What does Veles Digital build?

Custom AI tools, automation pipelines, conversational chatbots, and analytics dashboards. We engineer practical AI systems that eliminate waste, accelerate growth, and ship. Not demos. Every project is shaped around the specific workflows of a client's business.

How long does a custom AI project take?

A typical first-version delivery is two to eight weeks. Discovery and specification take about a week, a focused MVP runs two to four weeks, and larger integrations or analytics dashboards extend to six to eight weeks. We ship working software in short cycles rather than long-lead waterfalls.

Do you work with non-technical clients?

Yes. Most of our clients are not engineers. We translate business problems into AI solutions, handle the technical complexity end-to-end, and deliver tools your team can actually use. Onboarding includes training and clear documentation.

What industries does Veles Digital serve?

We're industry-agnostic. The common thread across our clients is teams who want concrete AI-driven outcomes like automation, better analytics, and customer support efficiency, rather than experimental AI projects. Each engagement is shaped around your specific workflows and operational reality.

How does a project start?

Book a free 30-minute discovery call via calendly.com/velesdigitalsolutions/30min, message us on WhatsApp at +381 69 251 1196, or email contact@veles.digital. We discuss your business, identify which workflows AI can improve, and propose a scoped first project with a fixed price and timeline. We respond within 24 hours.

Where is Veles Digital based and who do you serve?

Veles Digital is headquartered in Serbia and serves clients across Europe and worldwide. We operate fully remotely with English and Serbian language support, and most engagements happen via async tooling and scheduled video calls.

How much does custom AI development cost?

We don't publish fixed prices because every project differs in scope, integrations, and timeline. The discovery call is free, runs 30 minutes, and ends with a scoped quote based on your specific needs. Typical engagements range from focused two-to-four-week MVPs to six-to-eight-week multi-system integrations, with fixed scope and fixed price for the agreed phase.

What's the difference between Veles Digital and a ChatGPT subscription?

ChatGPT is a general-purpose assistant; Veles Digital builds production AI shaped around your specific workflows, integrated with your data, deployed on your own stack. ChatGPT helps individuals draft emails. We build systems that automate your support inbox, qualify your inbound leads, surface insights from your sales calls, or pull together a board-ready dashboard from your real-time data. Owned by you, not rented per seat from OpenAI.

Which AI platforms and models do you build on?

We're model-agnostic. We work with OpenAI (GPT-4o, GPT-4.1), Anthropic Claude (Sonnet, Opus, Haiku), Google Gemini, Mistral, and open-source models (Llama, Qwen) depending on the use case. Selection is driven by cost, latency, accuracy, and data-residency requirements, not vendor loyalty. We also use vector databases (Qdrant, pgvector, Pinecone) and orchestration frameworks (LangChain, custom agentic patterns) where they make sense.

Can you integrate AI with our existing CRM, ERP, or data warehouse?

Yes. Integration is the core of every project. We connect AI capabilities to Salesforce, HubSpot, Pipedrive, Notion, Airtable, Postgres, MySQL, BigQuery, Snowflake, and most modern SaaS APIs. If your stack has a documented API or webhook surface, we can wire AI into it. We don't ask you to migrate your data to a new platform.

Do you provide ongoing AI maintenance and updates?

Yes. After delivery, we offer optional retainers covering production monitoring, model upgrades, prompt tuning, and feature evolution. Most clients move to retainer after the initial build to keep their AI current as models and APIs evolve. What worked on GPT-4o today may be obsolete or expensive six months from now, and a small ongoing engagement keeps the system competitive.

How do you measure AI ROI for our business?

Every project starts with a measurable target like hours saved per week, response time reduction, conversion lift, support ticket deflection, or qualified leads per month. We instrument the AI system to report on those metrics from day one. If we can't measure the win, we don't ship it. Discovery includes a baseline measurement so the impact is provable, not anecdotal.

// AI Agents

AI agents that read every message. Humans handle what matters.

Conversational AI agents that take your inbound messages (WhatsApp, Instagram, email, or anything else customers actually use), draft responses, book what they can book, and route the hard ones to a human. Built in 2-8 weeks, owned by you.

