Newfire Partners

10/22/2026

The gap between planning and ex*****on is one of the biggest barriers to successful business transformation.

While upfront planning can help reduce risk, too much discovery can drain budgets and delay progress. Organizations don't need to resolve every unknown before development begins.

Learn how to make discovery more effective in our latest blog:

More planning isn't the same as less risk. See how focused discovery and early validation protect your budget and your delivery window.

10/01/2026

Recent statistics show a projected 10% rise in software developer employment from 2025 to 2035. As AI improves coding speed, many leaders believe adding more engineers to their team will further increase delivery speed of products. However, research shows that when it comes to improving delivery, prioritizing IT infrastructure and clearly defining roles and responsibilities have the biggest impact.

Read the full blog to learn how a growing organization can look more capable on paper while becoming slower in practice:

More engineers doesn't mean faster delivery. See why organizational design — not headcount — determines how fast software teams actually ship.

09/10/2026

74% of companies still haven't realized meaningful business value from their AI investments. One myth may be getting in the way: faster code means a better product.

AI can help teams write code more quickly, but architecture and engineering judgment still determine whether that code becomes a reliable product.

The teams that pair AI with these practices will be the ones that create lasting advantage.

Read our first blog in the Missteps series, The AI Shortcut Myth, here:

Generative AI writes code faster, but speed isn't a product. Why software architecture and engineering judgment decide what becomes durable software — and what becomes debt.

09/01/2026

"We need to move faster, so let's start coding."

It sounds like the right decision. But many of the choices that derail technology transformations start with good intentions, responding to pressure for speed, lower costs, or quick wins.

That's the idea behind The Missteps, our new six-part series exploring decisions that can undermine transformation. Drawing on real-world experience, each blog looks at how choices around priorities, accountability, operating models, and ex*****on shape business outcomes.

Read the series introduction from our CEO Steve Hau here:

Technology transformations fail through reasonable decisions that conceal hidden mistakes. Explore six missteps that consistently undermine even well-intentioned initiatives.

08/14/2026

When a clinical AI system fails, the instinct is to blame the model or rewrite the prompt.

According to John Puopolo, that framing misses the core issue. The model is rarely the problem, and prompt engineering is not enough.

In healthcare, AI systems operate across complex EHRs, lab systems, and pharmacy records. The real discipline is context engineering—a system-level architecture that governs data retrieval, tool permissions, session state, and clinical boundaries at runtime.

When a model gives a poor recommendation, it isn't a reasoning failure. It is the system doing the best it can with incomplete, stale, or missing data. In clinical environments, context engineering is a fundamental patient safety mechanism.

Read John's full breakdown on why healthcare must move beyond prompts to build true context-first AI architecture: https://hubs.li/Q04t1-Wj0

07/24/2026

When a large health plan built its first AI feature, the real risk wasn't whether AI could help care teams. It was whether clinicians would trust it enough to use it.

What our UX research uncovered changed the entire roadmap—and one finding became the north star for the whole product.

We break down what care teams actually need to trust AI in high-stakes workflows and how it led to a first AI investment de-risked before major spend.

Read the case study: https://hubs.li/Q04qFqkG0

06/26/2026

Before making an AI investment, it’s important to identify the signals that matter, amplify the right work, and turn readiness into business value.

Most technology-led conversations about AI focus entirely on automation as it relates to model selection, licensing fees, and output speed. Unfortunately, without a product roadmap clearly aligned to business value and an engineering operational review, automation will only amplify the underlying friction within unscalable workflows and legacy processes.

Our Strategic Advisory Lead, Patrick Schmid, mapped out how to stress test your delivery workflows before committing your technology budget, pulling directly from our internal implementation data.

Read the full operational breakdown here: https://hubs.li/Q04mQ6NX0

05/22/2026

Newfire teams across Europe, the Americas, and APAC simultaneously marked our 10th anniversary.
With engineering hubs across three continents, our strength is delivering complex technical solutions under any conditions. Celebrating together across time zones proves the unified culture powering this global ex*****on engine.
As AI accelerates change, we are proud of the framework we've built to guide clients through the next wave of tech innovation. We are ready for the next decade.

05/15/2026

We’re excited to welcome Mike DeSimone to Newfire as our new Chief Delivery Officer.

Mike brings 30 years of experience navigating the realities of healthcare operations and leading Fortune 500-scale transformations. He has a track record of turning $40M+ P&Ls around and guiding organizations through complex M&A, always with a focus on translating high-level strategy into ex*****on that actually moves the needle.

In an industry where the gap between ambition and results is wide, Mike’s leadership ensures our delivery stays grounded in measurable business outcomes.

Learn more: https://hubs.li/Q04gMZGt0

05/12/2026

The gap between a promising AI demo and a reliable production agent usually comes down to the data foundation.

As John Puopolo explains in our latest article, most platforms are built for humans to look backward at reports. AI agents need to look forward to reason and act. This requires a shift from static storage to runtime context.

When data platforms strip away sequence and provenance, agents lose the ability to provide traceable evidence. If your agents are stalling, the problem is likely your architecture rather than your model.

Read the full breakdown: https://hubs.li/Q04ggmVC0