Flowhub has released a new connector called Flowhub MCP, built on the open Model Context Protocol standard, that lets dispensary operators link their Flowhub account directly to AI tools including ChatGPT, Claude, Gemini, Grok, Perplexity, and Cursor. The point isn't just data visibility - it's execution. Operators can issue natural-language instructions and have Flowhub carry out the underlying work, from price adjustments to inventory transfers, inside the platform itself.
That distinction matters more than it might first appear. Most cannabis retailers are already drowning in administrative overhead - SKU management, vault-to-floor inventory moves, promotional deal creation, margin calculations - tasks that eat into the time store managers should spend on staff training, customer flow, and compliance. The ability to say "increase every vape SKU by 5%" and have the system act on that instruction, rather than simply surface the data, represents a different category of tool. For operators curious about how that kind of operational integration translates in practice, it's worth taking a moment to see how it works in the context of regulated cannabis retail environments where compliance and speed have to coexist. Flowhub's design requires operator approval before any changes are applied, which maintains the audit trail that seed-to-sale compliance demands - a non-negotiable in any licensed market.
The underlying protocol - MCP, developed as an open standard - is what gives this approach its flexibility. Rather than building a proprietary AI layer that locks operators into a single vendor's model, Flowhub is positioning itself as a connector that can work with whichever AI system a retailer already trusts. Kyle Sherman, Flowhub's founder and CEO, framed it plainly: "The best AI of 2026 won't be the best AI of 2028." That's not a throwaway line. Cannabis retailers have watched POS vendors, compliance software providers, and data platforms come and go. Betting on a single AI vendor today carries real switching-cost risk down the road.
Why Operational Overhead Is a Structural Problem for Cannabis Retail
Licensed cannabis retail is, by design, a compliance-heavy business. Every transaction touches a state-mandated tracking system. Inventory must reconcile against purchase orders, transfer manifests, and METRC logs. Pricing changes have to account for excise tax pass-through calculations, bundle deal restrictions that vary by state, and margin floors that determine whether a promotion actually generates positive contribution or just moves product at a loss. None of that is glamorous work, and most of it is done manually - or semi-manually - by managers who are also running a retail floor.
The result is a familiar problem for any multi-location operator: the people best positioned to make good business decisions are spending their hours on data entry instead. Flowhub MCP's stated use case addresses this directly. The example of a query like "Show me my slow-moving products and create deals to sell through them while maintaining at least a 30% margin" isn't a marketing flourish - it describes a workflow that, done manually, requires pulling an inventory aging report, calculating cost-of-goods on each SKU, setting a discount floor, and building the promotion inside the POS. That's not a five-minute task.
Early Operators Are Already Building Custom Workflows
Ankit Bhasin, owner of Cannabis Cowboy, described his setup directly: Claude connected to Flowhub through MCP, analyzing existing store data and generating actionable recommendations - including daily promotion suggestions. His goal, as he put it, is "one-click everything" for his management team. That's a reasonable operational target for any retail business, but it carries particular weight in cannabis, where differentiation on product assortment is limited. Most licensed dispensaries in a given market carry overlapping SKUs from the same licensed cultivators and processors. Margin, customer experience, and promotional execution are where stores actually compete.
The compliance angle deserves attention here. Any system that executes changes inside a licensed retailer's POS - price updates, inventory adjustments, deal creation - has to maintain a reliable audit trail. Flowhub's design, which requires operator approval before changes are applied, addresses that requirement. But operators integrating third-party AI tools will want to confirm how those interactions are logged, whether they satisfy state-level recordkeeping requirements, and how the system handles edge cases where an AI instruction might conflict with a compliance rule. Those aren't hypothetical concerns; they're the standard questions any compliance-minded operator should ask before connecting any external tool to systems that touch regulated inventory.
The Broader Implication for Cannabis Retail Technology
What Flowhub is doing with MCP fits a pattern emerging across regulated industries: open protocols that let businesses own their data and choose their tools, rather than being locked into a vertically integrated software stack. In cannabis, where operators already navigate complex integrations between POS systems, METRC, e-commerce menus, loyalty platforms, and payment processors, adding another proprietary layer is genuinely unattractive. An open connector approach - if it holds up under real-world compliance scrutiny - offers a more durable foundation.
The short version: AI that only talks to your data is a dashboard. AI that acts on your data, inside your operating system, with an audit trail and operator approval, is something closer to an automated back-office. For cannabis retailers operating on thin margins, facing rising compliance costs and competitive pressure, the difference is worth paying attention to.
