- Company
- Tavezio (formerly H&H Purchasing), a purchasing and management group based in Florida, United States
- Industry
- Procurement services — food and supply purchasing for clients including schools and summer camps across the US
- Process automated
- End-to-end invoice processing: file renaming, data extraction, verification routing, CSV export, storage, and exception handling
- Systems involved
- Zenphi, Google Workspace, Dropbox, Trello
- Previous approach
- Fully manual processing, with verification outsourced overseas and up to 9 temporary staff added each peak season
- Results
- 6× daily processing capacity, 90% cost reduction, more than $85,000 saved in one 3-month peak, same-day handling of every invoice, no overtime or temporary staff required
The problem was volume concentrated into a short season. Tavezio receives invoices daily from a wide range of suppliers, and the summer camp period generated thousands of documents that took months to clear by hand. The fix was an automated pipeline that renames each incoming PDF, extracts the key fields, routes the data through the team’s own verification logic, exports to CSV, files the document in Dropbox, and sends only the exceptions to Trello for a person to review.
Who Tavezio is
Tavezio, previously operating as H&H Purchasing, is a purchasing and management group based in Florida. The company manages food and supply purchasing on behalf of its clients, among them schools and summer camps across the United States, with a stated focus on maximizing savings through purchasing efficiency without compromising quality or integrity of service.
That business model puts invoice handling at the center of operations. Every client relationship and every supplier generates documents that have to be read, verified, recorded, and stored accurately.
Why manual invoice processing stopped scaling
Before automation, every invoice at Tavezio was processed by hand. The team read each document, keyed the data into their systems, verified it, and filed it. At normal volumes this was demanding. During the summer camp season, when thousands of invoices accumulated from a wide range of suppliers, it became a months-long backlog.
The workarounds were expensive. Up to nine additional team members were brought in each peak period to help process and verify the volume. Existing staff worked overtime. Invoice verification was outsourced overseas to keep up. And because people under deadline pressure make mistakes, accuracy suffered in exactly the period when volume was highest.
For a fast-growing company, the deeper issue was structural: a manual process cannot absorb growth. Every new client meant more documents and more hours, so the invoice workload would have capped how fast the business could expand.
What Zenphi automated
Josh Cohen, Owner and President, approached Zenphi during a peak invoicing period. His initial request was narrow: automatically rename incoming PDFs and store them in Dropbox for clients. What the team built covers the full process.
Each incoming invoice PDF is renamed to a consistent convention as soon as it is received, so documents are identifiable without being opened.
Zenphi pulls the fields the team needs out of each document, turning unstructured invoice PDFs into structured records.
Extracted data passes through Tavezio’s own verification workflow rather than a generic template, so the business rules stay theirs.
Verified information is exported to CSV for downstream use, and the source document is saved to the right location in Dropbox.
Anything the workflow cannot resolve is exported to Trello, so a person is brought in when judgment is genuinely required.
The design principle is worth noting for anyone evaluating a similar project: the automation handles interpretation and movement of data, while people keep the decisions that need review. That is what makes an exception queue small enough to be useful.
What changed after automation
| Measure | Before Zenphi | After Zenphi |
|---|---|---|
| Time to clear peak volume | Months of manual processing | Same day as receipt |
| Daily processing capacity | Baseline | 6× higher (reported 600% increase) |
| Extra staff at peak | Up to 9 additional people | None expected next peak |
| Overtime | Required to keep up | No longer required |
| Verification location | Outsourced overseas | Brought back in-house |
| Peak-period staffing cost | Baseline | More than $85,000 saved in 3 months |
| Process visibility | Limited | Full view of the process, including supplier submission errors |
For the first time in roughly five years, the interns hired to help with the peak were sent home early. The rest of the team stopped working overtime to keep pace. Josh Cohen expects no additional staff will be needed for the next peak at all, which accounts for the $85,000-plus in staffing costs avoided across a single three-month period.
Accuracy and visibility improved alongside throughput
Removing months of manual keying removed the errors that came with it. The team describes accuracy and efficiency as better than they have ever been. Automation also produced something the manual process never did: a complete view of the pipeline. That visibility surfaced supplier submission errors the company had not previously been able to see, which management can now monitor and address directly.
“Zenphi has a great UI that allowed us to customize workflows to fit our specific business needs. After we implemented their system, we not only saved time, we were able to reduce our workload significantly. Previously we were forced to outsource our Invoice verification process overseas, but with Zenphi, we were not only able to bring it back in-house, we reduced our costs and decreased our processing time significantly.”
Where Tavezio is taking automation next
With the peak invoicing period behind them, the team has five further projects lined up. The pattern Josh Cohen describes is a common one: a discussion about a problem in another part of the business turns into a workflow that solves it.
Where this pattern transfers
Invoice processing is a good first automation candidate because the documents are repetitive, the fields needed are consistent, and the verification rules already exist in someone’s head or spreadsheet. The same structure — extract, verify against your own rules, file, escalate the exceptions — applies to purchase orders, delivery notes, claims, applications, and contract intake.
The relevant reading if you are scoping something similar: AI automation for procurement covers the wider procurement picture, and AI-powered document workflow automation in Google Workspace explains how document validation removes manual workload from a team.
Questions about this case study
What results did Tavezio get from automating invoice processing?
Tavezio increased daily invoice processing capacity by a reported 600%, handling every invoice the same day it is received rather than clearing a months-long backlog. The company saved more than $85,000 in staffing costs during a single three-month peak period, eliminated the need for up to nine temporary staff and for overtime, achieved a 90% cost reduction, and brought invoice verification back in-house after previously outsourcing it overseas.
What was Tavezio’s invoice process before automation?
Every invoice was processed manually. Staff read each document, entered the data, verified it, and filed it. During the summer camp season, thousands of accumulated invoices took months to work through, requiring up to nine additional team members, overtime from existing staff, and overseas outsourcing of the verification step. Accuracy suffered under deadline pressure.
How does automated invoice processing work in this case?
The workflow runs five steps. Incoming invoice PDFs are renamed to a consistent convention on arrival. Key fields are extracted from each document into structured data. That data is routed through Tavezio’s own verification logic. Verified information is exported to CSV and the source document is filed in Dropbox. Anything the workflow cannot resolve is sent to Trello as an exception for a person to review, so human attention goes only to cases that need judgment.
Which systems did Tavezio connect?
The workflow runs on Zenphi and connects Google Workspace, Dropbox for document storage, and Trello for the exception queue. Extracted data is also exported to CSV for use in the team’s downstream processes.
Is invoice processing automation a good fit for high-volume seasonal businesses?
It suits them particularly well, because the cost of manual processing in a seasonal business is concentrated into short periods and is usually paid in temporary staff and overtime. Automating the repetitive extraction and filing work means peak volume no longer translates directly into headcount. Tavezio’s peak went from requiring up to nine extra people to requiring none.
What kinds of documents can this approach handle beyond invoices?
The same extract-verify-file-escalate structure applies to purchase orders, delivery notes, claims, applications, contracts, and other recurring document types where the required fields are consistent and verification rules can be stated. Tavezio is extending the approach to vendor and member onboarding and to inbound email handling in its confirmations department.
Have a document process that grows with headcount?
Bring one example — invoices, purchase orders, applications, or claims. The Zenphi team can map which steps a workflow can handle and where people should stay in the loop.
Schedule a demo