What we build

Veles builds custom AI agents that live in the same channels your customers already use. The agent reads incoming messages, understands what the customer needs, and either answers directly, books something into your system, or hands the conversation off to a human with the context already summarized.

Every agent is bespoke. We do not ship a generic chatbot with your logo on it. We design the conversation patterns around how your team actually answers customers today, and we keep humans in the loop wherever judgement matters.

Typical agent deliverables:

Inbound triage: the AI reads every message, classifies intent (pricing question, support ticket, booking request, spam), and routes accordingly
Conversational answers for the questions you answer over and over (hours, pricing, availability, shipping, service area)
Booking and ordering flows: the agent can book a slot in your calendar, take an order, or generate a quote, all through chat
Human handoff with full context. When the agent escalates, the human sees a summarized conversation, not a wall of raw chat
CRM and back-office writes: the agent updates your CRM, scheduling tool, or order system as conversations close

Channels we support

WhatsApp Business API. The big one for Serbia, the Balkans, and most of Europe outside the Nordics. We use the official WhatsApp Business Cloud API (not third-party scrapers or unofficial wrappers), which is the only viable path for scale. The agent reads inbound messages in real time, drafts replies, sends template messages where required, and respects the 24-hour customer service window. Best fit for B2B ordering, support, scheduling, and lead qualification.

Instagram DMs. Through the Instagram Graph API, the agent reads incoming direct messages on your business account, answers price and availability questions, books appointments, and escalates the hard ones to the owner. Best fit for local services, creators, and small retail brands buried under DM volume.

Email. The agent monitors a shared inbox (Gmail, Outlook, IMAP, anything), classifies each thread, drafts replies for human review or auto-sends low-risk responses. Best fit for support queues, sales inboxes, and partner coordination.

Your own channels. SMS, in-app chat, Slack, Discord, Microsoft Teams, contact-form submissions, voicemail transcripts. If it has an API and your customers use it, we can build an agent on top.

You do not have to migrate channels to use this. The agent shows up where your customers already are.

How it works

1. Discovery. A 30-minute call to understand the channels, the volume, the kinds of messages, and the handoff points. We listen for the questions you answer 50 times a week and the conversations that genuinely need a human.

2. Scope and estimate. We map the agent into milestones: which intents it handles, which channels it covers, which systems it writes back to, and where humans stay in the loop. Concrete fixed-price estimate with no surprise invoices. See our scoping methodology.

3. Build. A small senior team works in weekly sprints. You get a live demo each week. Before the agent touches a real customer we run dry-runs against a copy of recent message data so you can see exactly how it would have responded.

4. Launch and operate. We deploy and monitor closely for two weeks, tuning the agent's tone, thresholds, and escalation rules based on real conversations. After that, hand off the code or keep us on a monthly operate-and-iterate retainer.

Why custom over off-the-shelf chatbots

Generic chatbot platforms (Intercom, ManyChat, Drift, Tidio) are good at one thing: scripted decision trees. They fall apart when:

Customer messages are unstructured, multi-intent, or in mixed languages (which is most of them)
You need the bot to actually book, order, or update a system, not just suggest "talk to a human"
The handoff to a human matters and you need a real summary, not a transcript dump
WhatsApp Business API at production scale is required (most off-the-shelf bots default to the unofficial Web API)
You want to own the agent code and the conversation data, not rent them by the seat

Custom AI agents from Veles handle all of the above by default. The agent is yours, the conversation history is yours, and the system gets cheaper to operate over time as we tune it on your real data.

Who we work with

Three kinds of businesses get the most leverage from custom AI agents:

Small business owners drowning in messages. The owner of a 5-person business who personally answers WhatsApp, Instagram DMs, and email every day. The agent reclaims that time without losing the personal touch the customers came for.
B2B operators with high-volume repeat conversations. Wholesalers, distributors, parts dealers, agencies. The 80% of messages are the same 20 questions; the agent handles them, the owner approves the daily promo or escalation.
SaaS teams building AI features inside their own product. When you need an agent shipped into your product without hiring a full ML team, we build the AI layer and integrate with your existing stack.

