The Real Reason 110 Organizations Adopted AI Documentation in a Year

Most behavioral health leaders have lived through the same disappointment at least once. A new tool arrives with a strong demo and a compelling case. Leadership approves a pilot. A handful of clinicians try it, a few champions like it, and then, quietly, it fades. Six months later the contract is up for renewal and almost no one is using the thing. The technology was not the problem. The workflow was.

This pattern is so common that it has become the default expectation. When a clinical AI tool launches, the honest question is no longer “Does it work?” It is “Will anyone actually use it once the novelty wears off?” That question is what makes the first-year results of Videra Health’s Sidekick Notes integration with BestNotes worth paying attention to. Between May 2025 and May 2026, clinicians using the integration generated 167,899 notes and saved an estimated 47,571 hours, with 518 active users across 110 organizations. The time savings matter. But the adoption number is the more interesting story, because adoption is exactly where tools like this usually fall apart.

Why good tools die in the pilot

Behavioral health clinicians are among the most overextended professionals in healthcare. Caseloads are full, staffing is thin, and documentation consumes a meaningful share of the workday that should belong to patients. The American Medical Association has documented how much administrative and EHR work follows physicians home after hours, and the pattern is just as familiar to behavioral health providers charting on Sunday nights to catch up.

Ask a clinician in that situation to adopt a new tool and you are asking for something most of them cannot spare: time and attention. Every separate app is another login to remember, another window to toggle between, another system to learn during hours they do not have. Even genuinely useful technology becomes one more thing to manage. So clinicians do the rational thing. They revert to the workflow they already know, and the new tool quietly dies, not because it failed on the merits but because it never had a fair chance to become a habit.

This is the trap that catches most clinical AI. The model can be excellent and the value can be real, and none of it matters if the tool sits outside the daily flow of work. Capability gets the attention in demos. Friction decides what survives contact with a real caseload.

What changes when there is nothing to adopt

Sidekick Notes took a different path into the clinician’s day. Rather than arriving as a standalone product, it was built into the BestNotes EHR that behavioral health and addiction treatment providers already use. There is no second application to open and no separate system to log into. A clinician finishes a session and the documentation tool is already there, inside the workflow they were going to use anyway.

That design choice is what the first-year numbers actually reflect. Reaching 518 active users across 110 organizations in twelve months is not only evidence that the AI produces good notes. It is evidence that clinicians kept using it after the first week, because using it did not require them to change how they work. The tool met them where they were.

The note-taking itself is built for how behavioral health clinicians actually practice. They can capture a session through ambient listening or by dictating a summary aloud, and they can move between a narrative-focused progress update and a clinical, diagnostic-focused medical necessity review at any point, even after the session ends. If a clinician wants to add a detail the recording missed, a passing observation or a note about body language, they can dictate it afterward and the AI folds it into a regenerated note. The work fits the clinician’s judgment rather than forcing the clinician to fit the tool.

Time savings only count if clinicians show up

The 47,571 hours saved is a real and useful number. It reflects the time returned on each note, multiplied across nearly 170,000 of them. For a behavioral health workforce stretched thin, that adds up to meaningful relief.

But notice the dependency hiding inside that figure. The time saved on any given note only counts if the note is actually written in the tool. Time savings is not a property of the software sitting on a shelf. It is the product of adoption multiplied by efficiency. A tool that trims time on every note but gets opened twice saves nothing worth measuring. The reason the hours added up to 47,571 is that 518 clinicians reached for the tool again and again, and they reached for it because it was already in front of them. For many, it changed the shape of the week. Documentation that used to spill into Sunday afternoons now gets finished in the quiet few minutes after a session, while the conversation is still fresh and the next patient hasn’t walked in yet.

This is the connection leaders tend to miss when they evaluate clinical AI. The efficiency case and the adoption case are not separate. One produces the other. The most capable model in the category will not return a single hour to patient care if clinicians have to leave their workflow to use it.

What this means for behavioral health leaders

For anyone weighing AI documentation tools right now, the most important question to ask a vendor is not how accurate the notes are or how advanced the model is, though both matter. It is a quieter, more practical question: how much does this actually ask of the clinician? A new platform to learn, a new login, a migration, a separate system to manage day to day? Or close to nothing?

That distinction predicts almost everything downstream. It predicts whether adoption sticks after the pilot, whether the projected time savings materialize, and whether the investment produces returns a year later or becomes another line item no one can quite account for. Low friction wins, not because simplicity is elegant, but because every additional thing a clinician has to learn or manage is a place where adoption can break.

This is what the BestNotes integration gets right. Notes move straight into the chart inside the EHR clinicians already trust, so the lift on the clinician is close to zero. That is the deepest version of the experience, and the first-year results show what it produces at scale. The same principle guides how Sidekick Notes meets clinicians in any EHR they work in: the goal is always to fit the way clinicians already practice, not to ask them to practice differently.

The first year of Sidekick Notes in BestNotes is, in the end, a story about removing friction. The AI is built specifically for behavioral health, and it does its work without asking clinicians to learn a new system or change how they chart. Take away that lift, and what remains is a tool clinicians simply use. The hours returned to patient care follow from that.

For behavioral health organizations, the path to AI value is shorter than it looks. It runs through the way clinicians already work.