Behavioral Health Access in 2026: What Grow Therapy's Model Changes at Intake

TL;DR#

Roughly 47.9% of U.S. adults with any mental illness received no mental health treatment in 2024, and Grow Therapy's AI strategy — like the insurance-first business model underneath it — targets the administrative half of that gap rather than the clinical half. SAMHSA's 2024 National Survey on Drug Use and Health counted 61.5 million adults with any mental illness, of whom 52.1% received treatment. Among adults who perceived an unmet need, actively sought treatment, and still went without, 41.9% said they could not find a program or professional they wanted to go to and 26.1% found no openings — barriers of routing and availability, not of clinical skill. Grow Therapy reported 26,000 providers and 125+ health plan partners at its March 2026 Series D, up from 12,000 providers across 20 states at its April 2024 Series C. The transferable insight for a solo or small-group practice is that the intake step is an access decision: a slow, opaque, or high-friction inquiry process functionally denies care to people you had the capacity to see. The highest-leverage change most practices can make this quarter costs nothing — state your current insurance status and real availability at the point of inquiry, and answer every inquiry within one business day. Software cannot manufacture clinicians; HRSA data shows the mental health workforce meets about 27% of designated need nationally. It can stop a practice from losing the clients it is already able to serve.

The Behavioral Health Access Gap, Quantified#

The behavioral health access gap is the difference between the number of people who need care and the number who receive it, and in the United States it is measured in tens of millions of adults per year.

The three most checkable measurements:

The workforce figures come from HRSA's quarterly shortage-area summary, compiled by KFF's state health facts on mental health care shortage areas: 6,807 designated mental health Health Professional Shortage Areas covering 137,133,953 people, with 27.29% of assessed need met.

Read those two data sets together and a distinction appears that matters enormously for a practice owner. The HPSA numbers describe a supply problem. The NSDUH numbers describe something broader — because a meaningful share of untreated adults are not in a shortage area at all. They are people who looked for care and did not complete the handoff.

How Much of the Access Gap Is Administrative?#

A large share of the access gap is administrative rather than clinical, meaning it is produced by cost, coverage, navigation, and availability rather than by a shortage of clinical skill.

SAMHSA asks adults who perceived an unmet need and did not receive treatment why they went without. In 2024, among all adults in that group:

• 54.5% thought it would cost too much
• 43.6% did not know how or where to get treatment
• 38.2% could not find a treatment program or healthcare professional they wanted to go to
• 33.6% said their health insurance would not pay enough of the cost
• 31.2% did not have insurance coverage for mental health treatment
• 21.7% had problems with transportation, childcare, or getting appointments at times that worked for them
• 14.3% found no openings with the program or professional they wanted

Now isolate the subgroup that actually picked up the phone. Among adults who sought treatment and still did not receive it, the profile shifts hard toward supply and routing: 41.9% could not find a program or professional they wanted to go to, 32.4% did not know how or where to get treatment, and 26.1% found no openings. Those three answers describe the experience of contacting practices and not getting a usable response.

The provider-directory literature says the same thing from the other side. In a peer-reviewed analysis of Oregon Medicaid published in Health Affairs, Zhu, Charlesworth, Polsky, and McConnell found that 58.2% of directory-listed providers were "phantom" providers who filed no claims for Medicaid patients during the study period — 67.4% among prescribers. Comparing directory counts against claims produced a fivefold discrepancy in apparent provider supply: 4.0 prescribers per 1,000 enrollees on paper versus 0.7 in reality.

That is the administrative half of the access gap in one number. The care existed on a list. It did not exist at the moment a person reached out.

Grow Therapy's AI Strategy: What the Public Record Actually Shows#

Grow Therapy's publicly documented AI work is concentrated in clinical documentation and between-session client tools, not in intake triage — and the company has published no detailed AI roadmap for how people are screened, matched, or routed at first contact.

What the announcements do state, on the record:

• At its April 2024 Series C, Grow Therapy raised $88 million led by Sequoia Capital, citing more than 12,000 providers, operations in 20 states, and more than 75 payors across commercial, Medicaid, and Medicare lines. The same release reported more than 3 million patient encounters with 94% in-network, a 4-day average time to care, and a Net Promoter Score of 85, and announced a measurement-informed care system built on repeated clinical assessments.
• At its March 2026 Series D, Grow Therapy raised $150 million led by TCV and Growth Equity at Goldman Sachs Alternatives, bringing total funding to $328 million. That release cites 26,000 providers, expansion from 75 to more than 125 health insurer partners including Medicare and Medicaid, coverage reaching 220 million people nationwide, 7 million visits in 2025, and 10 million lifetime appointments. It reports an average client cost of $21 per visit with one in three paying $0.
• On AI specifically, the Series D release describes free clinically guided AI-assisted notes, with provider documentation time down by nearly 70%, plus client-facing AI tools that support self-reflection between sessions.

