Evening BCBA Carrie was in crisis mode for 1.5 weeks over a tough client case. Neither Jana nor Carrie escalated despite it requiring specialized training. Could have been resolved in 5 minutes with proper channels.
Result: Lost an RBT. Direct business cost from broken communication. This is the single strongest argument for the voice memo / escalation system.
Quarterly meetings with Shell aren't surfacing real issues. Employees say "everything's great" then quit immediately after.
Root cause: People don't feel safe being honest. By the time problems become visible, it's too late. The voice memo system replaces this with continuous, low-friction, optionally anonymous feedback.
Stop expecting everyone to do everything. Let people own what they're good at. Pair humans with AI systems that handle the rest.
People-facing role. Primary focus on mentorship and supervision. Removes mentorship burden from other BCBAs. Creates a clear career path.
BCBA + IO Psychology background — rare combo. Systems optimization, process improvement, AI tool adoption lead. Bridges clinical knowledge and business operations.
Her current role gets largely automated. The honest question: is the role retainable with AI augmentation, or does it naturally phase out? If retainable — what does the evolved version look like?
Staff record daily voice memos → Whisper Flow transcribes ($14/mo) → AI categorizes and routes → Tash gets daily digest. Replaces the failing quarterly check-in model with continuous feedback.
Company handbook, training materials, and procedures loaded into a chatbot. Staff ask questions anonymously. Reduces burden on senior staff. Reveals training gaps from what people are asking.
Purpose-built apps for each role (RBT, BCBA, Admin, Management) rather than training everyone on a general AI tool. Cleaner, more secure, more effective.
Scheduling, payroll, client communication, document generation, training material updates — all automated with human oversight.
Baseline metrics for operational costs. Staff sentiment tracking from voice memos. Capacity planning. Financial dashboards. Data-driven decision making.
| Tool | Cost | Purpose |
|---|---|---|
| Claude (HIPAA-compliant) | API pricing | Core AI engine — chatbot, apps, automation, document generation |
| Whisper Flow | $14/month | Voice-to-text transcription for staff memos |
| Agentic AI tools | Dev time | Automated workflows, scheduling, document generation |
| Custom web apps | Dev time | Role-specific interfaces (RBT, BCBA, Admin, Mgmt) |
One person (Matt) builds and maintains all AI systems. Staff interact only with purpose-built interfaces, never raw AI. This is cleaner, more secure, and eliminates compliance risk from untrained users touching AI directly.
Tash's strong recommendation: do NOT start free. Get BACB and KSBSRB approval first, then formalize compensation. This protects Matt's contributions and avoids the pattern of people taking free work indefinitely.
The rule: Proposal and planning only until board approval is secured. No building, no deploying, no free work beyond the initial consultation.
Matt's belief: 90% displacement within 18 months. Getting Tash's company on AI infrastructure now is an early mover advantage. Human-centered services (therapy, ABA, coaching) remain safer longer. Admin roles are the most immediately vulnerable — Jana's situation is a real-world preview.
This consulting work could formalize into a service offering beyond video production: WordPress fixes hands-free, full websites in 1.5 hours, automated business workflows. The rehab project is the first real case study.
Tash is carrying enormous weight — managing difficult staff while running a company. The Kim situation has a personal sting. Building systems that reduce Tash's direct management burden isn't just efficiency — it's sustainability.
Currently at 2 projects (vs. usual 6) — real bandwidth available. Spring break limits this week (Wed–Fri). Planning and async work only until next Monday, then building can begin (pending board approval progress).