You got a PDF. Nobody explained it. Here's what AI actually does with that information — and what it leaves to your doctor.
You had blood drawn. A few days later, a PDF appeared in your patient portal — or arrived by mail. You opened it and saw a table of numbers, cryptic abbreviations, and reference ranges you've never been taught to read. Some values have an asterisk next to them. Some say "H" or "L." A few are flagged in red.
Your doctor's office sent a message that said "results reviewed, no action required." But you still have the PDF open and you still don't know what your creatinine level means or why your MCV is slightly low or what the difference between LDL and VLDL actually is.
This is an extremely common experience. It's also one of the gaps that AI is genuinely well-positioned to help with — alongside some important limitations that are equally important to understand. This post explains both honestly.
Lab reports are generated by clinical laboratory software built for clinicians, not patients. The format prioritizes completeness and clinical utility — not readability for someone with no medical training. Every value is listed with a reference range, but reference ranges are population averages. Your "normal" might be at the high end of the reference range — or your clinically relevant change might be within the range but still meaningful relative to your personal baseline from six months ago.
The abbreviations compound this. CBC stands for complete blood count. Within a CBC, you'll find WBC (white blood cell count), RBC (red blood cell count), HGB (hemoglobin), HCT (hematocrit), MCV (mean corpuscular volume), MCH (mean corpuscular hemoglobin), MCHC (mean corpuscular hemoglobin concentration), RDW (red cell distribution width), and platelet count. These are not intuitive. A patient reading "MCV: 74.2 fL (L)" gets very little from that without context.
Beyond the abbreviations, there's the context problem. A single blood draw is a snapshot, not a story. An elevated white blood cell count means something very different in a patient with a known infection than in someone with no current illness. A thyroid-stimulating hormone result at the top of the reference range means something different for a patient already on thyroid medication than for someone whose thyroid has never been tested before. The numbers alone do not tell you which situation you're in.
When a doctor's office sends a "results reviewed, no action required" message, it typically means nothing is critically out of range and no immediate intervention is needed. It does not mean all your results are optimal, that trends are stable, or that there's nothing worth discussing at your next appointment. Patients who receive this message often close the PDF without ever understanding what it actually said.
Most patients receive their lab results through one of three channels: a patient portal notification, a follow-up call from the doctor's office, or the results arriving by mail. In all three cases, the format is essentially the same — a table of values, reference ranges, and flags.
The follow-up conversation, if there is one, is brief. Doctors are under time pressure. A nurse calling with results will communicate the clinically actionable items: your A1C is too high, your cholesterol needs attention, your kidney function is in range. What doesn't get communicated is the broader context — why these particular values matter, what organs or systems they reflect, how they've changed since your last test, or what questions you should be asking.
Patients who want to understand their own results have historically had limited options. They could search the internet, where they'd find information of wildly varying quality that often defaults to alarming interpretations. They could wait until their next appointment and hope to have time to ask questions. Or they could simply not understand and file the results away.
AI changes this — partially. The partial part matters as much as the change.
AI is genuinely useful for the translation layer. If you upload a lab report and ask what "elevated ALT" means, a well-designed AI health tool can tell you that ALT (alanine aminotransferase) is an enzyme primarily produced by the liver, that elevated levels can indicate the liver is under stress, and that common causes include certain medications, alcohol, fatty liver disease, or recent intense exercise. It can explain this in plain language, without abbreviations, in a way that helps you understand what you're looking at.
This is meaningful. For a patient who has never learned what a lipid panel actually measures, having an AI explain that LDL is the lipoprotein that carries cholesterol toward artery walls, HDL is the lipoprotein that carries it away, and that the ratio between them matters as much as either number in isolation — that's genuinely useful context for a conversation with their doctor.
AI can explain, in plain language, what biological process or body system each value in a lab report reflects. A CBC tells you about the composition of your blood — whether you have enough red cells, white cells, and platelets, and whether the cells are sized and shaped normally. A comprehensive metabolic panel (CMP) tells you about electrolyte balance, kidney function, liver enzyme levels, blood glucose, and blood protein levels. A lipid panel measures the distribution of cholesterol-carrying molecules. An A1C reflects average blood glucose over the preceding three months.
