AI delivery

Put AI to work.Keep control.

We take one real workflow from idea to production. Built in your accounts, tested against failure, secured, and handed to the team who will run it.

AWS Advanced Tier Services Partner · OpenAI Partner · 5 certified engineers
What you get

One workflow in production. Evidence your reviewers can inspect.

Your account. Your repositories. Scope agreed before work starts.

01 /
A scoped integration

An agent built around a real workflow.

02 /
Controls & evidence

Permissions, evaluation and agreed tests.

03 /
An operating plan

Monitoring, a runbook and a named owner.

A controlled AI-agent workflow
Conceptual workflow with scoped tools and operating evidence.
Our work

AI that shipped, and still runs.

View all work

We also build and operate software ourselves. tMinus1 built Automatum, a multi-tenant platform with marketplace integrations and metered billing. That work informs how we approach identity, integration and ongoing operations.

Platform engineering experience; this case study does not demonstrate an AI outcome.

Automatum case study
Start with the work

One useful workflow.
A clear reason to build.

We start with the handoffs, repetitive decisions and information gaps your team deals with every day. Then we decide where AI helps, where conventional software is enough and where a person stays in control.

01 / KNOWLEDGE

Find the answer.
Show the source.

Search internal documents and return answers with references. Respect source permissions, test retrieval quality and define what happens when the evidence is missing.

Example scope · Internal knowledge assistant
02 / OPERATIONS

Move the task.
Keep the approval.

Extract information, classify a request and prepare the next action in an existing system. Require human approval before sensitive or irreversible changes.

Example scope · Service request triage
03 / DOCUMENTS

Read the detail.
Flag the exception.

Turn incoming documents into structured records. Validate required fields, surface uncertainty and send exceptions to the person who owns the process.

Example scope · Document processing workflow

Illustrative use cases. We agree feasibility, data access and acceptance criteria with your team.

How we build it

A complete system.
Not just a model connection.

The model is only part of the system. We design the application, data, identity and operating layers together, with controls from the outset, so the workflow has a defined place in your environment.

01 / EXPERIENCE

Where your team works

A business application, internal interface or service workflow. Clear inputs, useful responses and an obvious route to a person when the system cannot complete the task.

02 / ORCHESTRATION

How the work moves

Task routing, model requests, retrieval and tool calls. Explicit limits on what can run, how long it can run and which actions need approval.

03 / CONTEXT & TOOLS

What the agent can reach

Approved documents, search indexes and application APIs. Permissions, source ownership and data quality are part of the design, rather than assumptions made by the model.

04 / PLATFORM

What keeps it accountable

Identity, secrets, deployment, logging, evaluation and cost visibility. Separate environments and a documented route to release, rollback and recovery.

Illustrative architecture. The components and deployment pattern depend on the agreed workflow and environment.

Connections and boundaries

Knowledge & retrieval

Identify the sources that are authoritative for the task. Work through document ingestion, metadata, retrieval and references, including how content is updated or removed.

  • Source ownership and freshness
  • Permission-aware retrieval design
  • Answers grounded in accessible evidence
  • A defined response when evidence is insufficient

Business system integration

Connect through agreed APIs and service identities. Separate reading from writing, validate inputs and decide which actions should be prepared for approval rather than executed automatically.

  • Explicit API and tool contracts
  • Scoped credentials and secret handling
  • Validation, retries and duplicate handling
  • Human approval for consequential actions

Data handling & environment

Map how information moves through the application, model and supporting services. Agree the handling of sensitive information and the records needed for your review.

  • Data flow and service inventory
  • Environment and account boundaries
  • Logging, retention and redaction decisions
  • Access responsibilities agreed with your team

Design principles

01

Access with a purpose

A scoped identity, approved data sources and explicit tool permissions. The agent only gets the access its task requires.

02

Evidence before confidence

Representative test cases, retrieval checks and agreed failure scenarios. Test prompt injection, data exposure and unauthorised tool use within an approved environment.

03

A way to stop and recover

Human approval points, exception handling, audit traces and monitoring. A named owner knows when to intervene and how to operate the system.

How we prove it

A good demo is a start.
A test set tells you more.

We agree what useful means before treating the workflow as ready. Evaluation combines representative work, difficult cases and the failures your process cannot afford to overlook.

EVALUATION PLAN / ILLUSTRATIVEEvidence → review → decision

Task quality

Can it complete the agreed job?

Check answers or structured outputs against representative examples and a documented review rubric.

Grounding

Can the answer be supported?

Inspect retrieved material, source references and behaviour when the relevant information is missing or contradictory.

Action boundaries

Does it stay within its authority?

Exercise tool permissions, approval gates and requests outside the intended task. Record blocked actions and unexpected behaviour.

Operating behaviour

Can the workflow run practically?

Observe latency, request volume, model usage and failure handling under agreed conditions. Use the findings to discuss service expectations.

Changes need evidence too.

A different prompt, model, document set or integration can change behaviour. We leave a repeatable evaluation approach so your team has a way to assess subsequent changes against the agreed baseline.

Test the paths you hope never happen.

Adversarial testing is scoped with your team. We agree the targets, environment, access and test window, then document the observed behaviour, its implications and the response.

INPUT /

Hostile instructions

What happens when a request or retrieved document tries to redirect the agent away from its task? Test whether untrusted content can influence tools, outputs or approval decisions.

IDENTITY /

Permission escalation

Can one request reach information or functions beyond the caller’s authority? Exercise the boundaries between the user, agent identity and connected service.

