I've had a few people ask me whether Think Beyond is becoming an EHR — or whether it could eventually replace theirs. Up until now, my answer has basically been no.
One of the biggest reasons is that I have zero desire for Think Beyond Practice to become the company sitting on a giant centralized database containing everyone's patient records. That's a level of responsibility, risk, and liability I have intentionally not wanted.
But I've been thinking about the problem differently.
This post is open to everyone — including free members. Free members can read it and comment on it, and I genuinely want your input too.
Before cloud EHRs, clinicians and practices were the custodians of their records. The software company wasn't necessarily the place where everyone's records lived.
Technology could potentially let us bring some version of that model back without asking clinicians to become their own IT department.
What I'm envisioning is something like this:
Your practice controls the longitudinal patient record.
Think Beyond would help set up the secure record infrastructure for you — encryption, backups, recovery, access controls, audit history, record integrity, export, etc.
Then Think Beyond AI becomes the clinical workspace that works with those records.
There is another really important reason I'm interested in this model:
Right now, your EHR is often a wall between you and better AI.
Most clinicians probably never have to think about this because you log into your EHR and can see the whole chart.
But an outside tool like Think Beyond usually cannot.
Your EHR may let another company access certain pieces of information, or none at all, but it generally does not just say:
Here is this patient's entire chart. Use whatever information the clinician has access to.
That creates a major limitation for clinical AI.
You can have years of valuable information sitting in the chart — prior diagnoses, medication trials, doses, side effects, screeners, outside records, old assessments, treatment plans, labs, and previous clinical reasoning — while the AI can see only a small piece of it.
So even a very capable AI can end up working with an incomplete picture.
That's one reason so many AI scribes are still primarily focused on what happens during today's visit. They may know what was said today, whatever you manually pasted in, whatever you uploaded, or whatever limited information the EHR allows them to retrieve.
But that's very different from having access to the patient's actual longitudinal record.
That's a big part of what I'm trying to solve.
If the practice controls the record and explicitly authorizes Think Beyond to work with it, Think Beyond could potentially use the same longitudinal clinical information you use.
Not because Think Beyond owns the record.
Because you control it and are allowing your clinical workspace to work with it.
Once the AI can work from the longitudinal record instead of one isolated encounter, it could:
And honestly, that list could keep going.
The point is that the AI stops functioning like a scribe attached to one encounter and starts functioning more like a clinical workspace that actually knows the patient's history and helps you work with it over time.
To me, that's a much more interesting use of AI in psychiatry.
I don't think Think Beyond necessarily needs to reinvent infrastructure that other companies already do well.
A possible setup could look something like:
That puts the rough working estimate somewhere around $330–$350/month for the core stack.
Payment processing would still be transaction-based, and things like telehealth, fax, labs, etc. could either integrate with existing services or be added separately depending on what someone actually needs.
These are working estimates, not pricing I'm announcing. I'm trying to figure out whether the model itself is interesting before getting too far down the road.
The part I find compelling is that this could potentially give you something that functions like a full clinical EHR/workspace without requiring Think Beyond to own your patient records and without locking your records inside another company's system.
And there is another benefit to the practice actually controlling its own records:
If you ever decide you don't like Think Beyond, you should be able to leave.
Your records don't disappear. They don't belong to us. You shouldn't have to beg a software company for a giant export, worry about what format you're going to get back, or stay with a product you no longer want simply because moving years of patient records feels impossible.
You would still have your records.
If another EHR came along that you liked better, you could move to it.
If another AI company eventually built something better than Think Beyond, you could use that.
If you wanted to change billing systems, prescribing systems, or other parts of the technology stack, you wouldn't necessarily have to uproot the entire clinical record just to do it.
The record stays with the practice. The software around it can change.
To me, that's a much healthier model than having your entire practice effectively trapped inside whichever EHR happens to be holding the data.
You could potentially:
No forced migration. No requirement that everybody practice the same way.
And importantly, no requirement that you stay with Think Beyond forever just because your patient records are there.
Before I spend a ridiculous amount of time building toward this, I'm genuinely curious:
Would this actually appeal to you?
Would having more control over your patient records be a meaningful benefit?
Would the ability for Think Beyond AI to actually work with the patient's full longitudinal record, rather than whatever fragments another EHR allows it to see, matter to you?
Would you trust a model where Think Beyond handled the setup, security, backups, recovery, etc., but your practice remained the custodian of the records?
Would ~$350/month for the overall core stack feel reasonable if it could replace most or all of what you're currently paying an EHR for?
Or do you actually prefer the simplicity of having an EHR company host and manage everything?
And maybe most importantly:
What would have to be true for you to seriously consider leaving your current EHR for something like this?
I'm not announcing that I'm building it. I'm trying to decide whether this solves a problem clinicians actually care about.
So tell me where I'm right, where I'm wrong, what I'm not thinking about, and what would make you immediately say absolutely not.