Agentic delivery for small and mid-sized product teams

Agents write the code. Your team makes the decisions.

Experienced engineers build an agentic pipeline from ticket to production inside your company in about 4-6 weeks, and teach your team to run it.

Behaviour tests (BDD) from day one

The problem

Almost everyone uses AI. Almost no one lets agents work on their own.

90%
of professionals use AI at work DORA 2025
61%
never use autonomous agent mode DORA 2025
30%
have little or no trust in AI-generated code DORA 2025

Writing code is not the bottleneck. Verification and decisions are. Agents code, test, do the first review and prepare the release; people make only the decisions a machine should not make.

Tests first

We start with tests of how your product behaves.

An agent is only as good as the thing it checks its work against. A Given / When / Then scenario is at once the acceptance criterion, the agent’s task and the test that verifies it.

features/invoices/export.feature
Feature: Export invoices to CSV   Scenario: Accountant exports invoices for a selected month    Given 3 invoices were issued in October    When the accountant exports October invoices    Then they receive a CSV file with 3 rows    And each row contains the VAT ID and gross amount
$ behave 1 scenario (1 passed) · 4 steps (4 passed)
01

The scenario is the contract

The agent implements and runs the scenarios. When they pass, the change goes to review. When they fail, it goes back to the agent.

02

The scenario is the documentation

An always up-to-date description of what the product does. New people and new agents learn from the same files.

03

The scenario prevents regressions

Every change, human or agent, runs through all existing scenarios.

Does your product have tests like this? If not, that is where we start.

Your current testsWhere we start
BDD scenarios or good functional tests Pipeline on the first ticket class straight away
Unit tests only About 1 week of BDD foundation for the first ticket class
No functional tests 1-3 weeks of BDD foundation for the key paths, partly in parallel with the pipeline build

Underneath sits the full pyramid: unit, integration, API contract and browser e2e tests. But the scenarios decide whether a ticket is done.

Process

Five steps. Pipeline in production in about 4-6 weeks.

We work with agents ourselves: our agents map your repo, tests and CI, draft the first scenarios and build the configuration from proven templates. People handle the conversations, design, decisions and accountability.

  1. We learn your product, company and workflow

    3-5 days

    Interviews with the team and decision makers. Our agents map repo, tracker, CI and scenario coverage in a few hours.

    Outcome Readiness report with a test assessment, pipeline design, first ticket class.

  2. BDD foundation

    1-3 weeks · if tests are missing

    Agents describe the existing product behaviour as scenarios and your team approves them. Only then do agents change code.

    Outcome Key product paths covered by scenarios and running in CI.

  3. We build the AI agent pipeline

    2-3 weeks

    Context for agents, quality gates, scenario and QA agents, implementation and review agents, a question layer, morning report, integrations.

    Outcome First agent tickets in week one of the build, an end-to-end pipeline on the first repo.

  4. Continuous delivery (CI/CD)

    1-3 weeks · in parallel

    CI on every merge request, preview environments, promotion development → staging → preprod → production, acceptance tests on selected environments, release checks, monitoring, fast rollback.

    Outcome Agent work reaches production regularly and safely.

  5. We train and hand over

    1-2 weeks · in parallel

    Workshops for developers, leads and product owners, pairing on running the pipeline, a runbook.

    Outcome Your team runs, tunes and extends the pipeline on its own.

  6. We tune and grow the pipeline

    ongoing, min. 3 months

    Prompts and gates improved from metrics, models matched to roles, more ticket classes and repos, more parallel agents.

    Outcome More tickets done without rewrites, lower cost per ticket.

Typical timeline of a small engagement

wk 1wk 2wk 3wk 4wk 5wk 6 Discovery and readiness report BDD foundation Pipeline on the first repo CI/CD Training and handover Tuning and care

Larger engagement (several repos or teams): about 6-8 weeks. Pace depends on access to repo, tracker and CI within the first 2 days and a decision maker who answers questions within one working day. Steps can be bought separately.

How it works

From request to production. Agents in the loop, decisions with people.

One flow, several layers. Pick a layer to see who does what: the development pipeline, the QA pipeline, the deployment pipeline and the points where your people decide.

From request to production. Tickets come from your people, production users and the QA agent, the PM agent reviews and assigns them, and agents work in loops up to the merge. Then CD promotes the change through the environments.

