Know the business.
Bring approved policies, customer records and operational data into the task—with source references and permission boundaries.
A core NyKinSky product · Concept preview
NyKinSky Workspace Agents
From conversation.
To coordinated
action.
An enterprise agent platform vision for the work between your systems. Connect knowledge, coordinate specialist agents, and keep people in control of consequential decisions.
Interactive prototype available below. Enterprise runtime and integrations are planned, not currently available for production use.
WORKSPACEAGENTS
01Enterprise contextKnowledge with provenance
02Specialist agentsOne coordinated workflow
03Human checkpointReview before action
PRODUCT ARCHITECTURE VISION
INTELLIGENCE INTO ACTIONExplore the product ↓
The product ambition
Not another place to chat.
A new way to get work done.
Today, a request can travel through five tools and three teams before anything changes. Workspace Agents is NyKinSky’s proposed coordination layer: understand the request, assemble the evidence, prepare the next step, and involve the right person.
Work as a team.
Give specialist agents a defined job. Share only the context they need, and pass a structured result to the next stage.
Keep decisions owned.
Make approval, escalation and stop controls part of the workflow. Measure completed work, not just generated answers.
Try the product idea
One request.
A visible path to action.
Choose a scenario, run the workflow and make the final decision. Every record, policy and result here is synthetic.
Workspace AgentsEXPERIENCE LABILLUSTRATIVE BUSINESS REQUEST
A customer’s order is late. What should happen next?
Review order DEMO-1042, explain the delay and propose an appropriate service recovery.
Workflow trace
- Load the illustrative workflow with “Run simulation”.
Evidence desk Synthetic
APPROVAL RULE
Customer-facing offers require a service lead’s approval.
YOUR DECISION · DEMO REVIEWER
Review the evidence before allowing this simulated action.
Ready to explore. All data and actions are simulated.
No external action is executed, even after approval.
Proposed agent portfolio
Specialist by function.
Connected by design.
Start with one bounded workflow. Expand only after its quality, ownership and business value have been demonstrated. These are planned product modules—not live integrations.
Service Agent
Assemble customer context, explain an issue, draft a response and prepare a resolution for the service team.
- Human checkpoint
- Refunds, exceptions and outbound commitments.
- Pilot measures
- Resolution quality, reopen rate and handling time.
Revenue Agent
Prepare account research, organize qualification evidence and suggest a next-best step for the opportunity owner.
- Human checkpoint
- Outreach, commercial terms and opportunity changes.
- Pilot measures
- Preparation time and accepted recommendations.
Finance Agent
Reconcile supporting records, highlight mismatches and prepare an exception pack for finance or procurement.
- Human checkpoint
- Payments, supplier changes and ledger postings.
- Pilot measures
- Exception accuracy, review time and rework.
People & IT Agent
Find approved policies, prepare onboarding checklists and triage internal support requests.
- Human checkpoint
- Access changes and sensitive employee decisions.
- Pilot measures
- Answer accuracy, escalation quality and completion.
Operations Agent
Connect order, inventory and delivery signals; identify exceptions and prepare a coordinated response.
- Human checkpoint
- Supplier instructions and production changes.
- Pilot measures
- Detection quality and time to approved resolution.
Knowledge Agent
Turn authorized research and internal documents into sourced briefs, comparisons and decision support.
- Human checkpoint
- Conflicting evidence and external publication.
- Pilot measures
- Citation correctness, coverage and retrieval quality.
The planned architecture
Your systems.
Your context.
Your boundaries.
The ambition is an agent layer that works with an organization’s existing technology estate, rather than requiring every process to move into a new application.
Each connection would need its own access review, implementation and testing. No third-party connectors are operational in this preview.
Map your architecture with us ↗EXPERIENCE
Request, review, collaborate
Employee workspace · Client channel · Approval inbox
WORKSPACE AGENTS
Coordinate the work
Task routing · Specialist agents · Bounded execution
CONTROL PLANE
Permissions before actions
Identity · Policy checks · Evaluations · Action records
CONTEXT & CONNECTIONS
Approved business systems
CRM · ERP · Service desk · Documents · Data platforms
MODEL & INFRASTRUCTURE OPTIONS
Fit the workload and environment
Model evaluation · Deployment review · Cost limits
Governance is part of the product brief
Trust needs something
you can inspect.
