Customer results
Maryland Wellness & Psychiatry

Six in ten patient calls never reach the front desk
Maryland Wellness & Psychiatry is an outpatient psychiatry practice in Frederick, Maryland. Katie, MedReception AI's receptionist, has answered their phone since April 2026: 1,774 calls, of which 1,043 were handled start to finish without anyone at the practice picking up. MedReception AI also rebuilt and now manages their website, where Google impressions have risen roughly 50% since measurement began.
- Practice
- Outpatient psychiatry and behavioral health
- Location
- Frederick, Maryland
- EHR
- Tebra
The phone: what Katie handled
1,774
Patient calls answered
Every one picked up, on the practice's own number
20 April 2026 to 18 September 2026
1,043
Calls handled without the practice
58.8% worked by Katie and never passed to a person
20 April 2026 to 18 September 2026
43.8 hrs
Time on the phone with patients
2,626 minutes Katie spent so staff did not have to
20 April 2026 to 18 September 2026
212
Calls outside office hours
12.0% before 9am, after 5pm, or at the weekend
20 April 2026 to 18 September 2026
819
Distinct patients calling
No single caller drives the volume. The busiest was 1.9%
20 April 2026 to 18 September 2026
Every inbound call to the practice's main number.
Share of categorised calls.
- Message for the practice47.6%
- Change an appointment13.6%
- Request an appointment11.3%
- Prescription refill7.7%
- Billing6.5%
- New patient4.5%
- Question for a provider4.4%
- Everything else4.4%
Katie went live on 20 April 2026, so April is a partial month and is marked as such. The chart shows complete months only; September is still in progress and is left out. Across the 109 weekdays between go-live and 18 September there was not a single weekday with no calls. The totals above cover the full period. Categorised calls between 1 August and 18 September 2026 (n=664 of 758 reviewed calls). Earlier months are not categorised in our review system, so this mix describes that window rather than the whole engagement.
Every call gets picked up
The practice's main number rings Katie first, so these are all of their inbound patient calls rather than an overflow line. She has answered that phone since 20 April 2026. Across the 109 weekdays between then and 18 September there was not one weekday without calls, and every gap longer than eighteen hours falls on a weekend or a public holiday. Patients are not queued, and they do not reach a voicemail box during the day.
Most calls never reach the front desk
Of 1,774 calls, 1,043 were worked by Katie and closed without involving the practice at all (58.8%). She takes the caller's name and date of birth, captures what they need, and passes it to the practice as a written message. The remaining 30.4% she routes to a person, usually inside the first minute. The practice's staff did not pick up the phone for roughly six calls in every ten.
Nearly half of all calls are a message, not a transaction
The single largest category is a patient leaving a message for the practice (47.6% of categorised calls). That is the work Katie removes most cleanly: instead of a staff member stopping to take a message, the message arrives already written down, with the caller identified. Appointment changes and requests together make up another quarter.
The phone is answered when the office is not open
212 calls (12.0%) arrived before 9am, after 5pm, or at the weekend. 106 came in before the office opened and 72 after it closed. These are the calls most likely to have gone unanswered, and they are answered.
The practice grew, and the front desk kept up
Over the same period, Maryland Wellness & Psychiatry scaled from two clinicians to six. Appointments kept on their schedule rose from 276 in January 2026 to 574 in August. This is the practice's growth, not ours. Most of it is the capacity they added. What it shows is that the call volume arriving alongside that growth was absorbed without the front desk growing with it.
The practice's own schedule, all clinicians, all booking channels.
Counted from the practice's own appointment schedule as of 2 September 2026: appointments not cancelled and not marked no-show, across all six clinicians. These totals include every appointment on the schedule regardless of how it was booked: by phone, by the practice's staff directly, or through the practice's own online booking. The schedule carries no booking-channel field, so the share MedReception AI handled cannot be separated from it and is not claimed here. Figures were counted on 2 September 2026, shortly after the August month end.
What MedReception AI runs for this practice
- Katie, the AI receptionist, answering every inbound patient call
- Messages, appointment requests and refill requests captured and passed to the practice in writing
- Routing to a person when a call needs one
- Website rebuilt on Next.js and managed by MedReception AI
- Technical SEO, structured data, internal linking and sitemap
- Ongoing search-performance monitoring in the practice's own Search Console property
In addition: the website we rebuilt and now manage

66,656
Google Search impressions
Across 124 days of measurement
16 May 2026 to 16 September 2026
1,802
Organic clicks to the site
Patients arriving from Google search
16 May 2026 to 16 September 2026
+51%
Daily search impressions
416/day → 628/day. Clicks over the same windows rose 17%, which is within normal variation.
16 May to 12 June 2026 vs 20 August to 16 September 2026
9 → 15
Pages earning Google visibility
Distinct URLs earning impressions per month
June 2026 vs August 2026
8.6
Average Google position
Impression-weighted, across all queries
16 May 2026 to 16 September 2026
Google Search performance, month by month
Times the practice appeared in Google results.
