Pulse for Sales: An AI Workspace across CRM, ERP, and Company Systems

Pulse (Pro & Enterprise) can turn the systems a company already runs into one AI workspace for sales teams. Salespeople work in natural language. Orchesty executes predefined processes across CRM, ERP, helpdesk, document storage, and external sources.

Pulse is a conversational working layer across the systems a salesperson already uses through the sales cycle.

Sales work spans many systems #

A typical salesperson rarely works in the CRM alone. Before a customer meeting, they may need:

• CRM: contacts, opportunities, previous activities, meetings, and customer-related communication
• ERP: orders, revenue, prices, invoices, stock availability, and delivery status
• Helpdesk: open incidents or complaints
• Document storage: offers, contracts, and product materials
• External sources: current information about prospects and customers

The information may already exist, spread across those systems. Salespeople spend time searching, switching context, and combining data before they can act. Pulse brings these tasks into one interface.

One interface, several sales capabilities #

A company can expose a set of sales processes through Pulse. The salesperson names the outcome. The topology knows which systems are involved and which calls to make. Each capability is a business action backed by an Orchesty topology:

• Prepare me for a customer meeting
• Evaluate a new prospect
• Check price, availability, or order status
• Create an offer or order
• Process a completed meeting

Prepare me for a customer meeting #

A salesperson can ask Pulse to prepare them for an upcoming meeting. The topology retrieves selected information from CRM, ERP, and helpdesk, can enrich it with public information, and combines it into a concise briefing. The output may include:

• recent customer activity
• previous and upcoming meetings
• open opportunities
• recent orders
• revenue development
• overdue invoices
• open support issues
• important changes since the last interaction
• relevant public information
• potential risks or sales opportunities

The salesperson gets the context for the meeting from one answer.

Evaluate a new prospect #

A salesperson can provide a company name, a website, or an incoming lead. The process can:

• identify the company
• check whether it already exists in the CRM
• retrieve selected public information
• compare the company with the ideal customer profile
• evaluate its relevance
• create the company, contact, or opportunity when appropriate
• add a structured note and a next step

The model handles interpretation, enrichment, and scoring. CRM operations stay explicit process steps.

Check price, availability, or order status #

A salesperson can ask "Is product X available for this customer?" or "What is happening with order 12345?" The topology identifies the relevant customer, product, or order and retrieves the data from CRM and ERP. Depending on the request, that can include:

• customer-specific pricing
• current stock availability
• expected availability
• order status
• delivery status
• invoice status

The result comes back as a short answer the salesperson can use immediately.

Create an offer or order #

Pulse can run sales operations as well as answer questions. A salesperson can ask Pulse to prepare an offer from a customer request. The topology can:

• identify the customer
• identify requested products and quantities
• retrieve customer-specific commercial conditions
• retrieve current prices and availability
• prepare structured offer data
• create the offer in the ERP or the CRM
• generate the required document
• retrieve related product materials
• prepare customer communication

The same pattern can cover order creation and other recurring sales processes.

Process a completed meeting #

After a meeting, Pulse can process a transcript or the salesperson's notes. A single model step can turn the meeting into structured output:

• summary
• customer requirements
• commitments
• next steps
• a CRM note
• follow-up content

Later steps can update the CRM, create activities, and prepare the next customer interaction.

End to end: tomorrow's meeting with Acme #

A salesperson opens Pulse and asks: "Prepare me for tomorrow's meeting with Acme."

Pulse selects the corresponding topology and starts the process.

1. Identify the customer. The topology searches the CRM for Acme and resolves the company record, then retrieves account details, relevant contacts, open opportunities, recent activities, previous and upcoming meetings, and customer-related communication.
2. Retrieve commercial context. The topology queries the ERP for the same customer: recent orders, revenue over a defined period, current order status, overdue invoices, and stock or delivery information relevant to open opportunities. These are separate system operations in the topology.
3. Check open customer issues. The topology queries the helpdesk for unresolved tickets, complaints, or other active issues linked to the customer.
4. Add external context. If configured, a dedicated step retrieves current public information about the company, such as recent announcements, expansion, management changes, or other events relevant to the meeting.
5. Select the data for the model. The topology now holds data from several systems. Each preceding action returns only the fields the briefing needs. Detailed invoice records can be reduced to an overdue amount and the oldest due date. CRM data can be limited to relevant activities and active opportunities.
6. Generate the briefing. The selected data is passed to a model step with a defined output structure: a short customer summary, changes since the previous meeting, open commercial topics, unresolved customer issues, risks, possible sales opportunities, and suggested points to discuss.
7. Return the result. Pulse presents the finished briefing in the chat.

The salesperson gets one consolidated view. The company keeps control over how the data is retrieved and what reaches the model.

Controlled execution across company systems #

Pulse exposes business capabilities. Each capability is a topology that defines:

• which systems are accessed
• which operations run
• which data is retrieved
• which data is passed to the model
• which write operations are allowed

The model is used where interpretation helps. System reads and writes stay explicit process steps. A topology may work with more data than the model receives, because each step can keep only the fields the next action needs.

Pulse also works with the identity of the signed-in user, so available processes and data access can follow existing groups and permissions. See Pulse chat: identity, authorization, and execution (Pro & Enterprise).

Existing systems stay the systems of record #

CRM, ERP, helpdesk, and document storage remain the authoritative systems for their data. Pulse is the common interaction layer above them. Orchesty is the execution and integration layer underneath.

A topology can both read and change data in these systems: creating a lead, updating an opportunity, creating an offer, storing a note, or adding a follow-up task.

Salespeople get one place to retrieve information, run recurring processes, and work across company systems in natural language. Pulse turns existing sales systems and Orchesty topologies into a practical AI workspace that reduces administrative work and gives sales teams more time to sell.

Where next #

• Employee portal through Pulse: the same conversational layer for holiday, attendance, and approvals.
• Pulse chat: identity, authorization, and execution (Pro & Enterprise): signed-in identity, topology-level permissions, and trusted identity outside the model.
• What Pulse is (Pro & Enterprise): how Pulse exposes processes to people and to AI agents.
• Prepare your business for AI-driven solutions: why identity, permissions, and deterministic processes come before a conversational layer.