Financial servicesTokenization platform
Case study available on request, under NDA.
AI · Platforms · Automation
We build the engines other businesses run on.
CRM engines, data platforms, enterprise AI agents and tokenization platforms. Designed, built and shipped to production by Agatech, and yours to keep.
- NDA by default
- IP assigned to you
- Senior engineers only
Enterprise AI agentsCustom platformsCRM enginesData platformsTokenization platformsWorkflow automationData pipelinesAnalytics layersLLM evaluationSmart contractsAdmin backofficesPublic APIs
Next.jsTypeScriptPostgresRedisKubernetesTerraformdbtAirflowClickHouseSolidityLangGraphTemporalAnthropicOpenAIpgvectorKafka
- Platforms shipped
- —counted from signed hand-overs
- Agents in production
- —counted from live deployments
- Senior engineers
- —the whole team, no juniors on client work
- Years building platforms
- —published once confirmed
What we build
Six kinds of engine. One team.
Each one is a product we have built before, not a service line we are willing to try. Pick the closest and tell us where yours differs.
Enterprise AI agents
Agents that read, decide and act inside your systems, with evals, permissions and an audit trail from day one.
- evals
- tool permissions
- audit trail
Custom platforms
Multi-tenant platforms built to be operated: auth, billing hooks, admin, observability and documentation included.
- multi-tenant
- admin backoffice
- documented
CRM engines
The customer record, pipeline and automation layer underneath your product, with AI scoring and enrichment built in.
- lead scoring
- pipelines
- enrichment
Data platforms
Pipelines, warehouses and analytics layers that make your operational data reliable enough to run the business on.
- pipelines
- warehouse
- analytics
Tokenization platforms
Issuance, custody flows and ledgers for tokenized assets, with the compliance seams designed in rather than bolted on.
- issuance
- ledger
- compliance seams
Workflow automation
Pipelines that move work between your systems reliably: retries, idempotency, monitoring and a human step where it belongs.
- idempotent
- observable
- human-in-the-loop
How we deliver
From brief to production in four moves.
The same four steps for every engine. Scope written down before anything is built, working software every two weeks.
- 01typically one to two weeks
Discover
We read your brief, ask the awkward questions and write down what we would build.
- A working session with the people who own the problem
- Constraints, integrations and risks listed before any design
- A scoped proposal in writing: what, how, and what it depends on
- 02typically two to four weeks
Design
Architecture, data model and a clickable prototype you can put in front of users.
- System design reviewed with your engineers, not handed to them
- Interfaces and contracts agreed before the build starts
- A prototype that settles the questions a document cannot
- 03scoped per project, agreed in writing
Build
Two-week sprints, a working demo at the end of each, and the code in your repository throughout.
- Senior engineers only; the people who scoped it build it
- Tests where failure would be expensive, observability from the first deploy
- A demo every sprint and a written status note you can forward
- 04ongoing, on your terms
Run
Hand-over with documentation and runbooks, or managed operation by us for as long as you want it.
- Runbooks, architecture notes and an onboarding session for your team
- A warranty period for defects at no additional charge
- Optional managed operation with an agreed response time
agatech — build
- agatech scaffold crm-engine --with ai-scoring
- schema accounts, contacts, pipelines, events
- auth roles, scopes, audit log
- agents lead-scorer, dedupe, enrich
- evals 142 cases, 0 regressions
- deploy preview → https://crm.preview.example
- done in 4m12s
Built the way engineers build
Production from the first commit.
The same tooling, tests and observability we would want to inherit, in your repository from day one.
- Platforms, not prototypes
Built to be operated on day one and understood on day five hundred.
- typed end-to-end
- tested
- observable
- documented
- Agents with guardrails
Autonomy where it is safe, a person where it is not, a record either way.
- evals
- tool permissions
- human-in-the-loop
- audit trail
- Your stack, your IP
No proprietary runtime, no lock-in, nothing you cannot take with you.
- code in your repo
- no lock-in
- hand-over included
- NDA by default
Outcomes
Work we can talk about.
Most of what we build sits behind an NDA. Case studies are published only with written approval, and attributed by role, never by name.
- Every engagement under NDA
- Signed before the first call if you prefer.
- IP assigned on hand-over
- The code is yours; we keep no licence to it.
- One senior team, start to finish
- The engineers who scope the work build it.
EnergyData platform
Case study available on request, under NDA.
B2B SaaSCRM engine
Case study available on request, under NDA.
LogisticsEnterprise AI agents
Case study available on request, under NDA.
InsuranceWorkflow automation
Case study available on request, under NDA.
MarketplaceCustom platform
Case study available on request, under NDA.
Questions
The ones we are asked first.
Do you work with existing codebases?
Yes. Most engagements start inside a system that already exists. We begin with a review of what is there, agree what to keep, and build around it rather than replacing it for its own sake.
Who owns the code?
You do. Work happens in your repository from the first commit and the intellectual property is assigned to you on hand-over. We keep no licence and use no proprietary runtime.
How does an engagement start?
With a brief. Send it through the contact page, an engineer reads it, and we come back with questions and then a scoped proposal in writing. Nothing is built before the scope is agreed.
Where is the team?
Agatech LTD is a remote engineering company working across European and UK time zones, with overlap arranged for teams elsewhere. We travel for kick-offs and launches when it helps.
Which models and clouds do you use?
Yours, by preference. We build on the cloud accounts you own and route AI work across providers behind a single interface, so the model can change without the product changing.
Tell us what you need built.
A short brief, a real engineer reading it, and a scoped answer within a few business days.