See the proof: Mobilni Market (B2B WhatsApp ordering), Kacimbas (local Instagram bookings), People Up (recruitment screening agent).

// Proof from production

Already running in production

AI Commerce Stack

Mobilni Market

B2B mobile-parts wholesaler. The AI agent reads incoming WhatsApp orders, checks stock, negotiates within owner-approved limits, and writes the order to the back-office system. The owner approves the daily promo, the rest runs itself.

Case study

AI Booking System

Kacimbas Autostyling

Local booking business buried in Instagram DMs. The AI agent reads messages, quotes price and availability, books the appointment into the calendar. The owner sees only what is already booked.

Case study

Recruitment OS

People Up

Recruitment platform with an AI agent that screens candidates, scores resumes, and sequences screening questions before any human gets involved. Humans interview and decide. Multi-day pipelines now run in hours.

Case study

// FAQ

Frequently asked

What exactly is an AI agent and how is it different from a chatbot?

A chatbot is usually a scripted decision tree: "Press 1 for pricing, press 2 for support." An AI agent is a system that reads unstructured messages, understands intent, takes action (book, order, update CRM, draft a response), and decides when a human needs to step in. Chatbots are good for FAQ pages. Agents are good for actual customer-facing work.

Will the AI sound like our brand or like a generic bot?

Like your brand. Part of Discovery is reviewing how your team actually answers customers today, so the agent's tone, vocabulary, and pacing match yours. Customers should not be able to tell which messages were AI-drafted and which were human, and the AI never pretends to be a person who has a name and feelings (which is both a UX best practice and required by EU AI Act transparency rules).

How long does it take to ship a custom AI agent?

Most agents we ship are 2-8 weeks end-to-end. A focused single-channel agent (one input channel, light CRM integration, no custom booking flow) lands in 2-3 weeks. A multi-channel agent with custom integrations and on-call escalation runs 6-8 weeks. We give you a concrete timeline at the end of Discovery, before any code is written.

Can the agent actually take orders or just answer questions?

Yes, it can take orders, book appointments, send quotes, update your CRM, and write to your order system. The Mobilni Market case study is exactly this: AI proposes the daily promo, the owner approves once, and the agent then negotiates and takes WhatsApp orders. One approval per day, the rest runs itself.

What happens when the AI does not know the answer?

It escalates. The handoff is the most important design decision in any custom AI agent. We design explicit escalation triggers (new customer, high-value inquiry, complaint signals, anything the agent is not confident about) and the human sees a summarized conversation with the customer's question, what the agent already said, and what action is recommended. No "the AI has gone silent" experiences.

Will this work for our small team, or do we need an IT department?

Small teams are exactly who this works best for. The teams who benefit most from custom AI agents are the ones where the owner or operator is the bottleneck answering messages. You do not need an IT department to operate one. We hand off documentation and runbooks, and most clients only need to touch the system when they want to change the agent's behavior or add a new intent. We are also available on a monthly retainer if you want us to keep the agent tuned for you.

How do you handle customer data and privacy?

We are GDPR-compliant out of the box (EU-based, Belgrade). Before any production data moves we sign a Data Processing Addendum. For sensitive workloads we can run inference on-premise or via privacy-preserving providers, so your customers' conversations never train anyone else's model. Specifics get nailed down with you during Scope and Estimate.

Can the agent speak Serbian, English, or both?

Both, plus most other languages. The underlying language models we use handle Serbian, English, German, Spanish, Russian, Arabic, and dozens of others natively. For a Serbian B2B agent we usually configure it to respond in the customer's language by default (Serbian Latin / Cyrillic / English) and we tune the brand voice separately for each. Multi-language is not a premium feature; it is included.

How we estimate

No tiered pricing because every custom AI agent build varies. After Discovery we map the agent into milestones (channels, intents, integrations, escalation rules), identify the hard parts, and give you a concrete estimate with no surprises. The methodology is the same whether you need a single-channel WhatsApp agent or a multi-channel system with on-premise inference.

See our scoping methodology →

Talk to us about your project

Bring the channel, the volume, and what good looks like. 30 minutes is enough to know if a custom AI agent is the right fit, or if no-code would save you money. We will tell you straight.

book a discovery call Or email us