DATA /

Unintended disclosure

Can sensitive information appear in a response, tool call or diagnostic record? Examine the agreed data paths and the controls around them.

FAILURE /

Unsafe continuation

Does the workflow stop, retry or escalate when a tool fails or an answer is uncertain? Check that incomplete work is visible and consequential actions stay controlled.

Testing reduces uncertainty within the agreed scope; it cannot establish that a system is free of vulnerabilities.

An example engagement

From scope to handover.
With review points throughout.

A typical build runs six to eight weeks. The example below shows a six-week shape for a contained workflow. We confirm the actual schedule after reviewing access, integrations, complexity and your approval process.

  1. WEEK 01

    Define the job

    Map the workflow, data and permissions. Agree success measures, exclusions and who approves the work.

  2. WEEKS 02–03

    Build & evaluate

    Connect the systems and develop the workflow with representative test cases and controls in place.

  3. WEEKS 04–05

    Challenge & harden

    Run agreed adversarial tests, investigate failures and refine permissions, guardrails and monitoring.

  4. WEEK 06

    Review & hand over

    Walk through evidence and remaining limitations. Your team decides deployment and takes ownership.

What you take away.

Source code, evidence and a handover. Your team keeps the implementation and the context behind the decisions.

BUILD /

An integrated workflow

Source code in your repositories, integrations with agreed systems and configuration in your environment. Acceptance criteria connect the build to the process you want to improve.

VERIFY /

An evaluation & test report

The scenarios tested, observed failures, changes made and remaining limitations. Your reviewers can see the evidence rather than rely on a demonstration.

CONTROL /

A security evidence pack

Identity and permissions, data flows, guardrails, logging and monitoring documented for your internal review. Deployment remains your approval decision.

OPERATE /

A handover your team can use

A walkthrough, runbook, ownership and escalation paths. Ongoing engineering or operations support can be agreed separately if your team needs it.

Someone has to own Tuesday morning.

The system needs an owner when a document changes, a tool stops responding or a user reports a poor answer. We plan that responsibility during the build. Your team can take ownership, or we can scope continuing engineering support.

Observe the workflow

Agree useful traces, error signals and usage visibility. Connect operational information to the team responsible for responding.

Manage changes deliberately

Version implementation and configuration. Re-evaluate material changes and maintain an agreed release and rollback process.

Keep the economics visible

Understand the drivers of model and platform usage. Review usage patterns and limits alongside quality and response time.

Maintain a human route

Document ownership, escalation and manual alternatives. Make it clear when the workflow should pause or return work to a person.

Ways in

Bring the problem.
Or bring the unfinished pilot.

You do not need to arrive with an architecture or a model selected. We can help define the first workflow, engineer a scoped implementation or assess what is preventing an existing pilot from moving forward.

DISCOVER

Find a viable first workflow

Work through the business process, available data, integration constraints and ownership. Leave with a proposed scope, acceptance criteria and the uncertainties to resolve before building.

Starting point: a process and a technical owner.

BUILD

Engineer the implementation

Deliver the agreed workflow with integrations, evaluation, controls and handover. Review progress against concrete examples of the work and the acceptance criteria.

Starting point: a bounded use case and access to the systems.

STRENGTHEN

Work through the production gaps

Inspect an existing pilot’s architecture, permissions, evaluation and operations. Prioritise the gaps and agree which engineering changes are needed next.

Starting point: a pilot, repository or system to assess.

These are engagement starting points, not fixed packages. Scope, fees and responsibilities are agreed in writing.

Two faster ways in.

Not sure which workflow comes first? Start with the free sprint. Already have a use case and an owner? Go straight to a pilot.

A good fit starts with an owner.

Bring a process owner, examples of the work and someone who understands the connected systems. We will work through feasibility and scope together.

Clear commercial boundaries

Scope before price. We quote a fixed fee after agreeing deliverables, assumptions and acceptance criteria. Changes are agreed in writing.

Your environment. Your code. Access and responsibilities are agreed before the build starts. The implementation stays with you.

Funding is conditional. Qualifying projects may be eligible for AWS support, subject to AWS assessment and approval. We confirm eligibility rather than assume it.

Useful answers

Before we start.

Discuss anything specific to your team or environment.

Why only one process?

Because scope is what kills these projects. One process gives us a real number, a contained blast radius, and a security review that can actually finish. Once the first one runs, the second is faster.

Will this pass our security review?

We prepare evidence for your reviewers: evaluation results, permissions, data flows, audit logging and monitoring. We test in an agreed environment and document remaining limitations. Your team makes the approval decision; approval is not guaranteed.

Which model or framework do you use?

Whichever one the process needs, and we will argue for the boring choice. Model selection matters far less than scope, permissions and monitoring, which is where these projects actually fail.

What if the adversarial testing breaks something?

We agree the environment, targets, permissions and test window in advance. Findings are documented and prioritised, with fixes and retesting included in the agreed scope.

What happens after the handover?

The system is yours, running in your AWS account. Some clients keep us on a monthly engineering and operations retainer so they do not staff for it. Others take the runbook in-house. Both are fine, and we will tell you which we think fits.

Can AWS fund part of this?

It may. We apply for proof-of-concept funding on qualifying projects and tell you in the first call whether yours is likely to qualify. Funding is at AWS's discretion and is never guaranteed.

Further reading

The engineering behind the decisions.

AI delivery — built, broken, secured, handed over running
Request a technical call