Case study · Cross-Industry

Five AI builds in five industries, each started with a secure connection

How Trellist started five AI engagements in five unrelated industries the same way, with a custom MCP server into each client's own systems, then layered on AI capability and handed over the controls.

Let’s Talk See the results

Client: Five clients across five industries
Industry: Cross-Industry
Services: AI Transformation AI Agents Custom MCP Servers Fractional AI Team Internal Assistants Workflow Modernization Infrastructure Engineering Integrations Workflow Automation
Platforms: LeadFlow AI

The challenge

Five clients came to Trellist with five different problems: manual reporting in marketing operations, admin work and reconciliation in live events, fragmented lead flow in B2B distribution, constant listing management in e-commerce, and interface-bound publishing in consumer content.

Lined up side by side, the problems shared a root cause. In each case the bottleneck was the interface itself, the screens and logins people had to work through, more than the process running behind it.

What we did

In every engagement, Trellist built a custom MCP server first: a secure, governed connection into the client's own systems. The AI capability came second, on top of a connection that already existed.

Marketing operations reporting. For a national marketing and creative services agency, a secure MCP connection between Claude and the project-management platform lets managers ask for utilization and burn against live project and staffing data.
Live events, two connections. For a live events and registration platform provider, we scoped two separately governed builds: one puts admin management behind a prompt, and the other brings Stripe data into the same conversation for end-to-end reconciliation.
B2B distribution lead flow. For a B2B distribution business, a system spanning two CRMs now owns the pipeline from intake form to distributor assignment. LeadFlow AI proved out here first.
E-commerce operations. For an enterprise reseller operation, full integration with the marketplace's MCP lets Claude run listing, pricing, order, and inventory management within guardrails the client's team sets.
Consumer content publishing. For a multi-brand consumer products company, custom MCP connections let content be created, updated, and published for every brand by prompt, with no trip through the CRM interface.
One methodology throughout. All five followed the same AI-Powered Transformation methodology (Define, Audit, Propose, Execute, Enablement), delivered by a Fractional AI Team embedded in each client's operation.

The impact, measured

Across all five clients, tasks that once took several steps in several systems now run as a single prompt, on connections each team controls.

Engagements, one build order

Marketing ops, events, distribution, e-commerce, content

Reporting that comes to the manager with utilization and burn available on request.
Separate systems in one place so admin management and financial reconciliation run in a single workflow.
An entire operating function run by AI agents within guardrails the client sets.
Publishing by prompt for every brand, every time.
Builds that last because each one ends with the controls handed to the client's team.

Let’s Talk