What the AI staff sells.
Five consulting offerings for small and mid-sized businesses. Two to eight weeks, scoped per engagement. The same agents and methodologies we run inside our own operations—pointed at yours.
Five offerings. Every one of them is something we have already built and run inside RememoryLab or BSoT MTG before selling it outward. We use the same Skills, MCP servers, agents, and tools we ship to clients to operate our own business day to day—so when we say a methodology works, we mean it runs in production, not just in a demo. Pricing varies by scope; the ranges below are typical. Any engagement can be delivered as a proof-of-concept with a fixed deliverable and timeline, where the output is working software that proves or disproves a hypothesis rather than a production system.
Narrow, durable AI capabilities scoped to a specific business function. A Skill is a packaged behavior an agent can invoke—a pricing analyzer, a contract reader, a domain-specific reasoner—built to be reused across workflows and durable across model upgrades. We ship the Skill, the prompts, the evaluation harness, and the documentation needed for your team to extend it.
Agent-accessible interfaces to your systems. The Model Context Protocol gives agents a clean, typed surface to read from and write to internal tools—databases, ERPs, internal APIs, vendor systems. We design the tool surface, build the server, harden the auth path, and hand over running infrastructure plus the source. Two-to-six-week engagements typical.
Customized agents that run continuously in a specific operational or personal context. Proactive personal assistants for busy professionals. Adaptive workflows for high-functioning neurodivergent operators. Operational agents that watch a queue, decide what to do, and do it. Each one is bespoke—the value is in the context-fit, not the underlying model. Built on the same runtime that runs RememoryLab.
Multi-agent setups running real workflows. A team of specialized agents—writer, developer, manager, reviewer—coordinated by a shared queue, charter, or orchestration layer. The pattern that runs our own operations, applied to a function or division of your business. Longer engagements; deliverable is a running team plus the operating documentation that lets you grow it.
General-purpose AI-integrated software. Production tooling that embeds agent capability into existing operations—internal dashboards, customer-facing applications, automation pipelines, scraping and data systems. Where the work is “build us a thing that uses AI well,” this is the offering.
MTG Price Arbitrage Tool
A wants-list-driven arbitrage system using a canonical card-identity layer, Playwright-based scraping across multiple live-commerce vendors, per-vendor condition offsets, and a confidence-scored matching engine. Surfaces the cards worth pulling from inventory before a live show, ranked by expected margin. Runs continuously against the live data.
MTG MCP Server
MCP server exposing trading-card operations and inventory to AI agents—lookup, pricing, condition grading, stock movement. The interface the BSoT MTG agent uses to read and act on the business. Built once, reused across every workflow that touches inventory.
Elliot Wave Theory Skill
A Skill encoding Elliot Wave Theory analysis for technical-trading workflows. Reads market data, identifies wave structures, surfaces trade candidates with annotated reasoning. The first Skill we shipped to an external party; the framing held up under real use, and the pattern generalizes to other technical-analysis frameworks.
Two to eight weeks, scoped per engagement. We start with a one-week discovery—reading your systems, talking to the people who know how things actually run, writing a short scoping memo with the recommended shape, deliverable, and timeline. If the memo lands, we ship. If it doesn’t, you’ve paid for the discovery and we part on good terms.
Client owns the deliverable. We retain rights to reuse generic patterns and publish anonymized case studies. Client-engagement work lives in a per-client repository provisioned at engagement start; nothing client-specific lives in RememoryLab’s main codebase.
Capacity is limited. We take a handful of engagements per quarter. If the queue is full, we’ll say so.
The fastest path is a short note describing the problem and the shape of work you’re imagining. Email [email protected] or call 612.405.3460. We respond within two operating days.