# Bitsclan IT Solutions — Agency that designs, builds and runs custom AI agents inside your CRM, helpdesk, ERP and phone lines.

> Source: The Agents Index — https://theagentsindex.com/bitsclan-it-soulution (structured, researched, re-verified)
> Facts last verified: 2026-09-02

Bitsclan IT Solutions is a US-registered software agency that designs, builds and maintains custom AI agents for support, sales, operations and voice workflows. Engagements run from workflow discovery through integration and launch to ongoing tuning, delivered by its own engineers.

| Fact | Value |
| --- | --- |
| Website | https://bitsclan.com/ai-agent-development-services |
| Price | from $1500/one time |
| Free tier | Partly |

Free discovery/consultation call and a public ROI calculator that needs no email ("no cost, no obligation"; "no email required"); no free or trial build of an agent.

## What it does

**Writes code** · Yes — People, not an agent: Bitsclan's own engineers write each agent against the workflow mapped in discovery, in a 2-4 week build stage.

> Our engineers build and test the agent against real scenarios from your business, not generic demos.
> — https://www.bitsclan.com/ai-agent-development-services · How We Work, step 03 "Build"

**Debugs and fixes** · Partly — Post-launch maintenance is contracted - models tuned, results monitored, features updated - but nothing describes diagnosing and fixing faults in a client-owned codebase.

> Our experts fine-tune models, monitor results, & update features regularly so your business always benefits from the latest AI agent development company innovations.
> — https://www.bitsclan.com/ai-agent-development-services · "AI Agent Development Services" list, "Continuous Improvement & Maintenance"

**Writes tests** · Partly — The build stage tests the agent against the client’s real scenarios; no test suite or test artefact is named as a deliverable.

> Our engineers build and test the agent against real scenarios from your business, not generic demos.
> — https://www.bitsclan.com/ai-agent-development-services · How We Work, step 03 "Build"

**Ships it** · Yes — The engagement runs past the build to launch: usage is monitored, the agent tuned in production, and a performance dashboard handed over.

> We monitor real usage, tune the agent's accuracy and tone, and hand you a dashboard to track performance long-term.
> — https://www.bitsclan.com/ai-agent-development-services · How We Work, step 05 "Launch & Optimize"

**Done-for-you service** · Yes — This is an agency engagement, not software you run: Bitsclan can own strategy through launch, or embed engineers in your team.

> Let us handle the whole project, from strategy & planning through launch. You set the vision & we will do the work, with full transparency & control.
> — https://www.bitsclan.com/ai-agent-development-services · Engagement Models, "AI Agent Project Outsourcing"

_Not established: Reviews code_

## How much it does unattended

**Runs autonomously** · Yes — The vendor draws its own line against chatbots on exactly this axis, and its voice build books appointments in-call with no human handoff.

> An AI agent reasons over your data, calls your tools and APIs, and completes the task without a human closing the loop.
> — https://www.bitsclan.com/ai-agent-development-services · Section "Chatbot vs. AI Agent", intro paragraph

**Multi-agent** · Yes — Multi-agent systems are named as an integration target, and the build stack cites AutoGen Studio and CrewAI.

> enabling smooth communication using advanced microservices, containerization, & API strategies for single-and multi-agent systems
> — https://www.bitsclan.com/ai-agent-development-services · "AI Agent Integration" service description

**Agent permissions** · Partly — Limits are set once, at design time: decision logic and escalation points are architected and signed off before code. No runtime permission model is published.

> We architect the agent's decision logic, integrations, and escalation points, and walk you through it before writing code.
> — https://www.bitsclan.com/ai-agent-development-services · How We Work, step 02 "Design"

## Which models it runs on

**Claude** · Yes — Claude is one of five technologies listed in the agent stack’s NLP tab, alongside Llama 3 and DeepSeek.

> Claude
> — https://www.bitsclan.com/ai-agent-development-services · "Tools & Stack Technologies" grid, "Natural Language Processing" tab, tile 3 of 5 (spaCy, Hugging Face Transformers, Claude, Llama 3, DeepSeek)

**GPT** · Yes — OpenAI models are named in the healthcare service description; the homepage stack grid also lists GPT-4, GPT-3 and OpenAI GPT.

> Our healthcare AI agent services include voice agents & ChatGPT-driven assistants.
> — https://www.bitsclan.com/ai-agent-development-services · Industries We Serve, "Healthcare" description

**Open models** · Yes — A delivered engagement ran open-weight models locally via Ollama; Llama 3 and DeepSeek also appear in the agent stack’s NLP tab.

> By combining n8n & Ollama, Bitsclan developed a powerful, scalable workflow that enables seamless interaction with locally hosted Large Language Models (LLMs).
> — https://www.bitsclan.com/engagment-model · Success Stories carousel, "How Self Hosted LLMs Delivered 100% Data Control and 40%..." card

**Model choice** · Partly — Bitsclan picks the model as a build step; the stack spans Claude, GPT, Llama 3 and DeepSeek, but no buyer-facing model switch is published.

> Selecting the right AI model
> — https://www.bitsclan.com/ai-agent-development-services · FAQ "What is the process of building an AI agent?", step 3 of 6

## Where you use it

**In a browser** · Partly — A performance dashboard is handed over at launch; the agent’s own surfaces are whatever the engagement builds - phone, chat or an internal tool - not a hosted product.

