Building TESSA: From Service Records to Biomedical Intelligence

TESSA stands for Technical Equipment Support and Service Assistant. That name has always described something larger than an equipment database.

The original concept was an AI-powered troubleshooting assistant that PBES could put directly into the hands of its customers. If a piece of medical equipment was giving a facility trouble, someone could explain the problem to TESSA and receive useful technical guidance immediately. In its most ambitious form, it would feel less like searching through a service portal and more like having access to a biomed who already understood your facility.

We tried to start there. And that exposed the real problem almost immediately.

Artificial intelligence can know a great deal about a patient monitor, defibrillator or autoclave in general. But useful biomedical support often depends on knowing much more than the equipment category or model number. Which exact device is this, and where is it located? What has happened to it before, and when was it last inspected? Has this symptom appeared previously — and if so, what did the technician find, and what actually solved the problem? Has the equipment moved? What else is commonly used around it?

Without that context, even a very capable AI is still largely starting from scratch. We had tried to build the destination before building the road to get there.

So TESSA changed — not because the original idea went away, but because we realized what had to exist first.

The work became the data layer

PBES was already generating the information TESSA needed. It was happening every day in the field. A technician identifies a device, confirms its serial number, performs an inspection, records a failure, photographs something unusual, writes a note, completes a repair and eventually closes the problem. Then PBES returns the following year and does it again.

Historically, information like this is easy to fragment. Some of it lives in a database. Some is buried in PDFs. Some sits inside service reports or email threads. Some exists only because the technician who worked on the equipment remembers what happened.

TESSA is being built to turn that stream of work into structured equipment history. The important distinction is that we are not asking customers to maintain another database simply to create useful data. The dataset grows as a natural byproduct of the electrical safety inspections, preventive maintenance and repairs PBES is already performing.

An inspection updates the asset. A repair expands its history. A photograph adds visual context, a technician note preserves knowledge, and a location update makes the facility map more accurate. The next service event builds on everything that came before it.

This is what TESSA is today: the operational layer connecting PBES service work to a continuously improving equipment record.

And that foundation is what makes the original idea increasingly possible.

An equipment list tells you what a facility owns. Context tells you how that equipment actually lives inside the facility.

As TESSA becomes more complete, the goal is to understand the relationship between the organization, its locations, its departments, its rooms and the equipment operating inside them. That means the record for a device is no longer isolated. It has a physical location and neighbors. It has photographs, a maintenance schedule, service events, open issues and a lifecycle.

Over several inspection cycles, that creates something closer to a living digital representation of the equipment environment than a traditional static inventory. And importantly, that history compounds. The second year should know what happened during the first. The fifth year should carry five years of operational context into every future decision.

The resolution is as important as the failure

One of the most valuable datasets TESSA can build is not simply a record of equipment failures. It is a record of what happened next.

Biomedical technicians solve small and large problems constantly. They observe symptoms, investigate possible causes, test components, change settings, replace accessories, install parts and determine whether equipment should remain in service. A traditional report documents that a repair happened. A useful intelligence system should begin connecting the entire sequence: what was reported, what was observed, what was tested, what was found, and what resolved it.

Over enough service events, those sequences begin forming patterns. A particular model may repeatedly develop the same issue. A certain accessory may fail more often than expected. One symptom may repeatedly lead technicians toward the same component. A device with an increasingly frequent repair history may be approaching the point where replacement makes more sense than another service call.

That is where accumulated service history starts becoming more than recordkeeping. It becomes knowledge.

Giving the technician the memory of the entire organization

This same information can change what it means for a technician to walk into a facility for the first time.

Today, experienced technicians develop an enormous amount of knowledge simply by being there year after year. They remember which device caused trouble. They remember the strange configuration in one room. They know where equipment tends to disappear, what was fixed last time and what the customer was concerned about. That experience is valuable, but experience locked inside one person's memory is difficult to scale.

TESSA can become the institutional memory around the equipment. The goal is that a PBES technician who has never personally visited an account can arrive with years of accumulated context already available. Before beginning work, TESSA could eventually brief the technician on the facility: its inventory, room locations, service history, unresolved items, equipment with recurring failures and notes left by previous technicians.

