Case Studies

A look at the kind of work we do and the problems behind it.

Each engagement started with an operational problem. We chose the technology to fit that problem.

Machining & Industrial Services · South Texas

Every call answered, with the foreman still on the floor

Situation: A South Texas machine shop with two decades of history serves oil and gas, agriculture, and industrial customers. The phone was answered by whoever was free. That usually meant the foreman, who also ran jobs on the floor.
Impact: Every call he took was time off the machines. Every call he missed went to voicemail, and a same-day callback depended on when he could get back to it. Robocalls added screening time on top of that.
Approach: Capstone built an intake system that answers every call, screens out auto-dialers with a short menu, and rings the foreman on his cell while the caller holds. An assistant takes over when he cannot pick up. Web leads and phone leads land in one place.
Result: A caller now reaches either the foreman or an assistant that records a structured message with their name, number, and what they need. If any one piece of the pipeline fails, the lead still reaches the shop by another path.

A caller hears a short menu: existing order, new work, or something else. Auto-dialers do not press digits, so the keypress itself filters them out.
While the caller holds, the system rings the foreman's phone and asks him to accept. If he picks up, the two are connected.
If he is on a machine or the call would reach his voicemail, the caller is handed to an assistant that records what they need.
Phone leads, web form leads, and after-hours messages all land in the same lead sheet the owner already checks, with a text alert for anything urgent.
If part of the pipeline goes down, the lead reroutes by email. The system was built so that missing a lead is the one failure it does not allow.

Membership Operations & Custom Software · Texas

Situation: A Texas membership organization was launching a paid, application-based program. The founder reviews every applicant personally. He needed an application funnel and a way to track every lead from the first day, and a full CRM subscription was more system than the organization needed.
Impact: Applications, assessment results, and inbound questions were about to arrive from several directions at once. Without one place to see them, follow-up would depend on memory during the busiest weeks of the launch.
Approach: Capstone built the public site, three self-assessment tools matched to the organization's own framework, and a private admin view that reads and writes against the spreadsheet the founder already used, with lead status tracking and follow-up email built in.
Result: The founder logs into one protected screen, sees every application and assessment as it arrives, tracks where each lead stands, and sends follow-up from the same place. The engagement has continued through weeks of revisions because the tooling is in daily use.

Visitors can take a public self-assessment without giving an email. The result page offers a follow-up by email as an option the visitor can skip.
Two further assessments are private and used as homework before a prospective member's first call, built around the organization's own framework.
A three-step application writes to the same records the lead list uses, so who applied and who is a lead never need reconciling.
The admin view sits behind a sign-in and is invisible to the public. Inside, the founder gets a live lead table, per-lead status, outbound email with a sent-mail log, and a dashboard that rolls up the whole funnel.
The back office runs on the spreadsheet the founder already used, kept because it fits how the organization works today.

Our Own Operation · Capstone AI

Situation: Capstone is a small team. At any point we are in conversation with more than twenty prospects across industrial, construction, government, and services work. Each conversation carries call notes, research, and a proposal in progress. A firm this size normally handles that with memory and a busy inbox, or hires someone to manage it.
Impact: Whatever we tell clients about running an operation with AI has to hold up in our own shop first. If our answer to growing workload was a new hire or a software subscription, our advice would be worth less.
Approach: We run the business out of a plain-text repository that AI agents operate under written procedures. Research on a prospect, preparation for a meeting, notes from a call, and the proposal that follows are each produced by an agent following a documented playbook, then reviewed by a person before anything leaves the building.
Result: Every prospect gets researched preparation before the first call and a written proposal shortly after it, from a firm with no operations staff. If we have met with you, you have already seen this system's output.

Before a first call, an agent researches the prospect's business: how many customers their public footprint suggests, where their workflows lean on manual effort, and what the strongest operators in their space are doing differently. Every claim in the research links to its evidence.
Meeting preparation follows a written template that starts from the prospect's point of view: what they need to hear, what to avoid, and what happens after the call.
Call recordings become transcripts, transcripts become structured notes, and notes become the draft proposal. A person reviews each step and owns the final word.
The pipeline itself lives in the same repository, with a written status for every conversation, so anyone on the team can pick up any engagement with full context.
The procedures are versioned documents. When we improve how we prepare for a meeting or write a proposal, every future engagement inherits the improvement.

Sound familiar?

Describe what is slowing your organization down. We will tell you honestly whether we are the right fit.