The Seat You Never Filled: Building One System to Replace Many

By the numbers

80%	faster to build than a conventional CRM implementation
80%	below typical mid-market transformation budgets
60%	reduction in recurring software subscription spend
12 hrs	of manual coordination returned to the team each week
14 mo	estimated payback on the full build

Figures are HeitechSoft estimates based on the client's prior platform stack and comparable engagements. Final baselines are being confirmed with the client.

---

The per-seat model assumes everyone who touches your business sits at a desk

Most business software is priced per user, per month, forever. That model suits companies where every person who needs access is a full-time employee with a permanent login.

Plenty of businesses do not work that way.

Language Advantage delivers language training through a core operations group and a much wider network of instructors. The instructors need to submit availability, confirm sessions, and file paperwork. They do not need a CRM licence. Under per-seat pricing, a business like this pays twice: once for the core team's seats across every platform, and again in the manual work of relaying information to and from everyone outside those seats.

Catherine Jodar came to HeitechSoft with a practical question rather than a technical one. She wanted fewer platforms, less re-keying, and a system shaped around how the business actually runs.

The objective is to enable our team to work more efficiently, streamline our processes, and reduce the number of separate software platforms we need to manage.

1. The problem was decisions, not storage

Language Advantage was not short on software. Marketing ran in Klaviyo. Delivery ran in Demiks. Both worked. Both produced activity.

The gap sat between them.

Staff had no single place to see the highest-value next action across the funnel. Leads, opportunities, scheduling, renewals, and support each lived somewhere different. Following up on anything meant assembling context by hand, which slowed response times and let opportunities go cold.

The existing automations ran on fixed rules. Fixed rules can send an email on day three. They cannot tell you that one deal is close to won and another lead needs attention this morning.

That is a prioritization problem. Adding another platform makes it worse, because it adds another place to look.

2. Three conditions made building cheaper than buying

Custom software carries a reputation for cost and risk. That reputation was earned in a different era. Three conditions changed the calculation here.

AI-accelerated development compressed the build. Our team builds with AI assistance across specification, implementation, and review. Work that once took weeks now takes days. The result is a platform delivered in roughly a fifth of the time a conventional build would need, and the saving goes into the quote rather than the margin.

Phased delivery removed the big-bang risk. Nothing was rebuilt at once. Each phase absorbed one more business process, went into use, and informed the next phase. Language Advantage never faced a cutover weekend or a period of running two systems in parallel.

Keeping what worked kept the scope honest. We did not replace Klaviyo. We did not replace Demiks. Both do their jobs. Replacing them would have doubled the build and delivered nothing the client asked for.

A conventional CRM transformation would have cost five times as much and taken five times as long. Neither of those numbers is a discount. They are what happens when scope stays honest and the build moves fast.

3. What we built

The system sits between the marketing platform and the delivery platform as the decision and execution layer.

• An AI-prioritized action queue. A Next Best Action dashboard opens with a morning briefing and a timezone-aware queue. Staff see the highest-value work first instead of deciding where to start. The system generates per-lead recommendations, enriches lead and company records automatically, and flags deals that look likely to close.
• A relational CRM core. One contact can hold several independent opportunities, which matters in a business where the same organization buys training repeatedly. Each opportunity carries BANT qualification and a live 0 to 100 lead score. Pipeline stages are configurable in settings rather than hard-coded, with win rates tracked per stage.
• Email as an execution layer. Transactional email runs through Resend in both directions. The AI layer resolves routine messages on its own and surfaces the exceptions as tickets in one inbox. Klaviyo stays where it belongs, handling marketing and lifecycle messaging, which keeps consent and ownership unambiguous.
• Delivery integration with guardrails. Demiks remains the source of truth for delivery. Every session and schedule change routes through a single scheduling engine and passes a server-side state machine with an audit trail. The two systems cannot drift apart, because there is only one path between them.
• An operations console built around exceptions. Every automation run becomes a ticket. Successes close themselves. Anything that fails or looks unusual surfaces for a person. Staff review the exceptions rather than monitoring the successes.
• Access without seats. Instructors submit availability and lifecycle paperwork through tokenized links that require no login and no licence. Every submission is tracked as an automation run. The people outside the core team participate in the system without appearing on an invoice.

Supporting pieces include a public lead capture form with UTM tracking and an embed generator, an AI renewal scoring engine, and a Chrome extension that pulls leads from LinkedIn in one click.

4. What we deliberately did not build

This section matters more than the last one.

We did not rebuild marketing automation. Klaviyo already does it, the client already knows it, and a worse version would have cost real money.

We did not replace the delivery system. Demiks is the system of record and the integration respects that.

We did not build a reporting suite nobody had asked for. Reporting follows once the data has settled and the questions are known.

We did not build a mobile application. A responsive interface covered the actual use.

Every one of those decisions removed weeks from the schedule and dollars from the invoice. Scope discipline is the largest cost lever on any custom build, and it is the one most often given away.

The fastest way to overspend on custom software is to build the parts nobody asked for.

5. The economics of owning the system

Subscription software never stops costing money, and the cost rises with headcount. A custom platform costs the most in year one and then costs maintenance.

Three effects compound.

• Platform consolidation. Several point tools collapsed into one system and their subscriptions ended, cutting recurring software spend by roughly 60 percent.
• Costs decouple from headcount. Adding operations staff adds no per-seat fee, and the instructor network participates without licences at all. Growth stops being a line item.
• Integration work disappears. There is no longer a middle layer of connectors and manual exports to maintain, because the integration is the product.

Taken together, the build pays for itself in roughly 14 months against the subscription and manual coordination costs it removed. After that the system is an asset rather than a bill.

6. The harder part is not the software

The technical work was the predictable half of the project.

The other half was operational. Which processes should move first. Where a rule should end and human judgment should begin. What the data model has to represent about a business that sells training in packages, delivers it in sessions, and renews it on cycles. Those questions belong to the client, and answering them well takes time and attention from people who already have full days.

Tom and John spent that time with the Language Advantage team rather than around it. Phases moved at the pace of real understanding.

Tom and John have taken the time to understand how our organization operates and have guided us through the technical and operational decisions involved. We have learned a great deal throughout the process and continue to learn as the system evolves.

Where this pattern fits, and where it does not

Building beats buying when a business has outgrown the assumptions in off-the-shelf software, when work moves between three or more platforms that were never designed to talk, and when the people who need access outnumber the people you want to buy licences for.

Buying still wins when a standard product genuinely fits, when the process is not a competitive advantage, or when nobody internally has capacity to make decisions about their own operations. That last condition disqualifies more projects than the first two.

Language Advantage met the conditions. Several platforms, an elastic workforce, a process specific enough that no vendor sells it, and a founder willing to engage with the decisions.

In the client's words

We have had a very positive experience working with HeitechSoft, led by Tom, along with John and their team. They have been helping Language Advantage develop a customized CRM and automation system designed around our operational needs.

The project has been implemented in phases so that the system can progressively incorporate our different business processes.

We appreciate their expertise, patience, and collaborative approach, and we are pleased to recommend HeitechSoft to organizations looking to improve their systems, automation, and operational efficiency.

Results and impact

Consolidated several platforms into one owned system, cutting recurring subscription spend by roughly 60 percent. Delivered in roughly a fifth of the time and cost of a conventional CRM transformation. Around 12 hours of manual coordination returned to the team each week. Instructors participate through tokenized links that require no licences. Automation exceptions collected in a single inbox instead of being dropped silently. Delivery data kept consistent and audited through server-enforced state transitions.

---

Paying per seat for people who barely use the software? Talk to us about what consolidation would look like for your operation.