Building Smarter: How this Minnesota based construction company is improving it’s AP productivity with Nanonets

We work on a variety of processes collaboratively and Nanonets is part of our daily flow here. Nanonets has been crucial in helping us scale processes as we grow.

 We’ve done a good job of getting a very high percent of our invoices. It remembers everything and so quickly, that's fantastic! Our service representatives love the easy and fast user interface that they were able to use immediately without any special training.

- Adam, CEO


Before Nanonets

Classic Build vs Buy Dilemma

Significant engineering time was spent on building a traditional data extraction architecture, as well as on experimenting with existing traditional data extraction platforms.

High expense on Manual Invoice Processing

An annual salary of over 100K to simply maintain a manual verification component — unjustified ROI.

After Nanonets

Focus on product = global scale

Renewed focus on onboarding more clients - more quickly and with more ease, customer experience and delivering the clients their favorite dream homes! 

Seamless Invoice Processing Solution

Automated Accounts Payable Solution improving team's productivity and value creation


increase in processing speed


work hours reprioritized


% Time Saved On Verification


Minnesota based Construction Company


Construction and Demolition

Key Stakeholders


Key Features Required

Table Extraction, Accurate Automation, Integration with Accounting Software

Long Term Goal

Renewed focus on customer experience and onboarding more clients

Table Of Contents

Who is the ClientWhat Industry Problems was the client facing?Challenges with Textract and AbbyyWhat was the NanoNets Solution?What was different about NanoNets?
What Long Term Goals was the client able to achieve?

Who is the Client?


Adam is the CEO of a construction company based in Minnesota. He explains his role in building homes quite simply: “I have to accomplish my short term day to day goals in order to accomplish my long-term goals. I need to be a people person who helps clients build their favourite homes and buildings!” He adds

But don't forget all that paperwork everyday. I suppose, don’t forget about the computer work too.


With over 400 successful projects under their belt, this Minnesota-based construction major has been leveraging world class building solutions along with 50 years of local construction knowledge to consistently rank among the top 100 metal builders nationwide. Involved in planning & design, budgeting, design-build & construction management for clients such as Polaris, AGCO, and Winnebago among others, this family owned construction company has been a regular at Butler Manufacturing’s annual awards program; picking up awards such as the Builder of the Year, National High Performance Builder, and The Butler Club over the years.

What Industry Problems was the client facing?


Construction is the second-least digitized industry in the United States, yet it accounts for 10 percent of the country’s GDP. A staggering figure! It is to be noted, however, that manual interventions and processes still play a huge role in this largely paper document driven industry. A lot of time and resources are spent in data entry & manual verification. 

Embracing automated document processing could therefore optimize processes and save a lot of time and money. Traditional automation solutions have so far been incapable of dealing with unstructured data; documents in the field come in varying shapes & formats and are often difficult to scan accurately due to marks/blemishes, handwriting or background. 

So while automation clearly provides benefits, the search is still on for a trustworthy, accurate & secure automation software that can extract and process unstructured data in all its diversity and present it in a convenient manner for further use within an organization (Accounting, HR etc.). And that’s where Nanonets can help!

A Standard Construction Invoice

This construction organization deals with a large volume of invoices that need to be processed and the processed data entered into an accounting software. These invoices present detailed charges for repairs and contractor assignments. The management implemented an automated solution (first with Textract and then with Abbyy) to capture all the information from these invoices in a more efficient manner; avoiding manual processes that are fraught with delays and errors. The automated solution was expected to pull out the following typical fields from the invoices:

  • The name, address and contact information of the contractee.
  • The name, address and contact information of the direct contractor.
  • The name and address of the construction investor, if applicable.
  • The location of the job site.
  • The name and address of the individuals who are receiving or will be receiving the work.
  • List and quantity of all supplies, materials and cost for each unit along with cost of labor.
  • Sales and state tax as applicable.
  • When payment is due and what payment type is accepted.
  • Information regarding when the payment is considered late and any penalties assessed by the contractor.
  • Information regarding Estimates and what occurs if labor hours are long or shorter or if materials and supplies change.
  • Deposit being requested, if applicable

What were the Challenges with Abbyy and Textract?


“You have to cut different pieces of invoices, and then sew them up well enough to keep financing running, smoothly and efficiently. After about 3 months of struggling with Textract and Abbyy, I feared we’d have to go back to entering data manually. It was a disaster” - Adam, the CEO.


