • Sales reports and dashboards
  • Linked to billing, marketing and support
  • Built to your process
  • No per-user fees
  • Change history

Best forCompanies whose sales or service process has outgrown a generic CRM.

Typical timeline: 16–32 weeks

Operations systems (ERP)

One system for orders, stock, buying, production or field work.

What’s included

  • Build, buy or extend, per part
  • Order, stock and buying processes
  • Linked to accounting, SAP or Zoho
  • Data moved across, with check reports
  • Operations
  • Stock
  • Linked systems

Best forGrowing businesses running operations on spreadsheets or a patchwork of tools.

Typical timeline: 12–20 weeks

One record per customer

One profile per customer, with consent, merged from your website, app, CRM and sales data.

What’s included

  • Customers matched across channels
  • Consent recorded and respected
  • Live audiences for email and ads
  • Ready-made product or your own build
  • Customer matching
  • Consent
  • Build or buy
  • dbt

Best forBrands whose customer data is split across tools that disagree.

Typical timeline: 6–12 weeks

Approval and request flows

Online forms and approvals that replace email chains. Each request reaches the right person, and every decision is recorded.

What’s included

  • Forms, hand-off rules and reminders
  • Approval steps, with cover when away
  • Notifications by email, Slack or Teams
  • Decision history and speed reports
  • Approvals
  • Deadlines
  • Decision log
  • Slack
  • Microsoft Teams

Best forFinance, HR, buying and compliance teams chasing approvals by email.

Typical timeline: 4–10 weeks

Staff tools and portals

Office tools that swap spreadsheets and hand edits for safe screens, access rules and a change history.

What’s included

  • Admin screens on your current data
  • Access by role
  • Bulk edits, imports and exports
  • Change history for every record
  • Back office
  • Access rules
  • Change history

Best forOperations and support teams doing repetitive work by hand.

02Data systems

4 offers
Typical timeline: 8–16 weeks

Data warehouse

One home for your business data, with clear rules, so reports and AI use the same numbers.

What’s included

  • The right warehouse for your data
  • Feeds from apps, CRM and ads
  • Tested data models, with full history
  • Access rules and cost alerts
  • Warehouse
  • One source
  • Clear owners
  • dbt

Best forCompanies with data in many tools and no trusted numbers.

Typical timeline: 4–10 weeks

Automatic data feeds

Feeds that move data from your systems into the warehouse, on a schedule or live, with failure alerts.

What’s included

  • Connectors for online tools and databases
  • Scheduled feeds that retry on failure
  • Quality tests and late-data alerts
  • Each number traced to its source
  • Feeds
  • Scheduling
  • Data quality
  • dbt

Best forTeams whose reports depend on manual exports and copy-and-paste.

Typical timeline: 8–14 weeks

Live data streams

Systems that act the moment something happens, such as an order or a failed payment, without waiting overnight.

What’s included

  • Event design and a shared catalogue
  • Kafka or managed streaming set-up
  • Changes copied live from your databases
  • Replays, failure handling and monitoring
  • Events
  • Sync
  • Live

Best forBusinesses where minutes matter: logistics, payments, stock and fraud.

Typical timeline: 4–12 weeks

Moving data

Move data from an old system to a new one without losing records. We rehearse first and keep a way back.

What’s included

  • Field mapping and clean-up rules
  • Rehearsed data moves
  • Check reports on counts and samples
  • Switch-over plan with a way back
  • Rehearsed
  • Checked
  • Reversible
  • dbt

Best forAny move to a new CRM, system or cloud.

03Reports and metrics

4 offers
Typical timeline: 4–10 weeks

Dashboards

Dashboards for leadership and teams, with every number defined once and traceable to its source.

What’s included

  • Key numbers agreed with their owners
  • Power BI, Looker or your tool
  • Access limited by team or region
  • Scheduled reports and alerts
  • Metrics
  • Power BI
  • Looker
  • Power BI

Best forLeadership teams reporting from conflicting spreadsheets.

Typical timeline: 4–8 weeks

One set of metrics

One definition of revenue, active customers and lost customers, used by every dashboard, sheet and AI tool.

What’s included

  • Metric definitions written as code
  • Tests on key figures
  • A glossary of every data field
  • Open to dashboards and analysts
  • One definition
  • Metrics as code
  • Data glossary
  • dbt

Best forOrganisations where two reports never show the same number.

Typical timeline: 6–12 weeks

In-app dashboards

Charts and reports inside your own product, so customers see their data without exporting it.

What’s included

  • Each customer’s data kept apart
  • Charts in your own branding
  • Access rules per customer
  • Usage tracking
  • Customer-facing
  • Data kept apart
  • Branded

Best forSoftware businesses whose customers ask for reports.

Typical timeline: 3–6 weeks

App usage tracking

See how people use your app, where they drop off and which features they return to.

What’s included

  • A plan of what to track
  • Set up on web and mobile
  • Drop-off and return-visit views
  • Tracking with consent
  • Tracking plan
  • Drop-off
  • Repeat use

Best forProduct teams planning features without usage data.

