Microsoft's OptimAL initiative leads partners toward improved Dynamics 365 Business Central performance

Microsoft is encouraging Dynamics 365 Business Central ISVs and VARs to take a more hands-on approach to improving application performance through a new program called OptimAL.

The company publicly revealed OptimAL at Directions North America 2026, after the inaugural group of companies had worked through it, and new cohorts of development teams are already underway.

The program provides training, code reviews, AI-driven optimization tools, and opportunities for partners to work with peers on common challenges related to developing for Business Central. The goal is to help partners find and fix coding problems, improve application performance, and learn how to use AI to build and optimize software more effectively, says Kurt Juvyns, principal product manager at Microsoft. Partners understand the importance of performance, he told MSDW, but many organizations have historically focused on adding new functionality rather than revisiting and improving older code:

Every partner in the world understands the importance of why you should focus on performance. If you focus on making your code base really performant, then you can get to bigger customers and bigger niches, and you've got more capacity and your customers are [happier]. A lot of them understand it … but not a lot of partners have a good culture of changing their code that they created in the past. They typically have a culture of adding functionality instead of revising functionality purely from a performance perspective.

Tracing performance issues

When a BC system experiences performance problems, developers and support teams start investigating. For Microsoft MVP Waldo (a.k.a., Eric Wauters), a veteran BC development manager who has been involved in OptimAL, this challenge has been around in various forms for decades. In discussing his involvement in the program, he wrote:

[H]ere’s the hill I’ve been (happily) dying on for about 20 years now: most BC performance issues don’t come from the platform. They don’t come from the baseapp.

They come from our code. Yes .. you heard me ..

Outdated code .. and patterns, sloppy PTEs, POC-code that somehow ended up in production, “works on my machine”-stuff .. and then the customer goes and blames “Business Central is slow” 🤷‍♂️.

Many such problems are eventually traced to coding choices partners made years earlier, Juvyns said. Microsoft believes the issue is becoming more important as D365BC environments grow more complex. Many applications now connect to multiple third-party products, cloud services, and integrations, making it more difficult for partners to understand how their code performs as systems expand. Improving performance can help partners take on larger projects, serve more customers, lower operating costs, improve customer satisfaction, and better prepare their applications for AI and automation. 

But many partners lack the visibility needed to understand how their applications are performing, according to Juvyns.

The world is becoming more complex. There are ISV solutions, custom extensions, integrations. … [And] not a lot of partners fully understand how their code is scaling. They don't have the numbers. A lot of partners still feel that they're sitting in front of a cockpit of a Boeing 747 and they see a lot of gauges and meters, but they've got no idea whether the plane is going up or going down. They're busy serving the customers in the back of the plane instead of looking at all of this. And then you've got the situation when it breaks, the blame game starts. And that is something that we wanted to fix.

To better understand the issue, Microsoft worked long-time consulting partner Ciellos and a group of D365BC partners to launch a pilot version of OptimAL. Their work showed that many companies were making similar performance-related coding mistakes. Ciellos uses proprietary automated code analysis tools that identified thousands of areas that could be improved, including inefficient database queries, poor use of indexes, and a lack of caching. Partners that participated in the pilot reported performance gains like faster page loading and shorter processing times. 

Matt Abbott, EVP at Ciellos, explained that the firm is a non-selling partner (they sometimes describe themselves as a “partner for partners”) that regularly helps other ISVs and partners with development work and other technical needs. That role positioned them well to develop this offering with Microsoft. 

Whether they're ISVs or resellers, partners need to zero in on what matters. Microsoft was trying to figure out, okay, how do we provide a program that really does not only teach and coach and bring up the level of capability of partners, but also do so in a way that is very hands-on, tangible, results oriented? That's where we came together at the end of calendar year 2025 and started really formulating this program with, I would say, a really unique combination of components

The final mix of elements included remote training sessions along with discussion, homework, in-class sharing, and technical consultation for each participating company. Ciellos used its own performance-focused tools for the technical assessments. Abbott explained:

[We are] using telemetry data and AL profiler information to really zero in on the areas that matter, and then we can provide two levels of recommendations. We can provide very detailed code-level recommendations. And on the strategic side we discuss how the product or the PTE (per-tenant extension) was built. So we’re coaching and guiding, but then we're giving recommendations to each participant in the program, and they have to do the refactoring. So we're helping them to learn by doing. 

The program costs $12,900 USD and includes the full assessment, the optimization work, online training, and other sessions.

By offering more than just tradition training, Juvyns says the results exceeded expectations:

The results that we got from this collaboration … were super, super good. When we looked over the patterns that affected performance in a negative way, we saw that it was all the same. Everybody made the same mistakes. We applied the [most important] 10 practices in our own booking code and ran it on top of 1 million records. By just removing the worst practices, the process became about 30 times faster. So our partners started fixing these things and then they saw these results.

Pilot participants reported significant performance improvements, according to Microsoft. For example, Microsoft shared that Binary Stream cut the processing time for a key batch job from more than five hours to less than 20 minutes, while PrintVis reduced page-load times by 93%. Other companies reported performance improvements of between 40% and 76%.

Following the success of the pilot, Microsoft formalized OptimAL into a larger program designed to help partners improve their performance skills, find and fix important coding issues, and build repeatable processes for optimizing applications. Abbott estimates that a hundred or more companies could go through OptimAL in the next year. They are building a North American cohort that they expect to launch in the summer, with plans for another Europe-based cohort aimed for autumn. 

An entry point for agentic development

What started as a program focused on improving application performance has also become a way for Microsoft to help partners learn AI-assisted software development. During the pilot, many participants used AI tools to review code, find performance problems, and make improvements more quickly.

Juvyns says performance provides a practical starting point for learning agentic engineering techniques:

We're going to build that skill, that capability so that partners can do it themselves in the future using the latest technology. … If you understand what you want to do with agentic engineering, then you don't try to solve the whole world with [it]. But you have this one thing in mind, which is performance, then you're going to start seeing other opportunities on what you can do elsewhere with agentic engineering capabilities.

Microsoft is also using what it is learning from the program to improve future D365BC quality efforts, according to Juvyns. The common coding mistakes identified during the pilot have been added to guidance for the broader D365 community and will help support future AI-powered development and code quality tools. 

Ciellos’s Abbott credits Microsoft with formulating the program into something unique for Business Central development teams. 

I give Kurt a lot of credit for how he's been thinking about the program because these components have really come together in a way that is, I think, pretty unique in terms of these types of programs. There is this combination of learning, hands-on doing, and a spotlight [on your work]. And the results have been awesome.