Visits to the site from those results.
Search Console coverage begins 16 May 2026, the day the practice's property was set up, so May is a partial month and is marked as such. The chart shows complete months only; September is still in progress and is left out. The totals above cover the full measured window through 16 September 2026.
View this data as a table
| Month | Days measured | Impressions | Clicks |
|---|---|---|---|
| May 2026partial | 16 | 6,232 | 194 |
| Jun 2026 | 30 | 16,102 | 411 |
| Jul 2026 | 31 | 15,473 | 453 |
| Aug 2026 | 31 | 18,854 | 500 |
Where the measurement starts
The practice had no Search Console property, so nobody, including the practice, could see what Google was doing with their site. Setting one up was among the first things we did, on 16 May 2026. Google does not backfill, so every figure on this page is measured forward from that date. It is worth being precise about what that baseline is: by 16 May the rebuilt site had already been live for roughly six weeks, so these numbers describe the trajectory of the site we built, not a comparison against the practice before we engaged. No such comparison exists.
What we built
We rebuilt the site on Next.js with a full service and conditions architecture, wrote and shipped the pages the practice actually needed, and fixed the internal link graph so every page sits within two clicks of the homepage. We added FAQ and breadcrumb structured data, an auto-generated sitemap that cannot drift from the filesystem, and canonical handling across the domain's host variants.
What the audit caught
In August 2026 a technical audit found a hosting misconfiguration that was collapsing most page URLs onto duplicate homepage content. The pages existed and were being built correctly, but Google was being served the homepage for nearly every address. We corrected it and re-verified all 97 live URLs as Googlebot. In the month that followed, the number of distinct pages earning Google impressions rose from 9 to 15.
Where the phones fit
Katie, MedReception AI's receptionist, answers the practice's inbound calls across six providers with materially different schedules: one works Sundays, one Saturdays, one runs an evening clinic. Call handling and the website are managed together, which is why search visibility and phone coverage are reported on the same page rather than as separate wins.
What we shipped, and when
April 2026
Rebuilt on Next.js
The single-page brochure site was replaced with a full multi-page architecture.
16 May 2026
Measurement begins
The practice's Search Console property was set up. Google does not backfill, so this is the first day any search data exists for the site.
June 2026
Service and location pages shipped
Service, condition and Maryland geography pages built out, with FAQ and breadcrumb structured data.
29 June 2026
Internal link graph repaired
Sixty broken internal links fixed and footer topic columns added, putting every page within two clicks of the homepage.
2 August 2026
Hosting misconfiguration corrected
A framework-preset issue collapsing page URLs onto duplicate homepage content was fixed, the CDN cache purged, and all 97 live URLs re-verified as Googlebot. Structured data was moved server-side on 91 pages so crawlers could actually read it.
7 August 2026
Service list corrected at the practice's request
Ten pages covering services the practice does not provide were removed and redirected. Search visibility continued to rise afterwards.
31 August 2026
Team and service imagery restored
Provider photographs restored and imagery added across the service pages.
“Since partnering with Medreception AI, their system has made a huge difference in managing our inbound call volume, giving our staff more time to focus on important daily tasks. Their SEO support has also helped drive meaningful growth in new patients.
Paul and Artur are incredibly responsive, supportive, and open to feedback. They listen, improve, and resolve issues quickly. Definitely worth it!”
What these numbers do and do not show
- There is no before. MedReception AI began measuring this practice's calls when Katie started answering them on 20 April 2026, so no record exists of how many calls the practice received, missed, or sent to voicemail beforehand. Every call figure here describes what Katie handled. None of them is a comparison against the practice's previous call volume, because no such measurement exists.
- Handled without the practice does not mean resolved. Nearly half of categorised calls (47.6%) are a patient leaving a message, which Katie captures and passes on for the practice to action. The work removed is the taking of the message, not the answering of it.
- 30.4% of calls were routed to a person, usually within the first minute. Katie is answering and triaging this practice's phone, not replacing the front desk.
- Where a call is routed, our telephony records show only that the destination picked up. They do not establish that a particular member of staff answered, so we do not claim it.
- 191 calls (10.8%) were hangups under fifteen seconds where the caller never spoke. They are counted in the 1,774 answered and excluded from the 1,043 handled, which is why the two figures do not simply differ by the transfer count.
- The 12.0% out-of-hours figure uses the practice's published Monday to Friday, 9 to 5. Measured against the wider hours their clinicians actually work, the same calls are 2.8%. We publish the larger figure because it is the rule Katie herself operates on, and note the smaller one here.
- Call and appointment figures are not connected. Nothing in our call records establishes that a call became a booking, and the practice's online booking runs through their own system.
- Search Console coverage begins 16 May 2026, the day the property was set up. Google does not backfill, so no search data exists for the period before that date. Every figure on this page is measured forward from then. None of them compares against the practice's performance before MedReception AI, because no such measurement exists.