That is the verified perimeter. Anything beyond it — including how the Grow Therapy platform decides which client sees which clinician — is not public.

Analysis, not disclosure: the structure of the model implies where the leverage sits. A network that contracts centrally with 125+ payers and enrolls clinicians onto those panels on their behalf has removed the single slowest administrative step in independent practice from the individual clinician's plate. A network with 26,000 providers can answer "who is actually taking new clients on this plan, in this state, this week" as a database query rather than a phone tree. Neither of those is an AI capability. Both are access capabilities, and they are the ones that move the numbers in the SAMHSA table above.

Which Parts of the Access Gap an Insurance-Based Therapy Network Attacks#

An insurance-based therapy network attacks the coverage, credentialing, and discoverability layers of the access gap while leaving the clinical layer untouched.

Three concrete mechanisms, mapped to the barriers people actually report:

1. Panel enrollment moves off the clinician. Credentialing with each payer individually is the classic reason solo practitioners stay cash-pay. Centralizing it converts a per-clinician administrative project into a network-level one, which is what lets a network claim 94% in-network encounters.
2. Coverage is confirmed before booking, not after. "Cost too much" (54.5%) and "insurance wouldn't pay enough" (33.6%) are the two largest reported barriers. A model that shows a plan-specific price before the first appointment converts a financial unknown into a decision. This is the same problem a private practice solves at the inquiry step — see the batch companion on insurance verification during therapy intake.
3. Availability is a filter, not a phone call. "Could not find a professional they wanted to go to" (38.2%) and "no openings" (14.3%) are directory-accuracy problems. Real-time availability is the antidote to the phantom-network effect.

Headway and Alma run structurally similar plays, and the intake implications differ in instructive ways — we broke both down in how Headway's network model changes intake and Alma's therapist-network approach to intake. For the enterprise-payer end of the same market, how Spring Health uses conversational screening at scale shows what the routing layer looks like when it is built deliberately.

Why the Intake Step Is an Access Decision#

The intake step is an access decision because the inquiry process is where a person who needs care either becomes a client or stops trying, and that outcome is determined by your operations rather than by their clinical need.

Consider what "did not know how or where to get treatment" (43.6%) means at practice scale. It rarely means the person could not find a therapist's website. It means they found several, could not tell from any of them whether the practice took their plan, whether anyone had openings, or whether their presenting concern was in scope — and then did nothing. Every one of those three questions is answerable at the point of inquiry, and most practice websites answer none of them.

The same logic applies to the 41.9% who sought care and could not find a professional they wanted to go to. Some of those people contacted a practice that was full. Some contacted a practice that had capacity and never replied. From the person's side, those two are indistinguishable. From the practice's side, only one is a supply problem.

This reframes intake quality away from a conversion metric. A slow inquiry response is not a lost lead; it is a person with a diagnosable condition who exits the system. That framing is why we argue that a client intake process should be designed not to lose clients, and why the therapy client intake form deserves a rebuild for 2026 rather than another field appended to the PDF.

Static forms make the problem worse in a specific way. They ask a person in distress to translate themselves into checkboxes before anyone has acknowledged them, and they collect the same twelve fields whether the answer is "anxiety, mild, flexible schedule" or "postpartum, acute, needs evenings, Medicaid." That is the argument behind AI-first cannot start with a web form and the drop-off patterns in counseling intake forms in 2026.

What a Practice Can Change at Intake#

A practice can close a meaningful part of its own access gap with four operational changes, none of which require a new EHR.

Step 1: Publish insurance status and real availability where people ask. Not "we accept most major insurance." Name the plans, name the states you are licensed in, and state whether you are accepting new clients this month. If the answer is no, say so — a fast, honest "no" plus a referral is better access than silence.

Step 2: Set and measure a response-time standard. One business day is a defensible floor. Measure it: median hours from inquiry received to first human or agent reply, tracked monthly. Practices that have never measured this are usually surprised, and the fix is often scheduling rather than staffing.

Step 3: Capture enough at inquiry to route the person well. Routing quality is what separates an inquiry process from a contact form. The minimum viable set:

The left column is a conversation, not a questionnaire — a person describing their situation in a few sentences yields better routing than eight dropdowns, because the follow-up question is what surfaces "evenings only" or "my last therapist wasn't a fit because…". That is precisely the job our therapy intake template is built for, and the broader pattern is documented in conversational screening for mental health practices. If you are choosing tooling, the batch comparisons on client intake software for therapists, telehealth intake tools, and mental health screening tools cover the landscape by workflow.

