Back to vendors
W

WIQ

Also known as: getwiq.ai, WIQ by Bardeen, WIQ agent blueprints, WIQ Agent

Visit site
Entry priceQuote only, through a demo; the Try page turns one recording into a free Claude skillFull pricing detail

WIQ, from the team behind Bardeen, watches how operations teams actually work through a scoped browser extension, ranks which processes are worth automating, and turns each one into a blueprint of steps, tools, approval points and escalations. Approved blueprints deploy as agents on WIQ's own agent or on Claude, Microsoft 365 Copilot, OpenAI Codex, Workato or any MCP platform, running locally or in the customer's cloud, with every run checked against its blueprint and logged.

WIQ is an enterprise agent platform from Bardeen, Inc., the company behind the Bardeen browser automation tool, run by Bardeen co founder Artem Harutyunyan and sold from its own site, getwiq.ai. It aims at the question most agent builders skip, which processes should get agents at all, and sells into customer support operations, payment operations, KYC and onboarding, procurement, fraud and financial crime, order to cash, logistics and commodities.

Discovery starts with a browser extension that records sanctioned work, limited by allowlists and blocklists for apps, domains and time windows so that nothing outside the policy is captured. WIQ combines those recordings with systems of record, documentation and expert answers to reconstruct how each process runs, showing the hours it takes each week and the systems it touches, and ranks workflows by automation value and readiness. Each chosen process becomes a blueprint that sets out the steps, the tools required and the approval points, which the team reviews before anything deploys.

WIQ's pipeline then builds an agent for each discrete task and deploys it to WIQ's own agent or to Claude, Microsoft 365 Copilot, OpenAI Codex, Workato or any MCP compatible platform, running locally where the work happens or inside the customer's cloud. Example blueprints work across ServiceNow, SAP Ariba, SAP S/4HANA and Outlook, and agents can run concurrent sessions to clear a ticket backlog in parallel. When a process changes or an agent meets a new exception, WIQ surfaces the change and the team approves the updated blueprint.

A control plane lists every agent run with its operator, platform and duration, scopes the tools and actions each agent may use, and holds the approval gates, escalation paths and checkpoints where agents pause for a person. Every run is checked against its approved blueprint, deviations go to the right person, and SLA thresholds are set per session. WIQ carries SOC 2 Type II, GDPR and CCPA badges. Pricing goes through a demo, and a free Try page turns one recording into a downloadable Claude skill without WIQ's guardrails.

WIQ fits operations leaders at large companies who want automation chosen from measured work rather than workshops and who already run an agent platform. Its case studies report an agribusiness freeing $5.2 million of working capital with 80% fewer handoffs, a medical equipment maker targeting 20,000 hours a year across more than 150,000 receiving orders, and a logistics company cutting exception handling time by 60% and demurrage and detention costs by $720,000 a year with 36 agents. It fits less well for teams wanting self serve prices, a named model, or an agent that operates screens itself.

Vendor details

Canonical URL

https://www.getwiq.ai

Category

Agent builder

Subcategory

Bardeen's enterprise platform that records how operations teams work, turns each process into an approved agent blueprint and deploys governed agents to the runtime the company already uses

Funding status

A product of Bardeen, Inc., backed by Insight Partners, FirstMark, 468 Capital, HubSpot Ventures and Dropbox.

Company status

first party product

Use cases & customers

Primary use cases

process discovery from observed workprocurement and purchase order automationorder to cash and payment operationslogistics exception handlingKYC and onboarding operations

Target customers

enterprise operations leadersprocurement and finance operations teamslogistics and supply chain operationsconsultants running automation programs

Deployment options

WIQ Agent runtimeCustomer runtimes: Claude, Microsoft 365 Copilot, OpenAI Codex, Workato or any MCP platformLocal runs where the work happens or inside the customer's cloud

Integrations

Deploys to the WIQ Agent, Claude, Microsoft 365 Copilot, OpenAI Codex, Workato and any MCP compatible platform; MCP connection for querying process analytics, checking tool availability and updating blueprints; example blueprints across ServiceNow, SAP Ariba, SAP S/4HANA and Outlook; discovery across Microsoft Teams, Excel, Notion, Gmail, Slack and Google Sheets.