CI fails review fixes unclear ticket answer new tickets user bugreports ← CD: promotion between environmentsreport: what to fix and whyno agent access Your people Ticket PM agent Scenario agent Developer agent CI on themerge request Reviewer agent Lead review Decision: merge Development Staging Preprod Production Clarification QA agent Monitoring agent Release approval CI fails reviewfixes newtickets user bugreports CD: promotion ↓no agent access Your people Ticket PM agent Scenarioagent Developeragent CI on themerge request Revieweragent Lead review Decision:merge Development Staging Preprod Production Clarification QA agent Monitoringagent Releaseapproval
agent human system no agent access

Agents work in an isolated zone with limited permissions and a full log of their actions. They have no access to production. Every ticket gets a note from the agent: what was done, how and why.

The full delivery chain

AreaWhat agents doWhat we oversee
Backend APIs, logic, integrations, migrations, unit and integration tests Architecture, contracts, security
Frontend Screens from the design system, states, forms, component tests Fidelity to design, accessibility
BDD tests Scenario agent: Given / When / Then scenarios from the spec, characterisation tests, full suite on every change. QA agent: edge and corner cases on staging and preprod Scenario quality, product owner sign-off, path coverage
Other tests Unit, integration, API contract, browser e2e, regression Test strategy, quality gates, test data
Review Checks against the ticket, fixes minor defects Final decision, review rules
Releases Release checks against tickets, acceptance tests on staging and preprod, pipeline monitoring with a report on what to fix and why, merge request hygiene Promotion between environments, release approval, fixes on higher environments, rollback
Security Work in an isolated zone with limited permissions, every action logged Production access for people and the pipeline only
Process Triage of user reports (ticket or answer), backlog prioritisation, clarifying questions, new tickets from tests, ticket updates (what was done, how and why), morning report Who decides and when

When an agent needs a decision

An agent does not guess. It asks in the ticket, with a ready answer.

When a requirement is unclear or there are several reasonable solutions, the agent moves the ticket to “Needs clarification” and writes one comment. You answer in a line, without a meeting.

#142 Validate VAT ID in form Needs clarification

developer agent03:40

Problem

The ticket says “validate the VAT ID”. A number can be checked by format only, or online in the EU VIES registry, which is sometimes unavailable.

Question

Should the order be blocked when VIES does not answer?

Options

  1. A

    Format check only

    Fast and never blocks, but accepts made-up numbers.

  2. B

    Format now, VIES in the background recommended

    The order goes through, an invalid number flags the invoice for the accountant.

  3. C

    Block until VIES confirms

    Safest, but orders fail whenever VIES is down.

Recommended path

B. The checkout scenario “order completes in under 3 s” would fail with C, and B follows how the payment webhook already retries. Meanwhile I am working on the other scenarios of this ticket.

product owner08:15

B. And email the accountant when an invoice is flagged.

developer agent08:16

Added the scenario “Invalid VAT ID flags the invoice and notifies the accountant”. Resuming.

01

The agent stops instead of guessing

Unclear requirements, several reasonable solutions or a change in product behaviour mean a question, not an assumption hidden in the code.

02

One comment, everything needed to decide

The problem in context, one concrete question, two or three options with their consequences and the path the agent recommends, with the reason.

03

The answer goes straight into the work

The agent turns the decision into a BDD scenario, updates the ticket and resumes. The decision stays in the ticket history, open questions appear in the morning report.

QA agent

The QA agent finds scenarios on its own. Including the ones nobody thought of.

This is not a standard tester. The scenario agent writes the main-path scenarios before any code. The QA agent works after development: it reads the specification, acceptance criteria and the scenarios in the repo, tests changes on staging and preprod, and deliberately hunts for edge and corner cases. Your team decides what the expected behaviour is.

QA agent pipeline

  1. 1

    Reads the specification

    Ticket, acceptance criteria, mock-ups, API contracts and existing scenarios.

  2. 2

    Checks the main path

    Runs the scenario agent’s scenarios on staging and preprod, with real integrations and test data.

  3. 3

    Hunts for edge and corner cases

    Boundary values, empty and invalid data, permissions, concurrency, time zones, integration failures.

