AI can improve product development by helping teams identify issues earlier, evaluate products from new perspectives, prototype ideas faster, and explore more possibilities before committing development resources. At , we experienced all of these benefits when we used Claude Fable 5 to review a new product before Microsoft Marketplace certification and later to prototype the next version of our configuration experience.
But our experience also reinforced an important point: getting value from AI requires more than simply adding it to the development process. Business leaders need to choose the right AI model for the assignment and combine its capabilities with human experience, technical expertise, customer knowledge, and accountability.
Key Highlights
- ForgeXRM used AI to uncover unresolved product issues before Microsoft Marketplace submission.
- AI can provide another perspective during product reviews without replacing human judgment.
- Different AI models are better suited to different types of product development work.
- The lowest-cost AI model may not be the least expensive choice if it misses an important issue.
- AI can accelerate prototyping while keeping product vision and accountability with people.
How Can AI Help Improve a Product Before Release?
AI can provide an additional layer of product review by examining the overall experience, questioning assumptions, identifying patterns, and highlighting areas that may not work together as well as intended.
We experienced this while preparing a new ForgeXRM product for Microsoft Marketplace certification.
Development was complete. The requirements had been addressed. On paper, the product was ready for submission.
But something about the overall experience still did not feel finished.
A product can be technically complete while still having issues with clarity, usability, or the way individual elements work together. Rather than submit the product and address any remaining concerns in a future release, the ForgeXRM team decided to conduct another end-to-end review.
This time, we used to evaluate the solution from beginning to end.
The review took hours rather than seconds. The result was also substantially different from the quick prompt-and-response interactions people often associate with AI.
It identified meaningful issues rather than simply suggesting cosmetic improvements.
More importantly, several of those findings aligned with concerns we had already sensed but had difficulty articulating. Of five issues we had been wrestling with, the AI-assisted review helped us resolve three. We are still working through the other two.
For ForgeXRM, that was a practical demonstration of where AI can add value to product development: not by making the release decision for us, but by helping us better evaluate the product before we made that decision.
Finding those issues before submission was far better than discovering them after Microsoft certification had begun or, worse, after customers were using the product.
Does AI Replace Human Judgment in Product Development?
No. Our experience at ForgeXRM showed us that AI can strengthen product development decisions, but it does not replace the experience, customer knowledge, collaboration, technical expertise, or accountability of the people responsible for the product.
The AI model did not know our customers the way our team does. It did not understand every development decision that had been made. It would not have to support the finished product.
Most importantly, it did not own the decision about whether the product was ready.
We did.
What AI provided was another perspective with the capacity to examine the entire solution, question assumptions, identify patterns, and explain where the overall experience did not completely hold together.
That distinction matters for business leaders considering how AI should fit into their organizations.
AI does not have to make the decision to improve the quality of the decision.
It can help experienced people see a problem more clearly, consider alternatives they may not have explored, and arrive at a better-informed conclusion.
Why Does AI Model Selection Matter?
Different AI models are suited to different types of work. One lesson we have taken from our AI work at ForgeXRM is that model selection should be based on the complexity and importance of the assignment rather than assuming all AI models will produce roughly equivalent results.
We sometimes talk about AI as if models are interchangeable: choose whichever tool is convenient, ask the question, and expect approximately the same outcome.
That has not been our experience.
Some models are well suited for fast answers, writing assistance, summaries, and everyday problem-solving.
Other models are better suited to complex assignments that require more context, sustained reasoning, and deeper exploration before reaching a conclusion.
Anthropic describes Fable 5 as designed for ambitious, long-running projects and complex work. That made it particularly interesting to us for the type of comprehensive product review we wanted to conduct.
We had already experimented with Fable 5 and learned some of its strengths. What changed during this review was the importance and scale of the assignment we were willing to give it.
Instead of asking isolated questions, we provided the context of the product, asked the model to work through the experience from beginning to end, and gave it the time required to conduct a deeper review.
That experience reinforced something we believe will become increasingly important as companies incorporate AI into product development:
Should AI Model Cost Determine Which Model You Use?
