Skip to content
  • There are no suggestions because the search field is empty.

Analyzing Documents with an AI Assistant

Start using the AI Assistant to uncover spending trends, notice gaps in your processes, and develop an efficient procurement strategy.

TABLE OF CONTENTS

What Can Precoro AI Assistant Do for Your Company

With the AI Assistant, any depth and angle of analytics is at your fingertips. Without any manual work and time spent manipulating spreadsheets, you can:

  • Learn how your funds are distributed across teams, vendors, categories, and time periods.
  • Find anomalies, duplicates, or glaring errors in your purchases.
  • Detect common holdups that indicate your processes need improvement and get personalised tips for updating them.
  • Get a supplier evaluation to see if they are reliable and punctual. The assistant can also suggest alternative vendors to ensure your business needs are always met.

Please note: This functionality is currently in Beta. Contact your Customer Success Manager to enable it for your company, try it out, and share your thoughts via the Open Product Plan.

How to Use the AI Assistant

The AI Assistant is available in all billing plans.

To access it, the user needs to have a purchase order or invoice role. According to their role, they can analyze POs, invoices, or both, including the suppliers, items, locations, and custom fields in the documents.

Once the Assistant is enabled, follow these steps:

  1. Click AI Assistant in the navigation menu at the top of any Precoro page. The chat window will appear in the bottom left corner.
  2. Select which type of documents you want to analyze (orders or invoices).
  3. Filter the documents. You can analyze purchases over a certain period, from a specific supplier, with a needed status, etc. Limiting the dataset will help you get more precise and relevant results.
  4. Enter your question at the bottom of the chat window. If you’re not yet sure what you can ask, check out the following section for examples. 
  5. Once you get the initial answer, ask follow-up or related questions to deepen your understanding and use the assistant’s capabilities to the fullest.

When you reopen the assistant, the chats will still be accessible from the Previous requests section. You can get back to them anytime if you have further questions or want to review the information. You can also delete any chat if you no longer need it. Inactive chats are automatically deleted after 60 days.

Use Cases and Request Examples to Get Started

The assistant can provide answers and insights on any topic based on your data. The vast possibilities may be overwhelming at first, so we’ve prepared a simplified list of what you can request. Please look through the topics and sample questions you can ask the AI Assistant as you start working with it.

Spend Analytics

  1. Learn how your spending patterns have changed over time.
  2. Request a spend breakdown by departments, projects, or locations.
  3. Identify top vendors by volume or value.
  4. Access the finest details when analysing spend by item.

Questions:

  • How much did we spend last month compared to the previous one?
  • Which department is spending the most this year?
  • What’s the variance between actual spend and budget this quarter?

Supplier Insights

  1. Find out if you depend on any suppliers and whether they are reliable.
  2. Assess supplier performance by analyzing the timeliness of invoice approvals and PO fulfillment rate.
  3. Request a price comparison if suppliers offer the same item.

Questions:

  • Who are our top 5 suppliers by spend?
  • Which suppliers deliver on time vs late?
  • Do we have multiple suppliers for the same item SKU?

Process Assessment

  1. Compare document creation and approval dates to detect bottlenecks.
  2. Gather information about payment or receipt delays.
  3. Evaluate if documents are rejected or cancelled too often.

Questions:

  • What’s our average PO approval time?
  • Where do approvals get stuck most often?
  • Which invoices are overdue for payment?
  • What’s the average time from invoice creation to payment?
  • What’s the ratio of approved orders to canceled ones?

Forecasting and Predictive Analytics

  1. Estimate your expected spend based on procurement trends.
  2. Forecast the cash flow accounting for the dates in the currently existing documents and payment terms.
  3. Notice and prevent emerging risks with budget consumption and supplier behavior prediction.

Questions:

  • What’s our projected spend for next quarter?
  • How much cash outflow should we expect in the next 30 days?
  • Which vendors put our operations at risk due to late deliveries?
  • If our current spending trends continue, will we exceed the annual budget?

Don’t hesitate to ask other questions and follow up after you receive the answer. This will ensure that you get the most precise responses relevant to your company.

FAQs

What if I encounter an issue with the assistant?

As the functionality is in Beta, it may not be completely stable. If you get an error message or the assistant doesn't respond, try creating a new chat. If you notice other issues or have ideas on improving the assistant, please let us know via the Open Product Plan.

Will the AI use my data without my permission?

No. The AI only works when you choose to use it, and any information you provide is handled according to our Privacy Policy.

Does the AI share my data with others?

When you use the AI features, your data may be processed by our trusted technology partners (such as OpenAI) to provide the service. This processing is done securely and only for the purpose of delivering the AI functionality. Your information is always handled according to the Precoro Privacy Policy and is never sold or shared for advertising.

Can the AI place orders or make commitments for me?

No. The AI is designed to assist with tasks like data analysis, but it can't create obligations or act on your behalf.

How accurate are the AI Assistant's answers?

The AI is designed to provide helpful insights and save you time with tasks like data analysis. While it's highly effective, no AI is perfect, so we recommend reviewing the results before making final decisions.