Book a call

Services

AI and software development services

Central Dev designs, builds and integrates AI agents, automations and chatbots, together with the websites, mobile apps and cloud systems they run on. Nine services, delivered by one team of 8 engineers in Chișinău, Moldova.

From the introductory call to production, you work directly with the engineering team responsible for your product, in English.

What is your goal?
AI SYSTEMS · 4PRODUCT ENGINEERING · 5AgentsAutomationsChatbotsDeep agentsWebMobileCloudDesignMVP
Left ring: AI systems, 4 services. Right ring: product engineering, 5 services. Highlighted nodes mark the recommended starting services for the selected goal.

AI features inside your product

Chatbots and agents built into your web or mobile product and integrated with your backend, with human approval on sensitive actions.

01— AI systems · 1 of 9

AI solutions & agents

Hand multi-step work to AI agents that operate inside your systems. We design agents that interpret a request, retrieve the relevant data, take the action and escalate to a person when a decision needs human judgment.

PLANACTCHECKREPEATAgentCRMEmailDocsYou approve sensitive actions
FIG. 01  Plan, act, check and repeat; sensitive actions wait for human approval.
Explore AI agent development

Who it is for

  • Operations teams in support, claims, sales and finance that move data between tools by hand
  • Product teams that want an in-product assistant to complete tasks for their users

What we build

  • Agents that triage and resolve incoming requests
  • Back-office agents that read documents and update your CRM
  • In-product agents that take actions on behalf of your users
  • Internal assistants that work across your company data

Client work

AdjustedDelivered

For Adjusted, we built an AI agent that is fully integrated with the backend and takes actions on the user's behalf.

Read the Adjusted case study

How it starts

It starts with an introductory call about the workflow you want to hand over. Scope, estimate and timeline are agreed with you afterwards.

02— AI systems · 2 of 9

AI automations

Remove repetitive manual steps from your workflows, from inbox and documents to your business systems. Automations follow fixed rules, with a language model only in the steps where rules are not sufficient.

AutomatedNeeds a personYOUR PROCESSInboxExtractValidateEnterNotifyExceptions to a personRules first; a model only where rules break down
FIG. 02  Routine steps run automatically; a person reviews only the exceptions.
Explore AI automation

Who it is for

  • Teams running a process by hand across email, spreadsheets and SaaS tools
  • Companies whose no-code automations no longer cover their process

What we build

  • Document intake for invoices, claims and applications, with data extracted into your systems
  • Routing of emails and incoming requests
  • Scheduled reports and summaries
  • Data synchronization between your CRM and other tools

How it starts

It starts with an introductory call about one process you want to automate. We identify the steps worth automating, then agree scope, estimate and timeline with you.

03— AI systems · 3 of 9

Chatbot development

Give customers and employees accurate answers from your own documents and data. We build chatbots that cite their sources, hand over to a person when uncertain and can later be extended into agents that take actions.

1CHATCan I return an order?Yes, for unused itemswith the receipt.1Source 1: Returns policy, p. 2Uncertain? A person takes over
FIG. 03  Illustration: answers come from your documents with the source cited; uncertain questions go to a person.
Explore chatbot development

Who it is for

  • Support teams that answer the same questions every day
  • Companies with extensive knowledge bases, policies or product documentation
  • Internal help desks in HR, IT and operations

What we build

  • Customer support chat for your website or application
  • Internal knowledge assistants for your team
  • Onboarding and pre-sales assistants

Client work

AdjustedDelivered

Adjusted's chatbot answered basic questions. We rebuilt it as a deep agent that understands the system and takes actions for the user.

Read the Adjusted case study

How it starts

It starts with an introductory call and a sample of real user questions, together with the documents that answer them. Scope, estimate and timeline are agreed with you afterwards.

04— AI systems · 4 of 9

Deep-agent development

Let your users complete work inside your product by asking for it. A deep agent understands your system, is integrated with your backend and takes actions on the user's behalf.

UserDeep agentUnderstands the systemAPIYOUR BACKENDDataRecordsActionsIntegrated with your backend; acts for the user
FIG. 04  The agent sits between your users and your backend and acts through an integration layer.
Explore deep agent development

Who it is for

  • Products with a chatbot that answers questions but cannot act in the system
  • Teams that want users to complete tasks inside the product by asking for them

What we build

  • Agents that understand your data model and each user's context
  • Integrations with your backend, so the agent can read data and take actions
  • Multi-step tasks that the agent plans and carries out for the user

Client work

AdjustedDelivered

We rebuilt Adjusted's chatbot as a deep agent that understands the system, is fully integrated with the backend and takes actions on the user's behalf.