The following controls are production requirements on our roadmap. The browser demo illustrates an approval checkpoint; it does not implement enterprise security.
Access with limits
Tenant isolation, scoped identities, source-system permissions and managed secrets. No blanket access to company data.
Evidence, not certainty
Source-linked answers, retrieval evaluations and explicit handling of missing or conflicting information.
Permission to act
Allowlisted tools, approval thresholds, previewable changes and escalation to accountable business owners.
Operational visibility
Action logs, quality and cost monitoring, incident response and an emergency stop before production rollout.
An original direction in enterprise agents
Built around your operating model.
Our intended differentiation is the combination of cross-function workflows, business-process consulting and hands-on implementation—not an unproven claim of feature parity or lower cost.
The category benchmark
Salesforce Agentforce offers enterprise AI agents, agent-building tools, orchestration, monitoring and connections to business systems.
Read Salesforce’s product overview ↗Reference reviewed 17 September 2026. Salesforce and Agentforce are their owners’ trademarks. NyKinSky is not affiliated with or endorsed by Salesforce.The NyKinSky product thesis
- Start from a measurable business workflow, not a chatbot deployment.
- Design around an existing mix of systems and model choices.
- Combine specialist agent templates with implementation support.
- Make evidence, approvals and ownership visible to the business.
A transparent path to the product
What you can explore today.
What needs to come next.
- 01 / AVAILABLE NOW
Concept & simulation
Product overview, three synthetic workflows, evidence panels and an interactive approval gate. Enquiries use NyKinSky’s existing website form.
- 02 / PLANNED
Controlled pilot
Implement a real model, one scoped connector, authenticated access, evaluation cases and logging in an agreed test environment.
- 03 / PLANNED
Enterprise foundation
Validate tenant boundaries, approval enforcement, observability, retention controls and operational recovery before expansion.
- 04 / FUTURE DIRECTION
Portfolio & scale
Extend validated workflows, agent templates and partner delivery. Production release depends on successful validation; no launch date is committed.
Design the first useful workflow
Bring a bottleneck.
Build a better way.
Tell us which team is affected, where work gets stuck and which systems hold the evidence. We will discuss whether a scoped agent pilot is a good fit.
Request a product consultation ↗YOUR PILOT CONVERSATION
Clarity before commitment.
- Workflow and baseline performance
- Data access and system dependencies
- Approval owner and risk boundaries
- Evaluation criteria and stop conditions
- Engineering scope and estimated investment
Estimated investment is scoped individually around integrations, data readiness, usage and support. No fixed fee or outcome guarantee is advertised.
For NyKinSky business offices →Future partner delivery or product access would require a separate agreement, training and access authorization; it is not automatically included with an office model.Before you begin
A few straight answers.
Is this already a live alternative to Agentforce?
No. This introduces NyKinSky’s proposed enterprise agent product and a working browser simulation. It is not a production agent service and does not claim equivalence to a deployed platform.
Does the demo call an AI model or connect to my systems?
No. It follows predefined scenario logic with synthetic records. It does not generate AI responses, authenticate to your systems, send messages or execute business actions. Reset clears its in-memory workflow state.
Which models and cloud services will it support?
Provider and deployment choices will be evaluated against each pilot’s requirements. No provider integration, partnership, certification or data-residency commitment is claimed by this preview.
How is this different from NyK & NyKi?
NyK & NyKi is the website’s guidance assistant for services and office models. Workspace Agents is the proposed enterprise product for coordinating business workflows. This preview does not change the website assistant.
How do we evaluate business value?
Agree a baseline and a representative test set first. Measure task quality, reviewer acceptance, completion time, escalation rate and total operating cost. Any improvement needs to be demonstrated with your workflow, not inferred from a demo.
Can I share confidential information to request a pilot?
Please share only a high-level description in the enquiry form. Access to confidential records, regulated information and credentials should be agreed separately through an approved secure process.