Two ways to start with Dialogo
Dialogo provides AI agents to handle recurring work across company tools. We are also developing a data partnership that helps companies assess existing records and selected examples from agent workflows for potential paid licensing to AI developers.
You can start with an automation project, a data assessment or both. Existing records do not need to have been created by Dialogo.
The data partnership is in development. This article describes the proposed service and the decisions required before any release or payment. It does not announce a buyer relationship, completed sale or guaranteed payout.
What could a company license?
Useful examples may show how a business task was completed: its initial context, decisions, corrections and final result. A company could assess its own operational procedures, resolved cases, approval histories and expert work records.
Selected agent workflows can add the original request, relevant inputs, tool calls, employee corrections and a checked outcome. Human-created, agent-generated and mixed records should remain distinguishable.
Repeated low-information logs, unverified outputs and records without sufficient rights may be unsuitable. The amount of data is not a valuation.
What the service would do
1. Map potential data
Identify sources, ownership, task types and possible restrictions. Start with descriptions and a narrow scope rather than broad access to every system.
2. Check buyer fit and permissions
Establish what an actual buyer needs and what the company has the authority to license. Permission to process a customer’s information for operations does not automatically authorize a different use or external licensing.
3. Prepare and review a dataset
Select useful records, preserve relevant context, remove or exclude sensitive material and check outcomes. Review provenance, consistency and the distinction between human and generated content. Privacy treatment must be assessed for the actual material; replacing names is not an automatic assurance of anonymity. The EDPB’s AI-model opinion also emphasizes case-specific assessments of model anonymity; it is not a certification of an exported dataset.
Dialogo AI
Assess existing company records and selected agent workflows for potential paid licensing.
A separate assessment, with your approval before any release.
4. Approve terms and release
The company reviews the dataset, buyer, permitted purpose, restrictions, delivery and retention terms. External sharing requires a separate agreement and approval.
5. Receive the agreed payment
A completed buyer agreement defines the licensing fee or revenue share and its payment conditions. Assessment and preparation costs are agreed separately. If there is no acceptable buyer agreement, there is no licensing payment to assume in the company’s budget.
A workflow example
An agent investigates an order delay, checks the ERP and CRM, prepares a response, receives an employee correction and waits for approval. The resulting record might explain how the correct account was selected, why the first proposed response was changed and how the final action was verified.
That structure can be useful for evaluating whether another agent completes the same kind of task. The original customer identity and confidential order details do not have to be the intended product. Which fields can be transformed, excluded or retained requires review against the real dataset and agreement.
The operational benefit comes from the completed workflow. Any external licensing opportunity is assessed separately.
Why investigate this market?
Prism, micro1 and Scale publicly describe operational-data licensing programmes. Ooak Data describes turning company data into agent-training environments. These are examples of a market category, not evidence that they are Dialogo partners or will purchase a particular company’s records.
Their offers, published payouts and eligibility criteria should not be treated as Dialogo terms. We need actual buyer requirements and commercial evidence for each proposed dataset.
What should you bring to an assessment?
Describe the workflow, the systems involved, the kinds of historical records available, who owns them and any restrictions you already know. A description of one representative task is a useful starting point. Do not send confidential records before agreeing the scope and handling arrangements.