Read the Adjusted case study

How it starts

It starts with an introductory call about your product and the tasks your users should be able to delegate. Scope, estimate and timeline are agreed with you afterwards.

05— Product engineering · 5 of 9

Websites & web apps

Websites and web applications built around how your team and your customers work. We build products such as CRMs and dashboards, and marketing websites your team can update itself.

Web appsWebsitesDashboardsSites, web apps, CRMs and dashboards
FIG. 05  Web applications for your product, websites your team can edit, dashboards for your data.
Explore website and web app development

Who it is for

  • Companies relaunching a website that no longer meets their needs
  • Teams that need a CRM or internal tool tailored to their workflow
  • Startups that need a web product and the website around it

What we build

  • Marketing websites your team can edit
  • Web applications and dashboards
  • CRMs and customer portals
  • Digital menus and other customer-facing tools

Client work

AdjustedDelivered

We delivered the web application and the CRM, alongside the AI agent.

Entera ToolsIn progress

We are building the CRM as part of the MVP. The project is in progress.

Tucano Coffee MoldovaDelivered

A digital menu that gives customers more information about what is on the menu.

View all case studies

How it starts

It starts with an introductory call about your current website or a short brief. Scope, estimate and timeline are agreed with you afterwards.

06— Product engineering · 6 of 9

Mobile apps

Mobile applications that work as one product with your backend and web application. We build Android and iOS apps with the same team that delivers the systems they connect to.

iOSAndroidONE TEAM · ONE BACKEND
FIG. 06  Both apps connect to one backend built by the same team, so features stay consistent across platforms.
Explore mobile app development

Who it is for

  • Startups launching a mobile-first product
  • Companies adding a mobile app to an existing web product

What we build

  • Consumer applications
  • Companion apps for web products and CRMs
  • Apps with AI features, such as an in-app assistant

Client work

AdjustedDelivered

We delivered the mobile application, built by the same team as the web application, CRM and AI agent.

Read the Adjusted case study

How it starts

It starts with an introductory call about the key screens and the systems the app connects to. Scope, estimate and timeline are agreed with you afterwards.

07— Product engineering · 7 of 9

Cloud architecture

Cloud infrastructure sized to the workload your product actually runs. We design and build cloud architecture for web products and AI workloads.

EDGEAPPDATACDN · firewallZone AZone BreplicationprimaryreplicaIllustration: keeps running if one zone fails
FIG. 07  An example cloud layout: two zones behind one edge layer, with the database replicated across them.
Explore cloud architecture

Who it is for

  • Startups moving from a single server to a production setup
  • Teams with rising cloud costs and no clear explanation for them
  • Companies that want AI workloads inside their own cloud account

What we build

  • Cloud architecture for new products
  • Migrations from other hosting providers
  • Infrastructure for AI workloads
  • Cost and reliability reviews

How it starts

It starts with an introductory call about your current setup and the issues you want to resolve. Scope, estimate and timeline are agreed with you afterwards.

08— Product engineering · 8 of 9

Product & UI design

Interfaces your users understand and your engineers can build from. We turn workflows into user flows, screens and a design system that keeps design and code consistent.

180 × 128Tokens → components
FIG. 08  Screens are built from tokens and components, so design and code stay in step.
Explore product and UI design

Who it is for

  • Founders preparing a first build
  • Teams whose product has grown one feature at a time
  • AI products that need clear review, confirm and undo steps for model output

What we build

  • User flows and information architecture
  • Interface design for web and mobile
  • Design systems
  • User experience for AI features

How it starts

It starts with an introductory call about your main user journey. Scope, estimate and timeline are agreed with you afterwards.

09— Product engineering · 9 of 9

MVP development

Take a defined idea to a launched first version of your product. We deliver MVPs at a fixed price when the scope is clear, or on an hourly basis while it is still evolving.

From scope to launch010203ScopeCore flowFirst versionLiveEach stage ends in a working build
FIG. 09  Scope, core flow, first version, then launch.
Explore MVP development

Who it is for

  • Founders with a validated problem
  • Companies testing a new product line
  • Non-technical founders who need a technical partner

What we build

  • SaaS and CRM MVPs
  • AI-first products
  • Mobile-first MVPs

Client work

Entera ToolsIn progress

We are building the CRM and MVP for Entera Tools. The project is in progress.