Step 4: Keep the clinical record where it belongs. Perspective is SOC 2 Type II and ISO 27001:2022 certified, with data encrypted in transit and at rest. Perspective is not HIPAA-certified — for workflows involving protected health information (PHI), contact us to discuss your requirements. The practical pattern for a private practice is to run the pre-intake inquiry conversation in a concierge agent, then hand structured output to the EHR or practice-management system that holds the record. Keep your EHR; replace the inquiry form. Practices that need a signed BAA for a specific workflow should evaluate vendors that offer one — several intake products in this category do, and we cover that evaluation in HIPAA-ready intake tools for private practice.

Once inquiries are routed well, the next leak is the gap between "booked" and "attended," which the companion post on reducing therapy no-shows at intake covers directly.

What Software Cannot Fix: Clinician Supply#

Software cannot fix clinician supply, and any intake vendor claiming otherwise should be discounted accordingly.

HRSA's designation math is blunt: about 6,800 additional practitioners would be required to lift the existing mental health shortage-area designations, and the share of assessed need met moved only from roughly 26% to 27% year over year. A network can make existing clinicians easier to find and easier to pay for. It cannot add clinicians to the labor market, and neither can a conversational intake layer.

Two honest limits follow. First, if a practice is genuinely full, better intake changes who gets the next opening and how fast the rest get referred out — it does not increase capacity. Second, faster routing can make a waitlist longer rather than shorter, because it surfaces demand that previously gave up silently. That is a better problem than the alternative, but it is a real one, and it is worth planning for before it arrives.

The realistic claim is narrower and still worth making: most small practices are not turning away clients because they are full. They are losing clients between the inquiry and the first reply. That is the part software fixes.

Frequently Asked Questions#

What is Grow Therapy's AI strategy?#

Grow Therapy's publicly stated AI work covers clinically guided AI-assisted notes for providers — which the company reports cut documentation time by nearly 70% — and client-facing AI tools that support self-reflection between sessions, per its March 2026 Series D announcement. The company has not published a detailed AI roadmap for intake, triage, or client-provider matching. Claims beyond documentation and between-session support should be treated as speculation rather than disclosure.

Is the behavioral health access gap caused by too few therapists or by administrative friction?#

Both, and the split is measurable. HRSA data shows 137.1 million people living in designated mental health shortage areas with about 27% of need met, which is a genuine supply constraint. But SAMHSA's 2024 survey shows that among adults who sought treatment and still went without, 41.9% could not find a professional they wanted to go to and 26.1% found no openings — outcomes shaped by directory accuracy, coverage clarity, and response time as much as by headcount.

How does an insurance-based therapy network improve access to care?#

An insurance-based therapy network improves access by centralizing payer contracting and credentialing so individual clinicians do not each negotiate panel enrollment, and by making plan-specific cost and current availability visible before a person books. Grow Therapy reported expanding from 75 to more than 125 health insurer partners between its 2024 Series C and 2026 Series D, with an average reported client cost of $21 per visit. The clinical work itself is unchanged.

What should a private practice capture at the first client inquiry?#

Capture presenting concern in the person's own words, state of residence, telehealth or in-person preference, schedule constraints, insurance carrier or self-pay intent, prior therapy experience, and urgency with a clear risk-escalation path. That set is enough to route well without collecting protected health information before a clinical record exists. Detailed history, consents, and assessments belong in the EHR after the person is accepted.

How fast should a therapy practice respond to a new client inquiry?#

Within one business day is a defensible standard for a solo or small-group practice, and the metric to track is median hours from inquiry received to first reply. Speed matters because a person who contacted three practices will start with whoever answers first, and because SAMHSA data shows a large share of people who seek treatment simply stop after failing to find a responsive provider.

Does Perspective AI replace an EHR like SimplePractice or TherapyNotes?#

No. Perspective replaces the form at the front of the funnel — the inquiry form, the "request an appointment" form, the pre-intake questionnaire — with an AI conversation that follows up on vague answers and returns structured output. It is not an EHR, practice-management system, billing engine, or documentation tool. It sits in front of those systems and feeds them.

Conclusion: Treat Intake as an Access Decision#

The useful lesson from Grow Therapy's AI strategy is not the AI. It is that a company chasing behavioral health access chose to attack credentialing, coverage clarity, and availability — the administrative half of a gap that leaves roughly 47.9% of U.S. adults with any mental illness untreated — because that is the half a business can actually move. The clinical half, the HRSA shortage math, does not yield to product decisions.

A solo or small-group practice cannot build a 26,000-provider network. It can do the same thing at its own scale: state which plans it takes and whether it has openings, answer every inquiry within one business day, and capture enough at first contact to route the person well or refer them out fast. Each of those is an access decision disguised as an operations task.

If you want to see what that looks like in practice, start with the therapy intake template or the broader patient intake template, read the walkthrough for private-practice intake for counseling clients, and look at how intelligent intake turns an inquiry into a routed, structured handoff. Practice managers and operations leads running this end-to-end should start at built for operations teams — that page maps the workflow to the people who own it. When you are ready to test it against your own inquiry volume, set up your first intake conversation and measure the response-time number you are working with today.