The experiments with Textract and Abbyy were not encouraging.

  • Setup effort - The setup for both tools was pretty elaborate and time consuming.  Abbyy required an extensive mapping of individual vendors to a particular invoice layout. Setting up Textract too was pretty challenging and rigid.
  • Ability to process multi-page invoices - Both Abby and Textract stumbled when dealing with multi-page invoices. While Textract faced a lot of difficulty when handling such invoices, Abbyy simply failed to process them. 
  • Ability to capture multi-line fields - This was just not supported by Abbyy or Textract which was extremely inconvenient as most detailed documents contain multi-line fields.
  • Accuracy - Both tools returned a lot of incorrect/inaccurate data that needed verification and rework; failing the core purpose of automation.
  • Unstructured data & imperfections - Documents with unstructured data and imperfections (blemishes, marks, blurred backgrounds etc.) had a high error rate (especially with Abbyy). Abbyy  & Textract also failed to identify key value pairs in such cases.
  • Verification UI - The verification UI was difficult to use, taking up to spending ~5 minutes per invoice, slowing down the team even with hard-ruled automation.

Finally it was impossible to train the algorithms to rectify these crucial flaws! The company was forced to drop Textract & Abbyy, and look for an alternative.

Could Nanonets handle The Four Critical Problems that had tripped up both Textract & Abby?

A quick trial run demonstrated Nanonets’ superior capabilities. In comparison to Abbyy & Textract, Nanonets was both easy and flexible to set up, requiring just about 1 day. It also handled multi-page invoices and identified multi-line items with ease; something that both the previous tools had failed at. Nanonets also customized column headers allowing it to process complex invoices more efficiently. Nanonets’ AI ensured a high accuracy while processing documents requiring minimal rework or revision. The automation handled unstructured data without much difficulty. The AI also extracted information from documents with imperfections & blemishes quite easily. 

And verification that would earlier take over 5 minutes for batches of documents could be blazed through in under 30 seconds with Nanonets’ modern user interface!

Any automated solution approved by the company had to address the following four critical problems:

Invoice and Material list data are provided together in PDFs. Both documents are in the same file and they look alike. Can the software recognize the document type and extract the correct data?

The Nanonets algorithm can filter, recognise and parse different types of documents such as invoices, material lists, packing lists, purchase orders etc. all at the same time! The Nanonets AI can segregate critical documents from non critical ones, and can selectively parse critical documents to extract the required key value pairs.

There are over 40 different suppliers with different invoice formats. Will the software need to be trained separately for each supplier? What if new suppliers are on-boarded every month? Does the AI need to be retrained for each new supplier?

Nanonets offers an AI based solution which is template agnostic. Our AI learns to understand the document based on a large number of features. Once trained on a diversified data set, the model can recognize the information in any new format.

Additionally, the AI retrains itself with the data you process. This ensures that the algorithm functions accurately even if you onboard new suppliers each month; helping you to scale efficiently while maintaining accurate results.

The Accounting Software can ingest the API or CSV only if the fields are presented in a certain format. Can the output be customized accordingly?

Nanonets is not bound by the template of the document at all. Custom validation rules allow you to reorganize data into convenient output layouts and formats that are easier for further processing. You can export the appropriately organized fields/data as an xml, xlsx or csv file.

Sometimes, the Product Description cell in the Invoice table also contains the Article Number and Weight associated with the Product Name. Can those be pulled out as a separate column?

With Nanonets, you can pull out specific fields under each row in the table as separate columns.

What was the NanoNets Solution?


"Instead of spending hours and hours engineering a workflow to sync key information from thousands of invoices to our software, we're able to automate this accurately and spend more time on highly strategic initiatives."

  • The AP team at the client site sends invoices to Nanonets using a simple API integration.
  • Nanonets AI automatically processes these invoices in real time and populates a csv that gets stored in the AP team’s Drive. 
  • The AP team, at the end of the day, downloads the csv and uploads it to their accounting software.

Extracting data from invoices provided by over 40 different suppliers (and growing) can be a challenging problem. Nanonets’ AI refined its models by using the client’s existing data; picking out the particular needs of the business. Unlike most document processing APIs out there that are quite rigid on the type of data they can work with, Nanonets isn’t bound by the template of the documents!

