Around 2002-2004, the CCM market went through a generational transition. Until then, building customer communications meant scripting: code written by specialists, in proprietary environments. Then a small number of vendors introduced visual, WYSIWYG, low-code composition, and the market reorganized itself around them. One or two players made that transition early, captured the opportunity, and defined the market structure that lasted twenty years. Those who stayed with scripting did not disappear overnight. They amortised a declining model, project after project, until the question was no longer whether to move, but how much the delay had cost.
Composition by Prompting is the same kind of transition, one level up. What WYSIWYG was to scripting, prompting is to low-code: the interface between human intent and platform execution moves up another level of abstraction, from code, to visual design, to natural language. It builds on low-code rather than replacing it: the visual and low-code layer remains the foundation, while prompting becomes the faster and more intuitive way to drive it.
The same market dynamic applies. The players who make this transition early will be well positioned to shape the next twenty years of CCM. Continuing to invest in a legacy model based on WYSIWYG design may well resemble the choice made by those who kept investing in scripting in 2004.
Every CCM professional knows this translation chain: A business requirement document is written by a line of business. The CCM team reads it, interprets it, asks for clarifications, and rewrites it as specifications. The specifications are then translated into templates, business rules and data structures using a proprietary design interface, by people with rare skills within the organization. Tests follow, then production.
Most of the friction lives within the early stages: between what the business meant and what the platform was told to do.
The market has learned to live with it because the traditional low-code CCM model is built on these constraints: scarce skills, proprietary languages, layered validation cycles. The cost is invisible in any budget line. It is dissolved into lead times, backlogs and a bottleneck concentrated on a handful of experts.
Webinar: Composition by Prompting
Prompting is replacing low-code platforms in CCM. Find out why and how it works in a real-world example.
The real bottleneck in CCM is not document design itself.
It is the translation of business intent into technical implementation.
Compart's Composition by Prompting removes that friction at its source. Composition becomes a dialogue instead of a translation:
A project lead or an integrator describes the communication project in natural language, including its structure, business rules and data requirements.
The Compart DocBridge® Communication Suite AI Assistant generates the template, the layout logic, the business rules and the data structures in minutes, using HTML5 and JavaScript.
Everything the AI produces is editable with standard skills. The developer reviews and validate the result and remains accountable for the work produced.
Developers can return to the visual Designer at any time to adjust and fine-tune the project.
The human does not leave the process. At every step, a person reviews and validates what the AI generates before it moves forward, and nothing reaches production without that approval. The human’s changes: from translating requirements into a proprietary syntax to reviewing and validating generated assets.
By transitioning to a Composition by Prompting approach and removing much of the traditional friction in the Customer Communication Management (CCM) process, three key changes can be expected:
There is a fundamental decision that platform buyers should weigh carefully. A CCM platform is amortized over five to ten years. An investment made today in a low-code platform without a credible roadmap for Composition by Prompting commits that entire amortization period to a production model the market is starting to move away from. Year after year, templates and rules accumulate in a proprietary language that will need to be migrated later. That is technical debt in the strict sense: work contracted now, at today's cost, that will have to be paid back at tomorrow's. The right question for any renewal or replacement is therefore broader than what the platform does today. It is whether its roadmap can accommodate the transition to prompting within the expected amortization window.
Three things will remain unchanged when adopting Composition by Prompting, and they matter as much as what changes:
1. Regulatory compliance remains unchanged.
Communications produced through Composition by Prompting meet the same compliance, archival and auditability requirements as before. Every version remains documented and traceable.
2. Team size remains unchanged, but the backlog shrinks.
CCM composition teams are small, typically five to ten people, and perpetually overloaded. Composition by Prompting converts waiting time into delivered work: the same specialists finally get through the queue of modifications, variants and versions that the business has been accumulating.
3. Governance is unchanged.
Human validation remains mandatory at every stage. The process is versioned and auditable end to end. AI proposes, humans commit.
Gen AI is powerful and continues to evolve rapidly. However, some points deserve clarity:
These limits reflect the current state of the technology, but that state is advancing rapidly. Progress in generative AI and Agentic frameworks moves at a pace where capabilities considered out of reach today can become standard within a few release cycles. Compart's platform is designed to evolve alongside these advances and integrate new capabilities as they mature.
Composition by Prompting addresses the friction between the first and second links of the CCM chain. The chain has a third link: business user modifications, the interactive phase where line-of-business users enrich and adapt documents built on previously created templates.
Today, this is done through interactive interfaces, operated by human agents within the line-of-business. Tomorrow, the same logic that transformed composition will apply here: modifications suggested by AI agents based on context, within the same governance framework of versioning, traceability and audit.
This is the direction of the first phase, which is available today.
The second phase, planned for the next DocBridge® Communication Suite release in 2027, will take this process further:
For those making platform decisions today, this industry paradigm shift deserves attention. The 2002-2004 transition proved decisive precisely because CCM platforms are amortised over five to ten years: the choices made during that window locked in advantages and constraints that shaped the market for a decade.
A platform choice is never a decision about where the market stands today, but about where it will be over the life of the investment. For every renewal or replacement decision in the coming years, the question is no longer whether this shift will happen, but on which side of it your next five-to-ten-year commitment will land.
The era of translation between business intent and technical realization is ending. The era of prompting in CCM has begun.
