Platform · RPA Bots
RPA bots for everything without an API
SAP GUI, AS/400, Citrix, QuickBooks Desktop, carrier and government portals, Excel and Outlook. Deterministic bots that run unattended on free Bot Runners, orchestrated from the cloud, and called as tools by agents. The cost is the same whether the job is 10 records or 10,000.
Updated for the October 2026 release.
Orchestrator · production fleet
invoice-agent
unattended · VM-04
as400-order-sync
unattended · VM-02
po-acknowledgments
unattended · VM-04
carrier-packet-intake
attended · ops-desk-3
month-end-recon
unattended · VM-01
In one paragraph
ElectroNeek RPA bots are built in Studio (a Windows designer with a recorder, a debugger and a library of activities for web, desktop, Citrix/RDP, terminals, Excel, email, PDF, databases and HTTP) and executed by Bot Runner, a free Windows application you can install on any number of PCs, servers or VMs. The cloud Orchestrator schedules runs, queues them when runners are busy, exposes webhooks and a REST API, keeps console logs and run history, and alerts by email on failure. Bots run unattended even when the RDP session is closed or minimized, and an AI workflow can run a bot as one of its steps. Pricing is per bot run; there is no fee per robot.
What it does — concretely
Systems without an API
Web apps through Chrome and Edge, Windows desktop apps, Java apps, Citrix and RDP through image-based surface automation, AS/400 and mainframe terminals, SQL databases, Excel and Outlook.
Unattended on a closed RDP session
Bots keep running on locked, closed or minimized remote sessions (Closed RDP add-on) — the failure mode that fills the Power Automate forums.
Free, unlimited Bot Runners
Install as many runners as you need on PCs, servers or VMs. You pay for runs, not for machines. White-label runner available for partners.
Orchestrated from the cloud — or on-prem
Schedules, launch queues, webhook links, a REST API, Zapier and AI-workflow triggers; run history with console logs; failure emails. Multi-tenant orchestrators for agencies. The whole thing on-prem with Sovereign Suite.
Called as a tool by an agent
Run RPA Workflow and Wait for RPA Workflow are steps in AI Workflows: the agent decides what to enter, the bot enters it, the result flows back.
Documents and AI inside a bot
Extract Document Data (IntelliDocs), Ask AI and Edit Text activities, cloud OCR engines and third-party extraction services, plus JavaScript, Python and command-line steps when you need code.
How it works
01
Build in Studio
Record desktop and browser steps, drag activities onto the canvas, add error branches, debug step by step. AI Code Assistance for scripts; the workflow generator is in beta.
02
Publish to the Orchestrator
Assign the workflow to Bot Runners, set a schedule or a webhook, store credentials in the Windows vault on the runner. Versions are kept with rollback.
03
Run and watch
Queued launches, console logs with millisecond timestamps, run history, output variables, email alerts on failure — and the same bot callable from an agent.
What it doesn’t do (yet)
Stated here so you don’t find out in a pilot. Roadmap items are marked as such.
- —Studio and Bot Runner are Windows-only (Windows 10/11, Server 2019/2022). No macOS or Linux runners, no containerized runners.
- —One workflow at a time per Bot Runner; parallel runs need more runners or Windows sessions.
- —Selectors are deterministic: when a UI changes, the bot needs an updated selector. AI self-healing and anchor-based selectors are in development, not shipped.
- —No native SAP GUI scripting or Citrix connector yet — SAP and Citrix are automated through the UI and image-based surface automation. Native SAP activities are in the roadmap.
- —Browsers: Chrome, Edge and Internet Explorer; no Firefox or Safari. Machine-vision activities work on the primary monitor.
- —No Git integration; versioning is in the Orchestrator and a local file history in Studio.
Common questions
How do I automate a system that has no API?
Build a bot in Studio that operates the application on screen — clicks, keystrokes, reading fields — deterministically, with a step log. Run it on a free Bot Runner, unattended, triggered by a schedule, a webhook, the API or an AI workflow. SAP GUI, AS/400, Citrix, desktop ERPs and web portals are the usual targets.
Why a bot and not an LLM that uses the computer?
For short, one-off screen tasks an agent is the right tool. For repetitive, high-volume, auditable work it is not: on OSWorld 2.0 (2026) the best model completes about one long task in five at roughly $25–70 per task, every screenshot costs tokens and every run differs. A bot runs job 10,000 at the price of job 1 and leaves a log you can replay. ElectroNeek uses agents for the decisions and bots for the execution.
Is RPA dead?
RPA-only platforms are. Gartner still rates RPA as the reliable option for legacy systems without APIs, Forrester expects most organizations to keep running deterministic automation through 2026, and UiPath and Automation Anywhere both re-architected their robots as the execution arm of agents. Bots are the hands; what changed is who decides.
Do we pay per bot or per run?
Per run. Bot Runners are free and unlimited; install one on every machine you need. Power Automate charges $150 per unattended bot per month ($215 hosted); UiPath and Automation Anywhere license per robot.
What happens when a bot fails mid-run?
The run is marked failed with the console log and the failing step; failure emails go to the owners; queued launches continue; you can relaunch from the Orchestrator or the API. Workflow-level retries and continue-on-failure exist in AI Workflows; replay-from-step for bots is on the roadmap.
Why does my bot break after a Chrome update, and what do you do about it?
Browser updates change page structure and extension behavior; every RPA vendor is exposed. Today: Chrome Native and Edge Native drivers with the ElectroNeek extension, updated with each release. In development: nightly testing against Chrome and Edge beta builds and anchor-based selectors so one UI change doesn't break many bots.