Is Generative AI tipping the scales in favor of building Enterprise Content Management (ECM) software, or will it ever get to that point? Indeed, the prospect of having code appear before your eyes is enticing. However, relying on Generative AI for coding projects comes with pitfalls that organizations must consider.
Among these pitfalls are customization limitations. Generative AI can produce generic code very quickly. However, it may also need help creating code that meets specific or complex requirements. In addition, AI-produced code can be inconsistent and produce varying levels of quality depending on the input provided and the particular AI model used.
Reliability issues and security vulnerabilities are some of the most significant considerations when using AI to help build in-house software. Generative AI can produce functional code, but it may contain subtle bugs that are difficult to detect. Regarding security vulnerabilities, Generative AI may inadvertently introduce attack vectors because it doesn't inherently understand the security practices or the specific security requirements of developing an ECM application.
Even with the advent of Generative AI, organizations must carefully consider the decision to build or buy software, especially regarding ECM applications.
ECM Build Vs. Buy
Many organizations use monitoring tools such as Splunk, NewRelic, Dynatrace, Solarwinds, Nagios, Zappos and others. However, regarding ECM, AI will be challenged to develop and maintain truly effective monitors that offer deep visibility and context.
As with any software development, readiness, scope, resources, opportunity, cost and time-to-value are all elements that must be entered into the decision equation. Before embarking on building your own ECM solution, consider the following:
1. Development Knowledge Gap and Costs — Building ECM monitors requires proficient developers who deeply understand ECM system architecture, interfaces and stack. Finding these individuals requires salaries, benefits recruitment costs, and hardware and software infrastructure. Companies must also consider the staff to produce and maintain such monitors because the software requires a team with specialized knowledge of multiple ECM systems and monitoring best practices.
2. Development Time and Effort — ECM development timelines can range from several months to years. Anyone considering developing an ECM solution in-house needs first to consider these requirements:
- Obtain Software Development Kit(s) (SDK) for each ECM platform.
- Develop or obtain developer knowledge of SDK(s) for each ECM platform.
- Develop detailed requirements and project plans.
- Develop or obtain developer-level knowledge of ECM platform stack, interfaces and operating behavior.
- Assign Project Manager.
- Design, Build and QA the following: Tests Monitors Conditions/Thresholds Displays Dashboards Reports Remediations/Automation Notifications
- Move through a DevOps process to the production state.
3. Ongoing Maintenance and Updates — Maintenance for the newly developed ECM solution will require resources to monitor the monitors, fix issues and introduce enhancements. Not carefully considering the maintenance costs will strain IT budgets and consume much of your staff’s time. It’s also important to note that these maintenance issues grow as ECM evolves with new features such as operating systems, databases and application services.
4. Scalability and Risk Challenges — Significant modifications are often required when attempting to scale a custom monitoring solution, and incremental funding for changing circumstances, such as supporting intelligent automation technologies like RPA and other ECM platforms or automating ECM recovery actions, may not be available.
Deciding whether to build or buy software requires a thorough analysis of cost, flexibility, scalability and compliance. In addition, the risk profile of a custom ECM monitoring solution increases over time compared to a proven "off-the-shelf" solution. This risk is especially true when the custom ECM monitoring solution authors are no longer in the same roles.
Even if organizations consider using Generative AI to help create an internal ECM monitoring solution, the facts still point to buying a solution as the more effective outcome. AI does help companies with the time, human resources and funds needed to build software solutions. However, when considering building ECM monitoring solutions in-house, the pros and cons analysis tips the scale in favor of the buy scenario — especially since developing practical ECM monitoring solutions is challenging for many in-house IT personnel.
Today, proven ECM monitoring solutions offer chargeback, capacity planning, content security, and many integrations to ensure an out-of-the-box solution at a fraction of the cost and time it takes a developer to produce.
Brian DeWyer is CTO and Co-Founder of Reveille Software. With more than 25 years of experience in technology, Brian DeWyer provides product strategy and technical leadership in his role as Reveille CTO and board member. Brian leverages his extensive knowledge from his tenure as a senior IT leader at Wachovia and previous role as a process consulting practice leader for IBM Global Services delivering on-premises and cloud-based solution implementations for Fortune 1000 commercial and government clients. He has led process change efforts within large organizations, building on content-driven solutions for high-volume transaction processing applications. He is a past board member of the Association of Image and Information Management (AIIM) industry association. Brian graduated from Virginia Tech with a BSME and holds an MBA from Wake Forest University.