> hand you a dashboard to track performance long-term
> — https://www.bitsclan.com/ai-agent-development-services · How We Work, step 05 "Launch & Optimize"

## Whose machine it runs on

**Self-hosted** · Partly — One published engagement runs the models on the client’s own machines; self-hosting is a demonstrated build option, not a stated default of the agent service.

> enables seamless interaction with locally hosted Large Language Models (LLMs). This innovative platform ensures complete control over AI
> — https://www.bitsclan.com/engagment-model · Success Stories carousel, "How Self Hosted LLMs Delivered 100% Data Control and 40%..." card

_Not established: Open source_

## What it costs to run

**How it meters** · Yes — Billed per engagement, not per seat or token: fixed price, time & material, or a dedicated team. The published automation audits are one-time fees.

> At Bitsclan, the most commonly used engagement models include Fixed Price, Time & Material (T&M), and Dedicated Team models.
> — https://www.bitsclan.com/engagment-model · FAQ "Which engagement models does Bitsclan offer?"

**Free tier** · Partly — A public ROI calculator needs no email, and the consultation is free - but every build starts from a call; there is no free or trial agent.

> Not ready for a full call yet? Get an instant estimate first — no email required.
> — https://www.bitsclan.com/ai-agent-development-services · Section "FREE TOOL - See What Manual Work Is Actually Costing You"

**API access** · Partly — The agent is wired into your stack by secure API per engagement, with auth and logging; Bitsclan publishes no API of its own to drive.

> Secure API connections to your CRM, helpdesk, ERP, or internal tools, with authentication and logging in place.
> — https://www.bitsclan.com/ai-agent-development-services · How We Work, step 04 "Integrate"

**Bring your own key** · Partly — A delivered build ran the client’s own local models through Ollama, so own-infrastructure inference is possible; no key or cloud BYO policy is published for a standard engagement.

> By combining n8n & Ollama, Bitsclan developed a powerful, scalable workflow that enables seamless interaction with locally hosted Large Language Models (LLMs).
> — https://www.bitsclan.com/engagment-model · Success Stories carousel, "How Self Hosted LLMs Delivered 100% Data Control and 40%..." card

_Not established: MCP server_

## Buying it for a team

**A company can buy it** · Yes — Sold to organisations only, in three shapes: staff augmentation, a dedicated team, or fixed-scope end-to-end delivery.

> We offer three models: staff augmentation when you need to fill a specific skill gap quickly, dedicated teams for long-term product work, and end-to-end delivery
> — https://www.bitsclan.com/ · FAQ "How do we engage can we just add engineers to our team?"

**Seat model** · Partly — The closest thing to a seat is an engineer: staff augmentation adds named engineers to your team, with no rate, floor or ceiling published.

> Need a couple of extra AI agent engineers on your team? Get senior ai voice agent developers who integrate directly into your internal team.
> — https://www.bitsclan.com/ai-agent-development-services · Engagement Models section, Staff Augmentation card

**Admin controls** · Partly — Constraints are agreed with you at design time and built in; no admin console for changing what the agent may do is described.

> We architect the agent's decision logic, integrations, and escalation points, and walk you through it before writing code.
> — https://www.bitsclan.com/ai-agent-development-services · How We Work section, step 02 Design

**Audit log** · Partly — Delivered agents log their activity into your systems - the voice build logged every call to the CRM - but no administrative audit trail is described.

> Every call logged and synced to the CRM
> — https://www.bitsclan.com/case-studies/voice-ai-never-miss-a-call · Section "What We Built", item 04

_Not established: Single sign-on_

## What happens to your code

**Opt out of training** · Partly — Personal information is contractually confined to the purpose it was collected for, which excludes reuse for training; client project and agent data are not addressed at all.

> We use personal information only for the primary purpose for which it was collected, or for a secondary purpose that is related to the primary purpose and which you would reasonably expect.
> — https://www.bitsclan.com/privacy-policy · Section 6.1 "Use" (APPs 6, 7 & 9)

**Data residency** · Partly — Data sits on Google Cloud in the USA and Australia and the company operates from Florida; no per-client region choice is offered, and the Terms’ governing-law clause is still an unfilled "[jurisdiction]" placeholder.

> We process and store personal and health information using Google Cloud, leveraging its certified Australian and global data centres.
> — https://www.bitsclan.com/privacy-policy · Section 7 "Cross-Border Disclosure of Personal Information (APP 8)"

**Getting out** · Partly — Engagements carry a full money-back refund, and personal data is accessible on request within about 30 days; no export path for a delivered agent’s data or code is published.

> We offer a 100% money back guarantee with a no questions asked refund policy, ensuring complete confidence & risk free engagement.
> — https://www.bitsclan.com/engagment-model · Engagement Model Comparison table, note beneath the table

**Certifications** · Partly — HIPAA and the Australian Privacy Act/APPs are claimed. SOC 2 is never mentioned, and ISO 27001/27018 appear only as certifications Bitsclan verifies in its cloud providers - not ones it holds.

> Health Insurance Portability and Accountability Act (HIPAA).
> — https://www.bitsclan.com/privacy-policy · Section 3 "Applicable Laws and Legal Frameworks", United States list

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

None recorded. Every fact above comes from the vendor and is not independently corroborated yet.

## What changed

- 2026-09-02 — Capabilities · Commercial terms · Free tier · Price

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent Bitsclan IT Solutions, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.