  4. 4

    Asks when the spec is silent

    If the specification does not say how the product should behave, the case goes to clarification instead of guesswork.

  5. 5

    Opens tickets on its own

    When a test exposes a bug in existing code or a coverage gap, the agent opens a ticket with a scenario that reproduces it. The PM agent reviews it, sets a priority and assigns it to other agents.

  6. 6

    Learns from your decisions

    Approved and rejected scenarios go into the project’s pattern library. Next proposals get closer to how your team thinks.

the next ticket starts with more knowledge

QA agent · export invoices to CSV

From the invoice export scenario, the QA agent derives cases the ticket does not mention:

  • edge Month with no invoices: file with just the header approved
  • edge Credit note with a negative gross amount approved
  • corner Invoice from 31 Oct, 23:30 UTC, accountant in Europe/Warsaw needs clarification
  • edge EU customer: VAT number instead of a Polish NIP approved
  • corner 20,000 invoices in a month: background export and an email link needs clarification
  • edge User without invoice permissions approved
  • edge Semicolon in the customer name breaks CSV columns bug → new ticket SHS-152
  • edge Invoice with no line items rejected: already covered

8 proposals · 4 approved · 2 need clarification · 1 new ticket · 1 rejected

The QA agent does not replace the tester on your team. It gives them broad coverage to start with, and a human decides which cases matter to the business.

Comparison

Classic team vs. agentic team

Typical scope: developing a web product, a dozen to several dozen tickets a week.

Classic teamAgentic team with us
Team 4-5 developers, often plus QA and DevOps 1-2 experienced leads plus agents; your team decides and reviews
More capacity Hiring 1-2 senior developers: a full salary each and 2-4 months to find them One-off engagement from about €26k, then about €250-600/month of AI budget per agent slot
First results A new senior needs weeks to learn the project First agent tickets in week one of the build
Working hours About 8 hours a day, working days Agents around the clock, people decide in their own hours
Tests Depends on discipline, often postponed BDD scenarios on every ticket, full suite before a human sees it
Review All work goes to people Agent does first review and fixes; a human gets a recommendation
Unclear requirements Developer guesses or waits for a meeting Agent stops and asks a written question in the ticket history
Project knowledge In people’s heads, leaves with churn In context files and configuration in your repo
Reporting Status meetings Morning report and monthly metrics: cost per ticket, lead time
Where a classic team is better Architecture decisions, working with an unclear product, security-critical code. Agents help here, but decisions and accountability stay with people.

We do not promise agents will replace your team. The model works best when your people shift from routine code to decisions, review and hard problems.

Tool setup vs. a pipeline in production

There are cheaper services that set up Cursor or Claude Code in your repo. That is a good first step, but a different product.

Tool setupPipeline with us
Outcome The tool is configured in the repo, a developer works with an agent in the editor Agents take tickets from the tracker and deliver a finished, tested branch
Tests Whatever the repo has today BDD foundation, scenario and QA agents, full suite on every change
CI/CD Unchanged CI on the merge request, promotion through environments, release approval
Unclear requirements The developer asks the agent in chat The agent asks your team in the ticket, the answer stays in its history
Measurement How the team feels Per-ticket metrics: cost, lead time, share without rewrites
When we leave Tool settings A pipeline, a runbook and a trained team that runs it

Benefits

What you get

  • 01

    Faster to production

    Agents work in parallel and overnight. Work is waiting for approval in the morning.

  • 02

    Lower delivery cost

    You pay for an experienced lead and an AI budget, not a whole team by the hour.

  • 03

    Scale without hiring

    More work means more agent slots, not new headcount.

  • 04

    Quality before a human

    BDD scenarios, tests, agent review and release checks on every ticket.

  • 05

    Living documentation

    Scenarios are always current, because every change has to pass them.

  • 06

    Transparency

    Every decision, question and change is recorded. You see the cost and time of each ticket.

  • 07

    Knowledge stays with you

    Configuration, prompts, gates and runbook as code in your repo.

  • 08

    More interesting work

    Less routine, more design and decisions for developers.

  • 09

    No lock-in

    Models from several vendors, your infrastructure, your tools.

Scaling

We scale agents with configuration, not hiring.

From a few to dozens of parallel slots, around the clock. We scale together with your budget and the pace of your decisions, because more agents also means more questions and more branches to review.