Read about the Entera Tools project

How it starts

It starts with an introductory call about the idea and its target users. Scope, estimate and timeline are agreed with you afterwards.

How we work

How an engagement starts

Share the context of your project on an introductory call. Scope, estimate and timeline are agreed with you afterwards, and you work directly with the engineering team responsible for your product.

  1. 01

    Introductory call

    A 30-minute conversation about your goals, constraints and existing systems.

  2. 02

    Scope and estimate

    We talk through the project and agree the scope, the estimate and the services that fit.

  3. 03

    Agreement and NDA

    We sign an NDA and agree the engagement model: dedicated team, fixed price or hourly.

  4. 04

    Kickoff

    Work starts on the date we agree together.

  • Working hours

    09:00–19:00 Moldova time (EET/EEST). Extended hours are possible, agreed per project.

Explore how we work

The filled ring marks kickoff. The same process applies to AI and non-AI projects.

Engagement models

Three engagement models

Every engagement is scoped individually. Choose the model that fits how defined your scope is; the estimate and timeline are agreed with you after the first conversation.

Three engagement models compared: dedicated team, fixed price and hourly (time and materials). Each project receives an individual scope and estimate.
CompareOngoing product workDedicated teamDefined scopeFixed priceEvolving scopeHourly
What it isA team of Central Dev engineers that works exclusively on your product, using your tools and processes.A defined scope, an agreed price and a clear list of deliverables, all set before work begins.You are billed for the hours worked, and priorities can change as the product evolves.
Best forCompanies building a core product without a full in-house team.MVPs, AI pilots and other well-defined builds.Evolving products, audits and additional capacity for your team.
You getA stable team that builds deep knowledge of your product.Budget certainty from the start.Flexibility to re-prioritize at any point.
Learn moreDedicated team in detailFixed price in detailHourly in detail

FAQ

Frequently asked questions

Answers to common questions about AI projects and working with us. Anything else can be discussed on a call.

Book a call

AI projects

How much does an AI agent cost?

We do not publish prices, because the cost of an AI agent depends on its scope. The main factors are how many systems the agent connects to, how reliable it must be and how much it will run. Scope and estimate are agreed with you after the first conversation.

What is a deep agent?

A deep agent is an AI agent that completes multi-step work inside a product, beyond answering questions. It understands the system, is integrated with the backend and takes actions on the user's behalf. For Adjusted, we rebuilt a question-answering chatbot as a deep agent.

What is the difference between an AI agent and an AI automation?

An automation follows steps you define, while an agent decides which steps to take. Automations are predictable and inexpensive to run; agents handle varied inputs that fixed rules do not cover. In practice, a system often combines both, with a language model only in the steps where fixed rules are not sufficient.

Do we need a chatbot or an AI agent?

Choose a chatbot if your users need answers, and an agent if they need tasks completed. A chatbot answers questions from your documents, while an agent takes actions in your systems, such as updating a record, and can ask a person to confirm sensitive steps. A chatbot can later be extended into an agent, as we did for Adjusted.

Working with us

Which service do I need?

The right service depends on the problem you want to solve. Select a goal in the service chooser at the top of this page to see the recommended starting services, or describe your project on a call and we will recommend the services that fit.

When can a project start?

The start date is agreed with you together with the scope and estimate, after the first conversation. It depends on the engagement model and the team your project needs.

Do you sign an NDA?

Yes. We sign an NDA and talk through how your data is handled in the first conversation.

Can you work with our existing team?

Yes. As a dedicated team or on an hourly basis, our engineers work in your tools and alongside your team. Our working hours are 09:00–19:00 Moldova time (EET/EEST), which covers the full working day in Central Europe, the UK and Israel. Extended hours are possible, agreed per project.

Book a call

Start a project

Let's discuss your project

Book a 30-minute call with our team or write to us. Scope, estimate and timeline are agreed with you after the first conversation.

Book a call

Choose a 30-minute slot on Cal.com. Times are shown in your own time zone.

  1. 01Choose a time slot and add a few lines about your project.
  2. 02Discuss your goals, constraints and the right engagement model with our team.
  3. 03Agree scope, estimate and timeline together.

Prefer email?

Describe your project in a few lines, and we will continue the conversation by email.

Hours
09:00–19:00 Moldova time