04Replacing old systems

3 offers
Typical timeline: 3–6 weeks

Old system review

A review of an old system: what it does, what it costs and the safest way to replace it.

What’s included

  • Review of code, data and links
  • Hidden rules found with its users
  • Options: keep, move, fix or replace
  • A costed replacement plan
  • Review
  • Options
  • Costed plan

Best forOrganisations with a critical system that few people understand.

Typical timeline: 16–40 weeks

Step-by-step rebuild

Replace an old system piece by piece while it keeps running, so there is never a risky all-at-once switch.

What’s included

  • A layer around the old system
  • Parts replaced one by one
  • Old and new kept in sync
  • Retirement plan for the old system
  • Part by part
  • No big bang
  • On-off switches

Best forSystems too important to switch off and too costly to keep.

Typical timeline: 6–16 weeks

Move to the cloud

Move an ageing app from your own servers to the cloud, with automatic releases and tested backups.

What’s included

  • The app packaged to run anywhere
  • Server set-up as code
  • Automatic releases
  • Backup and restore tests
  • Containers
  • As code
  • Releases
  • AWS
  • Microsoft Azure

Best forApps on ageing servers or hosting.

Your enquiry

Tick “Add to enquiry” on any offer, choose a plan, then continue. Or enquire about one offer.

How we work with you

Ways to start, from a question to a quote.

Ask a quick question, send your project details or request a quote. The person who will lead the work reads each one in full. Services in your enquiry go with it.

Or book a 30-minute call

What are you sending?

  1. 01

    About 2 minutes3 short answers

    For a first chat, a press request or anything not ready to plan yet.

    You get A reply from a named lead

  2. 02Recommended

    About 8 minutes5 short steps

    Goals, who it is for, budget and timing, so our first reply can outline the work.

    You get Options and prices after one call

  3. 03

    About 15 minutesYour documents attached

    Your documents, deadlines and the purchasing and security rules to meet.

    You get Confirmation and a named lead

How it is priced

Each plan shows how it is priced. Work starts once we agree a written scope and price.

  • Typical length
    1–3 weeks
    Pricing
    Fixed fee
  • Typical length
    3–9 months
    Pricing
    Fixed price per stage
  • Typical length
    4–12 weeks
    Pricing
    Fixed price
  • Typical length
    6–18 months
    Pricing
    Priced per written scope
  • Typical length
    Ongoing · 3-month minimum
    Pricing
    Paid by time used
Compare what each plan includes
What each plan includes and who it suits
PlanWhat you getBest for
SprintA short, fixed job that answers one clear question.
  • Scope and result agreed before we start
  • A senior lead and the right specialists
  • A weekly review with you
  • A clear answer or a test version
An audit, a brand check or a test version of an app
MilestoneA bigger build in stages you approve and pay for one at a time.
  • Stages, each with its own scope and approval
  • A go or stop review at each stage
  • Re-plan between stages as you learn
  • Pay for each stage you approve
Work too big for one contract, such as a large app
ProjectOne clear job for one fixed price.
  • A written scope: what you get and how we agree it is done
  • A named project lead and a fixed team
  • A shared plan with checkpoint dates
  • Source files, yours once paid for
A brand identity, a website, a simple app or a set of tools
EnterpriseA large, long project with its own team and agreed response times.
  • A project director and a decision group
  • A team for each part of the work
  • Response times, reports and a list of risks
  • Security, legal and purchasing checks in the plan
Large companies making changes across markets, brands or business units
SquadOur specialists join your team and work in your tools.
  • Named specialists matched to your plan
  • Works in your own tools and meetings
  • Grow or shrink the team each month
  • We teach your team from week one
Teams with a clear plan that need senior people fast

Questions · 8 answered

Custom software questions, from build or buy to handover.

What owners and tech teams ask before they start. Ask anything else on the first call.

Bring to the first call

  • The process that hurts most, as people run it today
  • A screenshot of the spreadsheet it lives in
  • The tools it touches, and who owns each one
Start a software project

Usually both, decided part by part. Buy where a ready-made product fits a standard process. Build where your process sets you apart. We write down why.

Yes, and it is often the right call. We keep your main system and add custom apps and two-way sync around it. We replace it only when extending costs more.

In your own cloud account, in the region you choose. We get only the access we need, and you can remove it. We copy nothing to ours. Personal data can stay in India or the EU.

One part at a time, behind a switch. We rehearse each data move and check old against new nightly. Any part can switch back until the old one is off.

Start with the data warehouse, which holds all your data. A customer data platform (CDP) keeps one record per customer and shares it with your tools, as consent allows. Add it once the warehouse data is trusted.

For routine data work, approved by a named engineer. Agents label columns and check data moves. They also draft documents and tests. They never approve their own work, release code or change access. Every action is logged.

Your team or ours. Your team can take over with our guides and a handover period. Or our engineers run it under Integration & Support. Explore Integration & Support

Typically within 2 months, for one working flow on live data. A first release of the main parts typically takes 12–24 weeks, depending on how many systems it replaces.

Tell us what you need built.

You will talk to the lead who would run your work.

Book a call

Three ways to start

We always agree a written scope and price before work begins.

Choose one of the three ways above