By Michael Sassano, Founder, Chairman GR, Interim CEO of SOMAÍ Pharmaceuticals
This is the first in a series. Each function gets its own deep dive.
Let’s be real about what medical cannabis actually is. Beneath the lush green canopy and the glossy brand jars, we are running a heavily regulated pharmaceutical machine built on endless paperwork, strict compliance, and mountains of data.
And guess what? That happens to be exactly what Artificial Intelligence eats for breakfast.
Over my next few articles, I’m taking us on a function-by-function breakdown across the entire enterprise — from finance, sales, and QA to cultivation, regulatory affairs, and the C-suite. We’re going to map out what AI obliterates, what it can’t touch, and what a hyper-lean, AI-driven medical cannabis operator actually looks like.
Spoiler alert: AI is taking over almost everything, except for the tiny sliver of operations where the law explicitly demands a warm body with a pen.
The Regulatory Floor: Why the Law Keeps Humans in the Building
Pharmaceutical regulators don’t care about your cool new AI tool. Under EU GMP Annex 11, every single computerized system touching product quality — your ERP, QMS, and LIMS — must be validated, generate an audit trail, and sit under a named, trained human who is legally on the hook for it.
The math is simple: AI takes the workload; an accountable human keeps the legal signature. That single rule is the filter for every department in this industry.
Department Breakdown: What Stays and What Shrinks
1. Finance: The Self-Driving Balance Sheet
Why do you have teams of people staring at spreadsheets when margins are razor-thin and vendors pay slow? AI runs the ERP, reconciles the books continuously, tracks aging receivables in real time, and hunts down cash flow gaps. You don’t need a finance department — you need one sharp administrator watching a dashboard.
2. Sales: AI Finds ‘Em, Humans Close ‘Em
The days of sales reps spending 80% of their week doing admin work are dead. AI bots identify pharmacy inventory needs, make first contact, and hand warm leads directly to a human closer. Once the deal is done, AI pushes the order through the pipeline and tracks the cash until it hits the bank.
3. Quality Assurance & Labs: Paperwork Gets Automated, Not the QP
An AI-driven QMS can track deviations and CAPAs 24/7. But under EU GMP law, the Qualified Person (QP) sign-off must remain human. AI flags the anomalies and prioritizes the issues, but a human QP takes the legal bullet for batch release. In the labs, AI speeds up testing analysis, cutting lab headcount down to the bare physical minimum.
4. Cultivation & Manufacturing: Data Over Instinct
Sensors and AI track plant health, environmental shifts, and yield predictions far better than a grower’s old logbook. While you still need humans on the facility floor for physical tasks (until robotics fully mature), decisions on genetics, lighting, and nutrient feeds will be driven purely by data — not gut feeling.
5. Warehousing: Robots Pick, Humans Validate
Robotic picking and packing take over the floor while AI manages shipping schedules and logistics paperwork. Under EU Good Distribution Practice (GDP), you still need one named Responsible Person (RP) to check compliance, but AI handles the heavy lifting.
6. Marketing: Fully Automated Except the Final Call
Labels, CRO-facing claims checks, and social media content all run through AI end to end. What stays human is one marketing architect who approves feel and direction — claims about a medicinal product still need a person accountable for what’s said publicly.
7. Pricing & Product: AI-Driven Feedback Loops
AI agents watch market pricing and demand trends across regional websites and feed that straight into decisions on which genetics to grow or which products to manufacture — faster and more accurate than a team of analysts.
The Executive, Legal, and Back-Office Reality Check
- Regulatory Affairs & Import/Export: AI can draft complex multi-jurisdictional filings for agencies like BfArM, Infarmed, or the MHRA, and auto-generate narcotics import/export permits for cross-border scheduled shipments. But the named Regulatory Lead and the licensed narcotics signatory still have to sign the final documents.
- Customer Service & Pharmacovigilance: AI handles routine pharmacy tickets and order updates. But the second a product complaint or adverse event hits, it routes to a human. EU law requires a named QPPV (Qualified Person Responsible for Pharmacovigilance) to legally report safety incidents — that role shrinks to one accountable person leaning hard on AI-drafted case reports, not to zero.
- IT Security & GDPR: Patient and prescriber data is sensitive-category data under GDPR. AI can monitor systems and flag anomalies continuously, but breach reporting and DPO-level accountability stay with a named person, not a dashboard.
- HR & Recruiting: Sourcing, resume screening, and scheduling are 100% automated. HR shrinks down to the human judgment calls — culture fit, disputes, and hard conversations.
- Investor Relations: AI can prep the numbers, draft updates, and flag trends worth raising. But the actual relationship with investors — the calls, the trust, the judgment on what to disclose and when — stays with whoever’s building the company. This one doesn’t shrink; it just gets better inputs.
- The C-Suite & Legal: Most executive work is just reading data and reacting. AI does that faster, cheaper, and without ego. Executive teams will shrink dramatically. As for legal? AI contract tools handle the NDAs and routine paperwork; lawyers only survive for the high-stakes battles that actually reach a courtroom.
The Bottom Line
The real question confronting global cannabis operators isn’t if AI is going to take jobs. The real question is: How few people do you actually need to run a multi-million-dollar, fully compliant medical cannabis company?
That’s what we’re going to map out in this series, function by function, with practical moves you can make today.
Which department in your business do you think is most vulnerable to an AI takeover first?
