The Core Problem With Remote-Only Certification
I have been training service dogs and coaching handler teams for fifteen years. In that time I have watched online-only training certification programs multiply at a pace that genuinely alarms me. The pitch is always the same: upload videos of your dog performing tasks, receive feedback from a remote coach, and earn a certificate confirming your team is public-access ready. It sounds efficient. It is not.
The fundamental problem is not about technology. High-definition video, real-time streaming and asynchronous review have all improved dramatically. The problem is that the variables that determine whether a handler team is safe and reliable in a public access environment cannot be transmitted through a camera lens. They can only be observed in physical space, by a trained evaluator standing close enough to catch what the handler does not know they are doing.
This matters because the handler team is the unit. The dog alone is not the unit. The certification question is never just "can this dog perform these behaviors?" The real question is "does this handler reliably elicit these behaviors under realistic conditions, across novel environments, while managing their own stress, movement patterns and timing?" That question cannot be answered remotely.
At TheraPetic® Healthcare Provider Group, our clinical and training staff work alongside handler teams in real environments precisely because we understand how much the physical context reveals. What I describe below is not theoretical. It reflects fifteen years of in-person observation that remote programs structurally cannot replicate.
Environmental Neutrality Cannot Be Tested Through Video
Environmental neutrality is the dog's ability to remain behaviorally stable across unpredictable, novel stimuli without handler cuing or compensation. It is one of the most critical public access competencies and it is the one most catastrophically misrepresented in online certification models.
Here is what video submission gets wrong. The handler chooses the filming location. They choose the time of day. They choose how many passes they attempt before submitting a clean take. They control what appears in frame. I have reviewed video submissions from remote programs where a dog performed a beautiful "ignore" of a passing shopping cart in a parking lot. What the evaluator could not see was that the handler had parked at 7 AM when foot traffic was minimal, filmed twelve attempts and submitted the one where nothing triggering happened to walk by. That dog had never been proofed in a crowded, unpredictable retail environment on a Saturday afternoon. That dog was not public-access neutral. That dog was video-neutral.
Genuine environmental proofing requires an evaluator to introduce stimuli the handler cannot anticipate and cannot script. I need to walk a team into a hardware store at noon, position myself where I can watch the dog's ear set, tail carriage and scanning behavior across fifteen minutes of uncontrolled stimulus exposure. I need to be present when a forklift moves unexpectedly in the aisle, when a child sprints past, when another dog barks from the garden center. No handler can stage that. No video can replicate it on demand.
The Public Access Test administered by credentialed in-person evaluators exists precisely because this kind of field assessment is non-negotiable. Organizations like the International Association of Canine Professionals (IACP), through which my own CSDT credential was issued, require in-person observation for the same reason. Environmental neutrality is a live-environment property. Certifying it from a submission reel is a category error.
Handler Mechanics Cannot Be Corrected From a Screen
Handler mechanics are the second area where remote-only programs fail systematically and where I have seen the most harm done to working teams.
Leash mechanics are the obvious example. Chronic leash tension is one of the most common handler errors I correct in person and one of the hardest to self-diagnose. A handler who consistently carries tension in their shoulder translates that tension through the leash into micro-corrections the dog receives constantly, without any conscious cuing intent from the handler. The dog learns that the default state is mild pressure and begins orienting to the handler's body rather than working independently. The result looks like an attentive dog. It functions like a handler-dependent dog who cannot generalize to a different leash handler or a moment of distraction.
I can watch a 90-second submission video and miss this entirely. The clip is too short. The handler knows they are being filmed and adjusts their posture. The camera angle shows the dog beautifully but crops the handler's elbow. I have caught this pattern in person by watching a handler walk a circuit for four minutes while I stood behind them with a clear sightline to their grip hand and shoulder. That observation position does not exist in remote evaluation.
Timing errors are equally invisible through video. Reinforcement delivered 1.2 seconds after a behavior versus 0.3 seconds after a behavior looks identical on a phone screen but produces dramatically different learned associations in the dog. In person, I can watch a handler's treat delivery motion and give real-time correction before the misassociation is reinforced. Remote coaches working from video can comment on timing in writing. The handler then films another attempt with an improved-but-still-off timing window. The coach sees improvement. The dog continues to build a slightly blurred behavior chain. Nobody knows until the behavior breaks down in a high-distraction environment.