Instead of replacing the technician's expertise, TESSA gives that technician a much larger memory to work from. And once the visit is complete, the new technician adds their experience back into the same system. Knowledge stops leaving with the person who learned it.

The final form: a biomed that already knows your facility

Everything above leads back to the original idea. The long-term vision for TESSA is straightforward: when a PBES customer has a question, concern or equipment problem, TESSA should already understand enough about their environment to help — not because it knows every possible answer, but because it knows the customer.

Imagine a staff member opening their TESSA portal and saying, "The monitor in Exam Room 4 isn't taking blood pressures." TESSA would not have to begin by asking which monitor they own. The facility record could already establish which monitor is located in that room, its manufacturer and model, when it was last inspected, previous service events, photographs, associated accessories and whether a similar issue has occurred before.

From there, AI becomes significantly more useful. The customer could eventually speak naturally to TESSA while using their phone's camera to show the equipment. Visual AI could help identify what the person is looking at, understand visible conditions and guide appropriate troubleshooting. A simple issue might be resolved without waiting for a service visit. A problem requiring a technician could arrive at PBES with far better information already attached — and when a technician does respond, the full interaction becomes another part of the equipment's history.

That is much closer to the meaning behind the name.

The same visual intelligence could eventually assist the PBES technician during inspections. A camera can become another input alongside measurements, technician observations and equipment history. Over time, AI-assisted inspection workflows could help identify equipment, document visible conditions, compare current equipment to previous photographs and guide technicians through standardized procedures — for applicable equipment and scopes of work, including workflows built around manufacturer procedures, facility requirements and relevant standards such as NFPA 99.

The technician remains responsible for the inspection, the testing and the technical judgment. The opportunity is to give that technician better context, better documentation and another intelligent tool alongside their existing training and test equipment.

Why PBES can build this from the field

There are plenty of ways to build software about medical equipment. Our opportunity is that TESSA is being built inside the service operation itself. PBES has direct access to the equipment, the technicians servicing it and the facilities using it. That creates a feedback loop that would be difficult to reproduce from software alone.

We do not need to invent scenarios simply to give TESSA more context; the normal operation of the business creates it. A new inspection adds another equipment record. A return visit adds history. A failure adds a problem, and a repair adds a resolution. A customer question exposes something the system should make easier, and a technician request exposes something the workflow should capture better.

Because many PBES customers are serviced repeatedly over multiple years, the dataset can become longitudinal rather than transactional. The system sees what changes, what repeats and how equipment moves through its lifecycle.

Our customer relationships are equally important. TESSA is not being designed in isolation and handed to healthcare facilities after it is finished. The facilities using PBES help expose the problems worth solving, while technicians continuously pressure-test whether the software reflects how biomedical work actually happens. That feedback loop is how we want to build the product.

Building toward the name

TESSA is not at its final form today. That is important to say clearly.

Today, TESSA is helping PBES organize equipment, inspections, service history, maintenance schedules, findings, photos, documentation and follow-up into a much more connected operating system. That work provides immediate value by giving PBES and its customers better equipment visibility and better records. But it also serves another purpose. It is building memory.

With enough high-quality history, the questions TESSA can help answer become more sophisticated. Not only "When is this device due?" but "What happened the last time this device had this problem?" — "What problems appear repeatedly on this model?" — "How have our technicians solved this issue before?" — "What equipment is consuming disproportionate service resources?" — "What should the technician know before walking into this facility?" And eventually: "Can you look at this equipment with me and help me figure out what is wrong?"

That progression is the larger TESSA project. The goal is not to bolt an AI chatbot onto an equipment database. The goal is to build enough structured understanding of the equipment, the facility, its history and the service work around it that AI can become genuinely useful.

We originally tried to build the answer first. What we are building now is the knowledge required to answer the question correctly. And with every inspection, repair, photograph, technician note, customer interaction and year of equipment history, TESSA gets another piece of the context it will need to eventually live up to its name: Technical Equipment Support and Service Assistant.