Indicative AI budget
1,500–3,600 € / month
Fits
Small team, one decision maker

Estimate, confirmed in step 1. Alert at 80% of budget, overage billed at cost + 20%.

Results

What we have already built

The first 5 weeks of a production pipeline we built from scratch: a dozen services, running around the clock, on a foundation of BDD tests.

684 tickets picked up by agents
63% brought to a finished, tested branch
53% without a single question to a human
85% of finished branches approved by the lead
20 production releases in 5 weeks
6,900 BDD scenarios as the main acceptance layer

This was possible because the product had an extensive BDD scenario suite. Second lesson: the real bottleneck is how fast humans make decisions.

Who does it

Experienced people, not a platform.

The founding team is two experienced engineers: full stack, DevOps, cybersecurity, project management and head-of roles. They spent years building and shipping products, and now design and run production agentic pipelines.

First Last

Delivery and agentic technology lead

15+ years

  • Backend
  • Agentic pipelines
  • BDD

First Last

Client and training lead

15+ years

  • Architecture
  • CI/CD
  • Training
  • A small engagement is led by one lead, a larger one by two people.
  • The same lead takes you from discovery to pipeline handover.
  • Every new team member goes through our training programme and works to the same playbook.

Integrations

We work in your tools.

We don’t change your tracker, repo or chat. Integration through APIs and MCP servers, with permissions kept to a minimum.

Trackers

  • Jira
  • Linear
  • GitHub Issues

Repo and CI

  • GitLab
  • GitHub
  • GitLab CI
  • GitHub Actions

Chat

  • Slack
  • Microsoft Teams

Monitoring

  • Sentry
  • Grafana
  • logs and metrics

Docs

  • Confluence
  • Notion
  • Figma

Cloud and models

  • AWS
  • GCP
  • Azure
  • Kubernetes
  • Bedrock
  • Vertex

Training

We teach your team to work this way.

For engagement clients this is step 4. For other companies it is a standalone product, including those building a pipeline themselves.

Ask about training

2 days, up to 12 people

Working with agents as a team

Developers

Context for agents, reviewing agent changes, working with quality gates. On your code.

€5,200

1 day, up to 12 people

BDD and tickets for agents

POs, analysts, testers, developers

Acceptance criteria as Given / When / Then scenarios an agent can do without questions and a machine can verify.

€2,100

2 days, up to 12 people

BDD in practice

Product teams

A workshop on your code: first scenarios for the most important product paths.

€4,700

1 day, up to 8 people

Agentic pipeline lead

Tech leads, EMs

Metrics, cost per ticket, who decides and how fast.

€2,100

1-3 weeks pairing

Pipeline takeover

1-3 of your people

Your people tune, scale and fix the pipeline on their own.

€5,900

2 days, groups up to 12

Open training

Individuals

The basics of the model and a chance to meet us. Remote, or on site in the EU.

€690 / person

Honestly

Where agents work and where they don’t

We don’t promise “10x”. We promise measurement: a monthly report with metrics that are easy to check.

Works well

  • Tickets with BDD scenarios or clear criteria
  • CRUD, API integrations, migrations, tests, docs
  • Bugfix with a reproduction
  • Screens from an existing design system

Works poorly

  • Vague requirements like “make it better”
  • Architecture decisions and cross-cutting changes across services
  • Performance, concurrency, security-critical code
  • Code without functional tests, until we build the BDD foundation

In the monthly report

  • behaviour coverage by BDD scenarios and its growth
  • share of tickets merged without rewrites
  • share of tickets with a question and response time
  • AI cost per merged ticket
  • lead time from ticket to production and release count
  • change failure rate and review time

What we need from you

  • a tracker with tickets
  • a repo and CI, or readiness to build CI in step 3
  • an environment where tests can run
  • a product owner who approves scenarios
  • one decision maker who answers agent questions daily

Pricing

A fixed price for the outcome of each stage, not for time.

The speed agents give us is our advantage, not your cost. Prices net of VAT.

A good place to start

€0

Readiness scan

A 30-minute call and a short questionnaire about your repo, tests and CI. Within 2 working days you get a readiness assessment and a recommendation on where to start.