AI consumption cost should be one consideration when selecting a model, but our experience suggests it should not be the only one. Business leaders should also consider the potential cost of using a model that is not capable enough for the assignment.
The cost of an AI interaction is relatively easy to see.
The cost of a missed product issue is much harder to calculate.
A missed issue might eventually result in:
- Additional development work
- Another round of testing
- A delayed release
- Increased support time
- A customer encountering a problem that could have been addressed earlier
In our case, a deeper AI review helped identify issues before the product was submitted to Microsoft.
Even if that review consumed significantly more resources than a quick AI review, the additional cost was small compared with the potential cost of revisiting the product after certification had begun or customers had started using it.
The same principle applies to prototyping.
A more capable model may cost more during a long working session. But if it helps a team explore multiple concepts, eliminate weaker directions, and give developers a clearer idea of what should be built, the overall economics can look very different.
This does not mean companies should always use the most powerful or expensive AI model available.
For simple work, speed and efficiency matter. For consequential work, depth may matter more.
Instead of asking only, “How much will this model cost to run?” we have learned to also ask:
“What could it cost us to use a model that is not capable enough for this assignment?”
How Can AI Accelerate Product Prototyping?
AI can accelerate product prototyping by allowing people to test ideas, question assumptions, change direction, and make concepts tangible before handing them to a development team.
We saw this shortly after our product review.
Ryan Plourde, owner of ForgeXRM, spent nearly twelve straight hours using Fable 5 to prototype ideas for a second version of our configuration experience.
Our current configuration screens already represent significant thought, design, and development from the ForgeXRM team. They are moving through Microsoft's certification process, and we are proud of where they landed.
The AI prototyping session was about what could come next.
Ideas could be tested, questioned, modified, or abandoned without losing the momentum of the original thought. Concepts that might previously have remained rough notes could become tangible enough for the development team to evaluate and improve.
The AI did not provide the product vision.
It provided a faster way to work through that vision.
For ForgeXRM, that is an important distinction in how we are approaching AI-assisted development. We are not looking for AI to independently decide what our products should become. We are using it to help experienced people explore more possibilities and bring better-defined ideas to developers for discussion, refinement, and implementation.
Can AI Help Software Teams Move Faster Without Sacrificing Quality?
AI can help teams move faster by identifying issues earlier and accelerating exploration, but speed should be measured across the entire product lifecycle rather than only by how quickly a product reaches release.
Software companies face constant pressure to move quickly. Release the first version. Get it into the market. Learn from customers. Improve it later.
There is real value in that approach. Products do not improve by remaining indefinitely in development.
But speed can also become an excuse for shipping issues a team already suspects are there.
There is a difference between releasing a focused first version with reasonable limitations and releasing a product before the overall experience has been adequately considered.
At ForgeXRM, the objective is not to release the largest number of features in the shortest possible time.
We want to build products that are repeatable, supportable, and able to perform in real customer environments.
That requires input from product, development, customer experience, and support. It also requires knowing when to move forward and when additional work is justified.
Delaying our Microsoft submission did not initially feel efficient. Development was complete, certification was waiting, and the team was ready to move on.
But submitting the product would not have eliminated the unresolved concerns.
It would only have moved them into a future release or into the hands of a customer.
In this case, slowing down may have been the faster decision in the long run.
What Is the Best Role for AI in Product Development?
At ForgeXRM, we see AI's most valuable role as complementing the people responsible for building the product. AI can provide breadth, persistence, rapid exploration, and additional analysis while people provide context, technical expertise, customer understanding, judgment, and accountability.
That combination is becoming part of how we build.
AI can help us explore more product possibilities, question assumptions, identify issues earlier, accelerate prototyping, investigate ideas that time or resource constraints might previously have kept out of reach, and bring better-defined concepts and problems to the development team.
People still decide what belongs in the product.
Developers still determine what can be responsibly engineered and supported. Our team still needs to understand the customer, evaluate tradeoffs, and recognize when an answer might be technically correct but practically wrong.
