01
Request for support · $3.15M over three years

Intervention engine built on truth.

PathwayStudio designs and re-routes stepped-care mental health treatment from validated session-by-session measures, with every recommendation traceable to graded clinical evidence. 


Built on Divine Mercy University's Catholic Christian Meta-Model of the Person, an integrated psychological, philosophical and theological account of what a human being is. Every pathway and every outcome measure orients around it.

Prepared by Thomas Cunningham · Divine Mercy University
Date August 2026

02
FIT Integration

A therapist measures the patient every session. 

Feedback-informed treatment is well evidenced and taught at DMU. Give patients a four-item outcome measure at each session, watch the trajectory, and dropout falls while outcomes improve. Thousands of clinicians already do this.

The measurement works. The score lands on a dashboard, the clinician reads it, pathway studio helps walk the therapist through potential interventions.

PCOMS + Pathway Studio

The strongest predictor of dropout in the feedback-informed treatment literature is a declining outcome score alongside a damaged therapeutic alliance. Both are measurable in ninety seconds. 

What the plan should do

  • Reroute when the score stalls
  • Repair the alliance first when that is the actual problem
  • Escalate on risk without depending on the clinician noticing
  • Step care down when someone is well, and free the slot
03
What we built

A care pathway that executes, rather than a document that describes.

A clinician designs a pathway on a visual canvas — assessments, interventions, referrals, spiritual practices, journaling, scheduling — and connects them with conditions. When a patient's score crosses a threshold, the pathway routes them down a different branch.

At each decision point the engine retrieves matching evidence from a curated corpus of clinical guidelines and proposes interventions, each carrying its source and a grade. The clinician decides. The engine records what was suggested, what was chosen, and what happened to the score afterwards.

That last sentence is the part that compounds. Every session adds to a record of which interventions moved which patients, under which conditions.

Humble beginnings.

  • 1,630 passages of clinical evidence from APA, SAMHSA, NICE and VA/DoD guidelines, each tagged by disorder and graded A, B or C
  • Retrieval and synthesis end to end — a score goes in, cited intervention suggestions come out in seconds
  • Visual pathway editor with fourteen activity types, conditional routing and version history
  • Server-side scoring for the ORS and SRS outcome and alliance measures
  • An audit trail capturing every suggestion, the evidence behind it, and the model that produced it
04
The market · comparison

The category is proven. The quadrant we occupy is empty.

Where we are
Blueprintand AI scribes generally
PathwayStudiograded evidence, score-driven
Documents care already given  →  Decides care not yet given
Model's own words  →  Cited, graded evidence

Blueprint has validated the two-part model we are proposing: pair a clinician tool with an intelligence layer, then license that layer to other systems. Over 50,000 therapists use it, and it runs the documentation infrastructure behind national provider networks, where it has cut charting time by more than 60%.

It is a strong product aimed at a different problem. Blueprint writes the note after the session from a recording of what was said. Its suggestions arrive inside the plan field of that note, in the model's own words, without a citation.

We can beat them to the goal. The entire industry wants to move to helping the therapist plan out interventions.

05
The comparison in detail

Same word, different machine.

Both products say "intervention suggestion." Below is what each one actually means by it. The final row is the one where they win.

BlueprintPathwayStudio
Input A recording of what was said in the session The patient's outcome and alliance scores, plus presenting problem
Evidence No citation
The model's clinical knowledge, unattributed
Cited and graded
Passage-level source and A/B/C strength from published guidelines
Time direction Backward — one note per session, after the fact Forward — a sequence that branches on the next score
Response to a stalling patient None
The plan is a static file; updating it is one of their most-requested unmet features
Reroutes
A declining score sends the patient down a different branch, or triggers alliance repair first
Model of the person None CCMMP
Toggleable off for secular deployments
Clinical documentation Their core strength
SOAP, DAP, BIRP and GIRP notes in seconds
On the roadmap
Funded work: notes whose plan section carries the citation and grade
Verified against blueprint.ai product material and an independent product review, August 2026.
06
Why it holds

Two things a competitor cannot copy by shipping a feature.

A model of the human person

PathwayStudio is built on Divine Mercy University's Catholic Christian Meta-Model of the Person, an integrated psychological, philosophical and theological account of what a human being is. Every pathway and every outcome measure orients around it.

This is not a content library that could be licensed away from us. It is the frame that decides what counts as improvement in the first place. A secular platform has an inferior tradition to measure flourishing against, so it measures symptom reduction and stops there.

The engine toggles cleanly to a secular configuration, so the same technology serves mainstream partners without dilution.

