👤 For process owners and procurement teams
🔐 Using the resources depends on your company’s plan, enabled capabilities, and permissions
🎯 For people who want to analyze process decisions and identify an improvement for their pipe
Inside the process
Checking information, creating a purchase requisition, and tracking its progress are connected activities. When every handoff requires manual intervention, someone has to keep the request moving between systems.
By the end of this article, you will understand how Algar connected these activities through MIA, its virtual agent integrated with Pipefy, and identify decisions that can guide your own process analysis.
Commenting on the initiative, Zaima Milazzo highlights the development of:
“[…] a culture of innovation built on the practical application of technology to solve everyday challenges.”
Zaima Milazzo, Director of Products and Technology at Algar
Translated from Portuguese.
That practical application is visible in how requests arrive, information is checked, and work moves forward. Let’s examine those decisions and what they can help you notice in your own pipe.
What you will learn:
Where the process needed to change
Algar’s procurement, contract management, and operations teams relied on email and manual work across SAP, Jira, and internal systems. This made information harder to trace and procedures harder to standardize.
This situation draws attention to the work between two activities. Finishing a check may still require someone to notify a colleague, copy information, and follow up on whether the next action happened.
In your process, look at these handoffs. When buying equipment, for example, the person receiving the request may need to locate the contract before forwarding the order. Record where each piece of information comes from and who needs it to continue.
In Pipefy, several resources can help you investigate that path:
- Map: shows the pipes and databases connected to the process.
- Flow View: brings phases, forms, automations, and AI Agents together in a diagram.
- Card Timeline: lets you follow the actions taken on a request.
To choose where to start, read For people orchestrating processes: see what happens. It connects these views to questions that come up during operations.
Why validation comes before the next action
When a request arrives through Pipefy, MIA checks contracts, balances, validity periods, and financial information before creating a purchase requisition in SAP.
The sequence is the decision to examine: validation prepares the next action. Receiving a contract number and confirming that the contract is still valid answer different questions. A field can be filled in while the information still needs checking.
In your pipe, choose an important action and identify what needs to be known before it happens. Also define how to handle missing or conflicting information.
When designing a process that uses AI in Pipefy, consider:
- Reference knowledge: the policies and documents that provide criteria for the analysis.
- AI Agent behaviors: the task, inputs, and expected output.
- Testing and human review: how to check the result and handle decisions that require evaluation.
The article AI Agents in practice: from triage to monitoring, available in Portuguese, helps you define the task and prepare your tests. These are options for your own design; the success story does not describe the specific configurations behind MIA.
To build on that preparation, take the Artificial Intelligence Specialist Certification. It covers process analysis, instructions, and AI Agent configuration.
How visibility helps teams decide where to act
Algar also uses dashboards to track approvals, owners, and request status. The success story reports improvements in control and traceability, along with less rework.
This part of the design raises a question: does monitoring help the team choose its next action?
In a purchasing routine, knowing how many requests are open helps you understand the workload. Identifying which ones are waiting for a particular person’s review helps you decide where to intervene.
In your pipe, start with the operational question. Then choose the data that can answer it.
Pipe reports let you select columns and filter cards, including by phase and assignee. Read How to create pipe reports to build that view.
Apply this analysis to your pipe
Choose a recurring request and follow its path to completion. It might be a purchase, a reimbursement, or a service request.
At each handoff, record:
- What information needs to be available.
- What check allows work to continue.
- Who acts when there is a discrepancy.
- What the next stage receives.
Choose one handoff to improve and decide how you will assess its effect. If requests keep coming back because information is incomplete, track that reason for return before and after the change.
To turn this assessment into an implementation plan, continue with the Process Implementation Certification, available in Portuguese. The course covers process mapping, business objectives, and process monitoring.
Before moving on, confirm that you understand:
☐ Where information is transferred manually.
☐ Which validations need to happen before an action.
☐ Which situations require human review.
☐ What change you will test and how you will assess its effect.
Take inspiration from Algar to connect your procurement process
Does your team also need to check information and keep requests moving across systems? Talk to the Pipefy team about connecting these activities through AI and orchestration, with your company’s needs in mind.