In practice

A logistics company records how its exceptions team clears delayed containers. WIQ ranks the work by hours and value, builds a blueprint across ServiceNow and Outlook, and deploys 36 agents that cut exception handling time by 60%.

A medical equipment maker maps its receiving order process, approves a blueprint with a checkpoint before any SAP S/4HANA posting, and deploys it to Microsoft 365 Copilot, where every run is checked against the blueprint and deviations go to a supervisor.

An operations analyst uses the Try page to record a vendor onboarding task and downloads the result as a Claude skill to test, before the team books a demo to add escalation paths and run it under WIQ's control plane.

Agentic Index coverage score

9.0 / 14 capabilities · 64%

Integrations & Tool Calling Full

Each blueprint names the tools its agent needs, and admins scope exactly which tools each agent can reach and which actions it may take there. WIQ's own example blueprint works across ServiceNow, SAP Ariba, SAP S/4HANA and Microsoft Outlook, and discovery also maps work in Microsoft Teams, Excel, Notion, Gmail, Slack and Google Sheets. Before deployment the platform checks whether the tools a blueprint needs are available in the chosen runtime, so the agent arrives with working connections rather than missing ones.

Sourcegetwiq.ai/deployread 2026-11-19

Workflow Orchestration Full

A blueprint describes how a whole cross department process runs over time: its ordered steps, the tools each step uses, the approval points and the escalation paths. WIQ's pipeline builds an agent for each discrete task in that process, so a procurement flow becomes several agents handing work along, and concurrent sessions let agents work a ticket backlog in parallel. Fixed screen steps on legacy systems can sit inside a blueprint as a tested script the agent calls when a decision has been made.

Sourcegetwiq.airead 2026-11-19

Knowledge Grounding & RAG Full

WIQ builds a lasting model of how the business runs. It combines recordings of sanctioned work, systems of record, documentation and answers from experts to reconstruct each process, with the hours it takes per week and the systems involved, and ranks processes by automation value and readiness. That process knowledge stays current: when the process changes or an agent meets a new exception, WIQ surfaces the change, and agents and other platforms can query the process analytics over MCP.

Sourcegetwiq.airead 2026-11-19

Human Oversight & Guardrails Full

People sign off twice. Teams review every blueprint before it deploys, and inside it they set approval gates, escalation paths and checkpoints where the agent pauses for a person's review before going on. Updates follow the same path, with WIQ surfacing a changed process for the team to approve before the new blueprint ships. These guardrails belong to WIQ itself, so a skill exported from the free Try page leaves them behind.

Sourcegetwiq.ai/deployread 2026-11-19

Security, Identity & Governance Partial

WIQ carries SOC 2 Type II, GDPR and CCPA badges, and its capture extension records only what policy allows, by app, domain and time window. Agents get least privilege scoping of the tools and actions they can use. For the people running the platform, though, there is no role model, single sign on or SCIM setting that decides who may approve a blueprint, deploy an agent or view recordings, and the legal entity behind the service is not named in its terms.

Sourcegetwiq.ai/deployread 2026-11-19

Observability & Auditability Full

A central control plane lists every agent run with its operator, timestamp, platform and duration, across WIQ's own agent and the outside runtimes it deploys to. Each run is checked against the approved blueprint and logged in full, so a deviation shows which step departed from the agreed process and who it was routed to. Discovery views add the hours each process takes per week, giving a before and after picture of the same work.

Sourcegetwiq.ai/deployread 2026-11-19

Memory & State Persistence Not documented

What carries forward between runs is the approved blueprint, a set of steps, tools and rules that people review and sign off, and WIQ revises it when processes change or new exceptions appear. Those are instructions the agents follow rather than anything an agent remembers on its own. There is no store in which agents keep context per user, customer or case from one session to the next, with a set lifetime and a way to clear it.

Sourcegetwiq.airead 2026-11-19

Deployment & Data Residency Full

Agents run locally where the work happens or inside the customer's own cloud, so data from SAP, ServiceNow or Outlook can stay inside the customer's environment. The same blueprint can also deploy to an agent platform the customer already operates, whether WIQ's own agent, Claude, Microsoft 365 Copilot, OpenAI Codex, Workato or any MCP compatible runtime. Where WIQ hosts its own discovery and control plane is not stated.