An outcomes record that only exists if you built it early

Because the engine logs what it suggested, what the clinician used, and what happened to the score afterwards, it accumulates something no amount of funding buys later: evidence about which interventions actually moved which patients.

A competitor starting this in three years starts their dataset in three years. The interventions in our library will by then carry a signal drawn from real trajectories rather than from guideline consensus alone.

This is also what makes the work publishable, and what makes the resulting pathway library useful to people who will never be our customers.

07
Where the project actually stands

The engine runs. It is not yet safe for patients, and that is what the money is for.

A full technical review of the codebase in August found what an honest prototype review usually finds. We would rather put it in front of you now than have you find it in diligence.

Real and demonstrable

  • Evidence retrieval, synthesis and citation, working end to end
  • 1,630 graded passages across depression, PTSD, anxiety, bipolar and substance use
  • Visual pathway design with conditional routing and immutable published versions
  • Validated scoring, computed on the server and never trusted from the client

Not yet built

  • Authentication and database-level access control — the prototype has neither, so no real patient data has ever entered it
  • Half the required diagnostic coverage: obsessive-compulsive, eating, personality and psychotic disorders are unsourced
  • Automated testing, which a validation study will require
  • The partner integration layer that makes the engine licensable at all

None of these are research risks. They are known engineering with known solutions, scoped into a twelve-week plan. The unresolved question in this project is clinical, not technical: which sequenced pathways actually help people. That is the study.

08
The study

Build the pathways, run them with real clinicians, keep what works.

The grant funds a multi-year validation study. We assemble the best available intervention evidence, implement it as stepped-care pathways grounded in the CCMMP, run those pathways at clinical partner sites, and measure outcomes for every person served.

Pathways that help are kept and refined. Pathways that don't are changed. After three years the output is an evidence-tested library of whole-person care pathways, with the trajectory data behind each one.

We are also publishing it. A library of validated pathways is useful to clinicians and institutions who will never license our software, and it does not exist anywhere today.

What gets measured

  • Outcome and alliance trajectories for every enrolled patient, session by session
  • Which suggested interventions clinicians used, and which they declined
  • Score movement following each intervention, linked back to the pathway branch
  • Where pathways failed — dropout points, stalls, escalations

The engineering that makes this measurable is in the first quarter of work. Get the data model wrong and three years of study data cannot be analyzed — so it is being fixed before any patient is enrolled.

09
Sustainability

Licensing revenue pays for the research after the grant ends.

Electronic medical record and practice-management companies want to offer their clinicians evidence-based intervention support. Most will not build it. Blueprint has already demonstrated that they will license it instead, and that an API-first intelligence layer can reach tens of thousands of clinicians without owning the clinician relationship.

PathwayStudio is built to embed in any system rather than to lock buyers into ours. Partner revenue funds the continuing study, which improves the pathway library, which makes the engine more valuable to the next partner.

Philanthropic money seeds an asset that pays for its own future. That is the case for funding it now rather than subsidizing it indefinitely.

Who buys

Behavioral health EMRs, practice-management platforms, provider networks, and faith-based counselling institutions that want the CCMMP configuration specifically.

Why they buy rather than build

The corpus, the grading, the clinical logic and the outcomes evidence take years to assemble. Licensing an engine that already carries them is faster and cheaper than reproducing them, and it comes with citations their compliance teams can audit.

10
The ask

$3.15M over three years, in three components.

We welcome support directed at whichever component is closest to your priorities.

One

Productize the engine

Turn a working prototype into a standalone, licensable module: authentication, database-level access control, an attributable outcomes record, automated testing, and the diagnostic coverage the corpus is still missing.  A versioned, authenticated, documented API so third-party medical record systems can embed the engine — plus clinical notes whose plan section carries the citation and evidence grade, which no competitor currently offers.

Two

Incorporate the research and best practices into the intervention engine

Beyond PCOMS and FIT common factors which is applied system wide, we then incorporate the many disciplines of counseling, psychotherapy, spiritual direction, and accompaniment into the suggested intervention pathways.

Three

Run the validation study

Implement CCMMP-grounded pathways with clinical partner sites, measure outcomes for every person served, refine what works, and publish the resulting library.

11
In short

The measurement already works. We are building the part that acts on it.

1,630
graded evidence passages already retrievable, with citations
4
national guideline bodies standing behind every recommendation — APA, SAMHSA, NICE and VA/DoD, cited passage by passage
3 yrs
to an evidence-tested pathway library that outlives the grant and funds its own continuation

A therapist who learns on Tuesday that their patient is getting worse should not have to wait for a quarterly review to change the plan. Everything here follows from that.