Spatial pressure, body blocking, handler footwork during task execution, the direction a handler faces during a crowd navigation exercise: all of these feed directly into how a dog reads the world around them. All of them require physical proximity to evaluate accurately. At Training Plus through officialservicedog.com, in-person session protocols are built around exactly these observable mechanics because we know remote observation produces incomplete data.
What Video Submission Models Actually Capture
I want to be precise here because I am not arguing that video captures nothing. A submission reel can confirm that a dog knows the command vocabulary. It can show approximate task performance in a controlled environment. It can reveal gross behavioral problems like obvious reactivity or severe avoidance. A skilled coach reviewing video can identify training gaps that a novice owner would not recognize themselves.
What video reliably captures is the dog's best behavior in a handler-controlled context. That is a useful data point. It is not a certification-quality assessment of a team's public access readiness.
The distinction matters enormously under current federal law. Under the Americans with Disabilities Act, a service dog must be trained to perform work or tasks directly related to the handler's disability and must be under control in public spaces. The ADA does not specify a certification standard, which is exactly why private organizations have developed public access evaluations. When those evaluations are remote-only, they certify video performance, not real-world behavior. A business owner or facility manager relying on that certificate to make an access determination is being given a document that reflects a narrower assessment than the term "certification" implies.
This is an ethics problem, not just a training quality problem. Families and individuals with disabilities invest significant time, money and emotional energy in getting their service dog team certified. They deserve to know what that certificate actually assessed and what it did not.
When Remote Coaching Is Legitimate
Remote coaching has a real and valuable role in service dog training. I use it. My training team uses it. The error is not using remote methods. The error is using remote methods as a terminal certification pathway rather than as a support modality within a larger in-person program.
Remote coaching is legitimate and genuinely useful in the following contexts.
- Between in-person sessions, when a handler needs feedback on a homework drill they are practicing at home and the behavior is simple enough that a wide camera angle captures the relevant data points.
- For task shaping in the early acquisition stages, before the behavior has been proofed to public-access standards, when the handler is working in a familiar controlled environment and the coach is supplementing in-person guidance rather than replacing it.
- For geographic accessibility where a handler lives in a rural area and in-person evaluation access is genuinely limited. In this case, remote coaching should bridge sessions to an in-person evaluator, not replace the in-person evaluation entirely.
- For reviewing recorded team interactions when both parties agree that the footage is exploratory rather than evaluative, meaning the coach is looking for discussion points rather than rendering a certification judgment.
- For continuing education, where a handler who has already been certified in person wants to maintain skills, refine task mechanics or learn new behaviors under guidance between annual evaluations.
The pattern that legitimizes remote coaching is always the same: it is a tool within a broader program that includes physical-space evaluation, not a replacement for it. Any program that offers a terminal public access certification without requiring in-person observation at some point in the process is not describing a different model of training. It is describing an incomplete one.
What the Industry Owes Handler Teams
The service dog training industry in the United States is not federally licensed. There is no government body that approves or revokes training organization credentials. This means the ethical floor is set entirely by professional associations, credentialing bodies and the individual integrity of trainers and program directors. That is a significant responsibility.
When I earned my CSDT through the IACP, the evaluation process required demonstrating in-person training competencies across multiple disciplines. That standard exists because the IACP understands that credentialing is a public trust instrument. A certificate represents a claim to the public, to families with disabilities, to facilities making access decisions, and to courts in the event of a dispute. That claim had better reflect what the evaluation actually measured.
What the industry owes handler teams is honest representation. If a program delivers remote-only coaching and issues a certificate, that certificate should say what it assessed: "This dog demonstrated task performance in handler-controlled video environments." It should not say or imply "This team has been evaluated for public access readiness" unless an in-person component was part of the evaluation.
Families seeking service dog support are often navigating disability, medical complexity and financial strain simultaneously. They are not always positioned to interrogate the methodology behind a certification program they find online. The responsibility to be transparent sits with the programs, not with the families.
I hold a CSDT credential. My organization is a 501(c)(3) nonprofit healthcare provider. The standard I apply to our training and certification work is the standard I would apply if every family we serve knew exactly what our evaluation captured and exactly what it did not. That standard requires boots on the ground. It requires a trained evaluator in the physical environment. It requires watching a handler team move through the world rather than watching them perform for a lens.
Online-only training certifications fail handler teams not because the trainers delivering them are necessarily incompetent but because the format structurally prevents the evaluation from reaching the evidence threshold that public access readiness actually requires. Families deserve better. The dogs deserve better. The standard should reflect that.