- The baseline is not the practice at zero. By 16 May the rebuilt site had been live about six weeks, the homepage was already ranking at an average position of 4.0, and pages predating our rebuild were still earning impressions at positions 6 to 9. This page measures the trajectory of a site already in flight.
- Most of the impression volume sits on the practice's homepage, where we can show the change but cannot name the mechanism that caused it. The portion we can attribute to specific work (the service, condition and location pages we built) accounts for roughly 40% of the gain.
- Growth figures depend heavily on which windows are compared. Across the methodologies we tested, impressions growth ranges from +13% to +203% on identical data. The +51% figure published here uses the two longest equal, non-overlapping, weekday-balanced windows available, and excludes both partial months.
- Click growth over the same windows was +17% and is not statistically significant (p=0.21). We publish it as flat rather than as growth.
- Click-through rate declined over the window, from 3.04% to 2.36%, and average position was essentially flat. More people saw the practice in search results; the share who clicked did not rise.
- A two-day burst of desktop impressions on 15 and 16 June 2026 produced no additional clicks and appears to be a reporting artifact. It is visible in the June bar and is excluded from every growth calculation on this page.
- Part of the measured improvement is the correction of a hosting misconfiguration that was itself affecting the site during the earlier period. We report it in the timeline rather than presenting the recovery as pure growth.
- The practice's online appointment booking runs through their own scheduling system, not through MedReception AI. Bookings made that way are not visible to us, so this page makes no claim connecting search traffic to appointment volume.
- The practice's appointment growth is largely capacity. Over the window they went from two clinicians to six, and the founding clinician's own monthly volume was flat to slightly down. We report appointment volume as the practice's growth and make no claim to have caused it.
- The review quoted on this page refers to growth in new patients. That is the practice's own assessment. We cannot measure it: their online booking runs through their own system, and their schedule records no booking channel, so we publish it as their words rather than as a result we are claiming.
- This is one practice's result. It is not a typical or expected outcome, and results vary with specialty, geography, competition, existing web presence and how long a practice has been live.
How these figures were measured
Call figures are counted from MedReception AI's own per-call records for the practice's phone number, covering every connected call from go-live to 18 September 2026. A call counts as handled without the practice when it was not routed to a person and the caller stayed on the line longer than fifteen seconds. Category mix comes from our call-review process, which covered 81.5% of this practice's calls overall and 99 to 100% from July onward. Search figures come from Google Search Console for the practice's own property, using Google's finalised data. Totals are counted by day rather than by search term, because search-term reporting covers only a fraction of clicks. Average position is impression-weighted, as Google reports it. Growth compares the first 28 complete days of measurement with the most recent 28, matching the number of each weekday in both windows and excluding partial months and a two-day reporting anomaly in June. Statistical significance is tested on daily values.
| Figure | Window | Sample | Verified |
|---|---|---|---|
| 1,774 Patient calls answered | 20 April 2026 to 18 September 2026 | 1,774 connected calls of 1,799 pulled; 25 excluded as no-answer or zero-length | 2026-09-19 |
| 1,043 Calls handled without the practice | 20 April 2026 to 18 September 2026 | 1,774 calls; excludes 191 sub-15-second hangups where the caller never spoke | 2026-09-19 |
| 43.8 hrs Time on the phone with patients | 20 April 2026 to 18 September 2026 | Sum of call duration over 1,774 connected calls | 2026-09-19 |
| 212 Calls outside office hours | 20 April 2026 to 18 September 2026 | 1,774 calls, bucketed in the practice's local time against their published Mon to Fri, 9 to 5 | 2026-09-19 |
| 819 Distinct patients calling | 20 April 2026 to 18 September 2026 | 819 distinct originating numbers across 1,774 calls | 2026-09-19 |
| 66,656 Google Search impressions | 16 May 2026 to 16 September 2026 | 124 consecutive days, no gaps | 2026-09-19 |
| 1,802 Organic clicks to the site | 16 May 2026 to 16 September 2026 | 124 consecutive days, no gaps | 2026-09-19 |
| +51% Daily search impressions | 16 May to 12 June 2026 vs 20 August to 16 September 2026 | 28 days per window (56 total), weekday-balanced | 2026-09-19 |
| 9 → 15 Pages earning Google visibility | June 2026 vs August 2026 | Counted per month across every page on the site | 2026-09-19 |
| 8.6 Average Google position | 16 May 2026 to 16 September 2026 | 124 days, 66,656 impressions | 2026-09-19 |
Published 2026-09-19. Last verified against the primary record 2026-09-19. Published with the practice's written permission.
See how it fits your practice
Every MedReception AI implementation is built around how a practice actually operates: its schedules, routing, escalation rules and existing systems.
MedReception AI results and traction
Platform-wide call volume, integrations and growth.
Katie AI, the AI medical receptionist
How Katie answers, routes and books for a practice.
Tebra integration
How MedReception AI connects to Tebra.