€2,000

Discovery and design

3 working days, a readiness report with a test assessment and a pipeline design. Credited towards the engagement if you start it within 60 days.

from €5,000

BDD foundation

Scenarios for the key paths, running in CI. Valuable on its own.

from €12,000

Pilot with a success criterion

3-4 weeks, one ticket class, one repo. We write the success criterion into the contract after step 1. If we miss it, you don’t pay the last instalment. The report covers tickets, stories, bugs, cases from the QA agent, automated checks and human decisions.

Small engagement

5-20 developers, 1-2 teams, 1 main repo, one decision maker

about €26-29k

Time
about 4-6 weeks
Then
care €3,900/month, AI budget up to €1,200/month included
From us
1 lead supported by our agents
Goal
Ship the roadmap faster with the same team.
What’s included
1. Discovery and design €2,000
BDD foundation (if tests are missing) from €5,000
2. Pipeline build €11,500
3. CI/CD from €4,000
4. Training and handover €4,000
5. Tuning and care (min. 3 months) €3,900/month
Start with step 1

Larger engagement

20-80 developers, 3+ teams or several products, several decision makers

about €52-59k

Time
about 6-8 weeks
Then
care €7,500/month, AI budget up to €2,400/month included
From us
2 people supported by our agents
Goal
One way of working with agents across the organisation.
What’s included
1. Discovery and design €3,800
BDD foundation (if tests are missing) from €10,000
2. Pipeline build (up to 2 repos) €19,000
3. CI/CD from €9,000
4. Training and handover €7,500
Each additional repo or team from €7,500
5. Tuning and care (min. 3 months) €7,500/month
Start with step 1

Training

For client teams and companies building a pipeline themselves

from €2,100

Time
1-2 days or 1-3 weeks of pairing
Then
open training €690 per person
From us
Led by the founders
Goal
Your team works the agentic way without us.
Ask about training

The variant depends on the scope of the engagement, not the size of the company. A company with 60 developers can start with a small engagement in one team.

  • Care lite after handover: €1,300/month. Up to 8 hours of work a month, a monthly metrics review, replies within 2 working days, AI budget on your side.
  • First three companies: 15% off in exchange for permission to publish metrics and a case study.
  • Scaling agents beyond budget: cost + 20%, alert at 80% of budget.
  • Regulated variant: prices without AI budget + 10% management fee, models on your own cloud account.
  • We don’t bill per merged PR, because “merged” doesn’t mean “correct”. Results show up in the report, not the invoice.

FAQ

Frequently asked questions

We already have Cursor or Claude Code. Why do we need you?

A tool is not a process. We build the system around it: context, quality gates, questions, reports, CI/CD and team training. You’d get there by trial and error in a few months; with us it takes a few weeks.

We have more than 50 developers. Is this for us?

Yes. We most often work with companies that have from a few to a few dozen developers, but we are happy to talk to larger ones. In a big organisation we start with one team and one repo, then expand.

Will agents replace our developers?

No. They change their work: less routine, more decisions, review and hard problems.

We have no functional tests. Does this make sense?

Yes, and that is exactly where we start. The BDD foundation comes first: we describe product behaviour as scenarios, and only then do agents change code. That stage alone gives you value, even without further steps.

Why BDD and not just unit tests?

Unit tests check the code; BDD scenarios check that the product does what the business wants. An agent can write correct code that does the wrong thing. A scenario catches it.

Who is responsible for what goes to production?

A human on your side always approves the merge. An agent never pushes code to the main branch on its own and has no access to production. Only the pipeline deploys to production, after a human approves.

Can training and the engagement be remote?

Yes, we run most of the work remotely. For workshops and the engagement kick-off we can come to you, in Poland or elsewhere in the EU.

Is our code safe?

We work under a data processing agreement (DPA). For regulated companies, models run on your own cloud account.

How much does the AI cost?

The budget is included in the package; any overage is billed transparently with an alert at 80%.

What if we stop working together?

The pipeline, configuration and documentation live in your repo, and your team is trained.

Which models do you use?

We match them to the task (Claude, OpenAI models, models on Bedrock or Vertex). We are not tied to one vendor.

Contact

Book a call

The first call takes 30 minutes. You tell us how you deliver today; we tell you where to start and what it costs.

kontakt@shipsteer.eu

Number of developers
Topic

We use form data only to contact you about this request. This site uses no tracking cookies.