Adapting Nanonets for the specific business rules required by the client was easy and straightforward. Nanonets can capture as many fields of text/data that you like and present it in any desired fashion. For instance, translating the date format to ISO 8601 or extracting the currency of the total invoice, before populating the csv was easily enabled by Nanonets’ plug and play webhooks. Captured data was presented in tables or line items with custom validation rules. 

Businesses often face dynamically changing requirements; for example, the company brings on board a new vendor or is presented with a new invoice format. Nanonets OCR API has enabled Adam and his team to tackle such situations with ease. Nanonets can easily re-train existing models with new data. Adam and his team have since been able to:

  • Retrain their OCR model
  • Add new fields
  • Integrate the API without any assistance from Nanonets

What was different about NanoNets?

After moving to Nanonets, the construction company automated the most labor intensive steps of the document collection process; resulting in a 10x increase in processing speed & 7200 work hours being reprioritized. Nanonets’ expertise in working with construction companies and familiarity with industry standards regarding documentation made it an ideal fit for Adam’s company!

The benefits of using Nanonets go just beyond better accuracy, experience and scalability. Here are 8 reasons that highlight the Nanonets advantage:

  • Training & working with custom data - Most document processing APIs out there are quite rigid on the type of data they can work with. Nanonets isn’t bound by such limitations. Nanonets uses your own data to train models that are best suited to meet the particular needs of your business.

  • Easy to use & flexible - Adapting Nanonets for your specific business needs is easy and straightforward. From creating custom OCR models & retraining them to adding new fields & handling integrations, Nanonets can handle it all.

  • Learns & retrains continuously - Businesses often face dynamically changing requirements and needs. To overcome potential roadblocks, Nanonets OCR API allows you to easily re-train your models with new data. This allows your OCR model to adapt to unforeseen changes.

  • Customise, customise, customise - Nanonets can capture as many fields of text/data that you like and present it in any desired fashion. Captured data can be presented in tables or line items or any other format of your choice with custom validation rules. Always remember that Nanonets is not bound by the template of your document!

  • Requires almost no post-processing - While most document processing APIs simply grab and dump data, Nanonets extracts only the relevant data and automatically sorts them into intelligently structured fields making it easier to view and understand. This does away with a lot of time spent in revision and verification.

  • Handles common data constraints with ease - Nanonets leverages deep learning & object detection techniques to overcome common data constraints that greatly affect text recognition and extraction. Nanonets AI can recognize and handle handwritten text, images with low resolution, images with new or cursive fonts and varying sizes, images with shadowy text, tilted text, random unstructured text, image noise, blurred images and more. Traditional OCR APIs are just not equipped to perform under such constraints; they require data at a very high level of fidelity which isn’t the norm in real life scenarios.

  • Works with non-English or multiple languages - Since Nanonets focuses on training with custom data, it is uniquely placed to build a single model that could extract text from documents in any language or multiple languages at the same time.

  • Requires no in-house team of developers - No need to worry about hiring developers and acquiring talent to personalize Nanonets API for your business requirements. Nanonets was built for hassle-free integration. You can also easily integrate Nanonets with most CRM, ERP or RPA software.

What Long Term Goals was the client able to achieve?

long term goals

Levie, Adam’s business associate, weighs in on one aspect of this project: 

“The ‘location of the job site’ data is a new field in the contracts. Earlier, I would pull up all invoices individually or email the clients to add this info; which was a horrible experience! Instead now, Nanonets automatically  pulls out this field and feeds it into my Contractor template using Zapier.”


With a streamlined system in place to process documents, the company is now implementing Nanonets for other projects. Next up is an efficient way to auto-populate Contract Agreements with Nanonets through a Zapier integration. The goal is to organize fields extracted by Nanonets in an appropriate manner and create an official agreement for potential customers.

Automating the process of providing invoices, quotes or agreements is an effective way to optimize the sales funnel and plug leaks. Automation helps convert more Prospects into Sales by removing unnecessary delays, reducing turnaround times, and avoiding human errors.

Adam’s construction company now spends less time worrying about paperwork and more on their core competency: building their customers’ dream projects! 

The best document processing solution for your business

Cookies Preferences
Close Cookie Preference Manager
Cookie Settings
By clicking “Accept All Cookies”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage and assist in our marketing efforts. More info
Strictly Necessary (Always Active)
Cookies required to enable basic website functionality.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.