And we remain responsible for what ultimately ships.
This fits into the broader philosophy behind . We believe the technology foundation matters, but the value comes from how that technology is used to solve real business problems.
AI is increasingly becoming another part of that equation.
How Should Business Leaders Adopt AI for Product Development?
Business leaders should approach AI as a capability that needs to be matched thoughtfully to the work rather than as a single tool that should be used the same way for every task.
Based on what we have learned at ForgeXRM, five principles are shaping how we approach AI in product development:
- Match the model to the assignment. Routine work may not require the same AI capabilities as a consequential product decision.
- Consider the cost of a weak answer. Consumption cost is only one part of the economics. A missed issue can create development, testing, support, and customer costs later.
- Give AI sufficient context. A comprehensive product review is different from asking a series of isolated questions.
- Use AI to expand human capabilities. AI can help experienced people explore ideas and articulate concerns faster rather than removing them from the process.
- Keep accountability with people. AI can contribute analysis, but the team still needs to decide what is realistic, responsible, supportable, and ready to ship.
These principles are not a finished AI playbook. They are lessons we are developing as we put AI to work on increasingly meaningful assignments.
How Is AI Changing the Role of Experienced Product Leaders?
One of the more unexpected lessons from this experience was not about AI technology itself. It was about how AI can change the way experienced people contribute to product development.
For Ryan, AI did not make his role less relevant.
It gave him more opportunity to focus on the parts of the process where his experience was most valuable:
- Recognizing when something does not feel right
- Asking the next question
- Evaluating tradeoffs
- Connecting ideas that may not initially seem related
- Imagining what a product could become
- Working with developers to determine what is realistic
- Choosing the right tools and models for the importance of the assignment
- Deciding when a product is ready to leave the team's hands
During the nearly twelve-hour Grid Control prototyping session, AI made it possible to stay deeply involved in product design while exploring possibilities much faster than before.
That may be one of the most important lessons ForgeXRM is taking from our adoption of AI.
The future of AI-assisted product development does not have to be about stepping away and letting AI build products for us.
It can be about giving experienced people better ways to think, explore, question, prototype, and collaborate so teams can build better products together.
Frequently Asked Questions About AI in Product Development
How can AI improve product development?
AI can help product teams identify issues earlier, question assumptions, explore alternatives, accelerate prototyping, and bring better-defined ideas to development teams. At ForgeXRM, we have used AI for both an end-to-end product review and prototyping, while keeping final product decisions with our team.
Can AI replace product managers, developers, or product experts?
AI can contribute another perspective, but it does not replace the customer knowledge, technical expertise, judgment, collaboration, or accountability of the people responsible for a product. At ForgeXRM, AI contributes to the process, but people remain responsible for what is built and released.
Should companies always use the most powerful AI model?
No. Different assignments require different levels of AI capability. Fast, efficient models can be appropriate for routine tasks, while more complex or consequential work may justify models capable of deeper analysis.
Is a more expensive AI model worth the additional cost?
It can be when the potential cost of a missed issue is significantly greater than the additional AI consumption cost. Companies should consider both the direct cost of the model and the potential business cost of an inadequate result.
How can AI help with software prototyping?
AI can make it easier to explore concepts, test ideas, change direction, and make ideas tangible before committing development resources. ForgeXRM used this approach while exploring ideas for the next version of its Grid Control configuration experience.
How ForgeXRM Is Combining AI With Human Judgment
Our experience using AI for product review and prototyping did not convince us to hand product development over to AI.
It convinced us to become more thoughtful about where AI belongs in the process.
Choosing the right model, providing the right context, asking better questions, and applying experienced human judgment can help us uncover problems sooner and explore possibilities faster.
For consequential work, the goal should not simply be to use AI as cheaply or quickly as possible.
The goal should be to use it where it can help people make better product decisions and ultimately build better products.
See how ForgeXRM is applying this approach to business applications built on Microsoft Dynamics 365 and Power Platform, including and other solutions from .
Also read:
