Bellarine Psychology: Less Time on Reports, More Time With Clients – Delivered in Two Days

Brief

In two days of development, we built a suite of tailored AI prompting systems on BastionGPT for Bellarine Psychology that automate the most time-consuming parts of the reporting workflow. Intake forms become narrative summaries, transcripts get mapped to DSM criteria for ADHD and Autism, and assessment data gets extracted into editable HTML tables. Clinicians get more time with clients.

Problem

Bellarine Psychology’s clinicians were spending considerable time on the documentation side of their work, not because the work wasn’t important, but because much of it was highly repetitive. Every report required turning intake form responses into a clean narrative summary, reviewing transcripts to identify behaviours and symptoms relevant to diagnostic frameworks, checking alignment with DSM criteria for ADHD and Autism, and manually transcribing assessment data from NovaPsych into a usable digital format. Individually each step is manageable, but across a full caseload, it compounds into a genuine bottleneck, with hours that could be spent on clinical judgement and client care going instead to documentation that follows predictable patterns every time.

Solution

We designed a set of tailored prompts for BastionGPT, Bellarine Psychology’s existing AI platform, covering three distinct areas of their reporting workflow. Narrative Summarisation processes intake form responses through a prompt series that outputs a clean, readable narrative summary written in the clinical style already used by the practice, ready to use with no reformatting or rewriting. Transcript to DSM Mapping breaks meeting transcripts into sections and surfaces relevant behaviours, symptoms, and observations, then aligns those insights with DSM criteria for ADHD and Autism. Automated Table Extraction passes NovaPsych assessment data through the system to automatically extract tables and convert them into editable HTML for direct use in reports. The development process was collaborative, with outputs reviewed at two checkpoints with the clinical team to ensure everything was genuinely useful and aligned with how clinicians already work.

• Built on their existing platform: no new tools, no new logins, everything runs on BastionGPT, which Bellarine Psychology already uses
• Clinically calibrated outputs: narrative summaries match the style clinicians already use, and DSM mapping gives structured clinical foundations rather than raw AI outputs that need significant editing
• Collaborative development: two review checkpoints ensured the system worked for real clinical use, not just in a demo
• Delivered in two days: fast, focused, immediately useful

Result

Bellarine Psychology now has AI prompting systems that handle the most repetitive parts of their reporting workflow, active and in use within two days of the engagement starting. Narrative generation and ADHD report support are fully operational, with foundational work for Autism mapping and table extraction in place and further iterations planned to extend the system’s capabilities. Clinicians spend less time converting raw inputs into report-ready content and more time on the clinical judgement that actually requires their expertise. Turnaround times for reports have improved and consistency across documentation has increased, with the same structured approach applied every time rather than varying by workload and fatigue.

Under the Hood