Sourcegetwiq.ai/deployread 2026-11-19

Prebuilt Agents / Templates / Packs Partial

WIQ's agents are generated from each customer's own recorded work rather than taken off a shelf. It organizes its offer by function, with solution pages for customer support operations, payment operations, KYC and onboarding, procurement, fraud and financial crime, order to cash, logistics and commodities, and a resources section of templates and guides. None of these is a ready agent a customer switches on without first mapping its own process.

Sourcegetwiq.airead 2026-11-19

Triggers & Channel Coverage Partial

Runs in the control plane carry the name of the operator who launched them, and teams start batches of concurrent sessions to work through a queue such as a ticket backlog. Beyond that, how an agent is set off depends on the runtime it is deployed to, whether a Workato recipe, a Copilot or Claude conversation or WIQ's own agent, and WIQ itself names no schedules, webhooks or event triggers that start a blueprint when a new order, ticket or email lands.

Sourcegetwiq.ai/deployread 2026-11-19

Model Flexibility & Routing Not documented

Choosing a runtime is not choosing WIQ's model. A blueprint deployed to Claude, Microsoft 365 Copilot or OpenAI Codex runs on the customer's own assistant, which picks its own model. Inside WIQ, a proprietary AI pipeline builds the blueprints and agents and the WIQ Agent runs them, and neither the models behind them nor any picker, bring your own key option or routing control for the customer is named.

Sourcegetwiq.ai/deployread 2026-11-19

APIs / SDKs / MCP Extensibility Partial

Agent platforms connect to WIQ over MCP, and from inside them a team can query process analytics, check whether a blueprint's tools exist in that environment and update blueprints, so WIQ can be driven from Claude or another MCP client. Its endpoint, sign in method and full list of tools are not given out, and there is no REST API, SDK or developer documentation alongside it.

Sourcegetwiq.ai/deployread 2026-11-19

Testing, Debugging & Optimization Full

Every agent run is checked for conformance with its approved blueprint, and teams set conformance requirements and SLA thresholds for each session, so a run that skips a step or misses its time limit is caught and routed. Because WIQ measured the same process by hand during discovery, it compares an agent's handle time, SLA performance and conformance with the observed human baseline, which shows whether the automation actually beats the work it replaced.

Sourcegetwiq.ai/deployread 2026-11-19

Browser / Computer-use Not documented

WIQ's browser extension watches people work, recording the apps and steps they use within allowed domains and times, and that capture feeds discovery rather than letting an agent drive the screen. The deployed agents act through the tools and runtimes a blueprint names. Where a legacy screen has no interface, the blueprint calls a tested script instead of having the agent click through a browser or desktop session itself.

Sourcebardeen.ai/answers/what-is-an-ai-agent-blueprintread 2026-11-19

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

Quote only, through a demo; the Try page turns one recording into a free Claude skill

Not stated

Trial available

What is public

WIQ sells through a demo and a discovery session with an AI architect, and its pricing link leads to the demo request form. A free Try page turns a recording of one task into a downloadable Claude skill.

Billing mechanics

WIQ does not say whether it charges by process, agent, run or seat.

Cost watchouts

Agents that run on Claude, Microsoft 365 Copilot, OpenAI Codex or Workato consume the customer's existing licenses or usage on those platforms alongside the WIQ contract.

Variable cost rationale

WIQ gives no billing unit, and agents deployed to Claude, Copilot or Codex also run on the customer's own platform subscriptions.

Additional watchouts

The free Claude skill leaves out the escalation paths and guardrails that WIQ's control plane adds.

Sales call required

Yes, required for paid access

Free / trial

Try page converts a recording of one task into a downloadable Claude skill, without WIQ's escalation paths or guardrails

Commercial notes

Agents deploy to the customer's chosen runtime, from WIQ's own agent to Claude, Microsoft 365 Copilot, OpenAI Codex, Workato or another MCP platform.

Missing data

WIQ gives no price for its platform or for deployed agents.

Agentic Index verified 2026-11-19

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.