Understanding what a test measures is the prerequisite for understanding what an abnormal result means. AI is well-suited to provide this layer of explanation clearly and consistently.
A single data point is a snapshot. Two or more data points over time reveal a trajectory. If your LDL was 110 mg/dL six months ago and is 138 mg/dL today, that's a trend worth noting — even if both values fall within the "borderline" range rather than the "high" range. AI that has access to multiple lab results over time can flag when values are moving in a concerning direction, even if they haven't crossed a clinical threshold yet.
This longitudinal pattern recognition is something that gets missed in typical clinical workflows. A lab result reviewed in isolation, with no reference to the prior result, loses important context. AI that reads your uploaded lab documents over months and years can surface these trends in a way that a single-appointment review cannot.
This is where tools like Cureva's AI health brief go beyond simple explanation. If you've been tracking your blood pressure readings and your medication schedule alongside your lab results, an AI that has access to all three data streams can identify relationships that would otherwise require your doctor to manually correlate across disparate records.
Eva reads uploaded lab report documents and connects what they show to your medication history and vitals logged in Cureva. If your creatinine has been trending upward over three tests and you've also been consistently logging high blood pressure readings, Eva surfaces that pattern as something worth raising with your doctor. She explains what each value means, what direction it's moved, and what context your other health data adds — then recommends bringing the full picture to your next appointment.
This is longitudinal health intelligence — the kind of picture that used to require a proactive physician, years of retained medical records, and a long appointment to review. AI doesn't replace that physician review, but it brings the connected picture to the surface so the appointment conversation can be more productive.
The capability description above comes with an equally important boundary. AI can explain. AI can identify patterns. AI can surface connections across data. AI cannot diagnose.
Diagnosis requires clinical judgment that accounts for factors no AI health app has access to: physical examination findings, clinical history, the full context of a patient's presenting symptoms, the physician's pattern recognition from years of practice, and the ability to order and interpret additional tests in real time. An AI that tells a patient "your elevated ALT may indicate liver disease" without knowing that the patient went to the gym six times last week and took ibuprofen for a muscle injury is providing information that sounds ominous and is likely irrelevant.
Explanation: "Your MCV is slightly below the reference range, which means your red blood cells are smaller than average. This pattern is sometimes associated with iron deficiency or certain vitamin deficiencies, and your doctor may want to check a few additional values to understand the cause."
Diagnosis: "You have iron deficiency anemia."
The first is appropriate for an AI health tool. The second is not — it requires a clinical assessment and often additional testing to confirm. The distinction is not academic. Acting on an AI diagnosis without clinical confirmation can lead to unnecessary treatment or missed alternative causes.
AI health tools should never tell you to start, stop, or change a medication based on a lab result. They should never tell you that you have a specific condition. They should never discourage you from seeking medical care by implying that the situation is resolved or doesn't warrant attention. The appropriate output is explanation, pattern identification, and a prompt to discuss with your healthcare provider.
The goal of AI-assisted lab result interpretation is not to replace the clinical conversation — it's to make that conversation more productive. A patient who arrives at their appointment knowing what their CBC measures, what their elevated neutrophil count could indicate, and what their trend across three blood draws looks like is a patient who can ask better questions, understand the answers, and participate in their own care.
That is the gap AI is suited to fill. Not the clinical decision, but the comprehension that makes the clinical decision meaningful to the person whose health it concerns.
Cureva's AI health brief brings together lab uploads, medication history, and vitals to create the kind of longitudinal picture that gives both patients and their care teams a more complete view. It explains what the data means, flags patterns worth discussing, and always routes clinical questions back to the physician who has the full clinical context to answer them. Beta opens September 15, 2026, with founding pricing at $5/month for your first year.
Medical Disclaimer: This article is for informational and educational purposes only. It does not constitute medical advice, diagnosis, or treatment. The description of what AI can do with lab results refers to educational explanation and pattern identification, not clinical interpretation or diagnosis. Always consult a qualified healthcare professional regarding your lab results and health conditions. Cureva is a health tracking and AI companion tool — not a medical device — and does not provide medical advice, diagnosis, or treatment recommendations.
Eva reads your uploaded lab reports, connects them to your medication history and vitals, and explains what the data means — in plain language, over time.
See Pricing →