Enterprise AI, Data, & Application Services | Improving

## Improving: Enterprise AI, Data, and Application Services

Improving is an enterprise technology services firm with 21 offices across 7 countries and 3 continents. Their core work sits at the intersection of artificial intelligence, data engineering, application modernization, and cloud platform delivery. The company positions itself explicitly as an alternative to what it calls the trust problem in the technology industry, the pattern of overpromising, understaffing, and scope creep that buyers in enterprise IT know well.

That framing is not just marketing. The company backs it with a reported NPS score above 90, which is unusually high for a technology services firm, and an average partnership length of 7.5 years. Those two numbers together say something real: clients are not rotating them out after the first engagement. Whether you weight those figures heavily depends on how you feel about self-reported metrics, but the client list (Norwegian Cruise Line Holdings, Berkshire Hathaway Energy, McKesson, Medtronic, Lakeshore Learning, Simpson Strong-Tie) gives them grounding. The claim that 28 Fortune 50 companies use Improving is notable if it holds up in a reference call.

### What They Actually Do

**Artificial Intelligence.** Improving owns the full AI lifecycle: strategy and planning, training and enablement, agent creation and deployment, and machine learning built on proprietary client data. They use a proprietary 8-stage IOS maturity model, which they position as a way for clients to know exactly where they are in an AI program and what comes next. That kind of structured framework matters in AI engagements because most failures are organizational and process failures, not technology failures. The company cites Lakeshore Learning as a case where an agentic AI system went from pilot to production in 90 days, and Simpson Strong-Tie as a $1.6M machine learning engagement delivered in production. Both are attributed directly on the Improving website and are worth pressing on in a discovery call to understand scope and team structure.

**Data Engineering and Platforms.** Improving builds data pipelines, governance infrastructure, and analytics platforms. They claim the most certified Confluent data streaming engineers in the Americas as of 2025 and were named Confluent's Enablement Partner of the Year. Real-time data pipelines for Medtronic are cited as a production example. For organizations trying to use AI seriously, data readiness is usually the bottleneck, and Improving's positioning here is coherent: they are not selling AI on top of a fragile data layer.

**Application Development and Modernization.** The applications practice covers AI-enabled development, legacy to cloud-native modernization, DevOps and platform engineering, enterprise integration and APIs, and ongoing production support. The 7.5-year average engagement length is especially meaningful here: monolith decomposition and multi-year modernization programs require a vendor who can hold context across long timelines. Improving's own website cites a multi-year partnership with Norwegian Cruise Line Holdings and 60% faster deployment for Berkshire Hathaway Energy as named examples from this practice.

**Cloud and AI Platform Infrastructure.** Improving designs, builds, and operates cloud, data, and AI infrastructure across the three major hyperscalers: Microsoft Azure, Google Cloud Platform, and AWS. The company's website cites 70+ certified Kubernetes engineers and a 10x faster deployment speed improvement for McKesson as examples of platform delivery outcomes. They also hold a 2023 Google Cloud Partner of the Year award for the Public Sector, which is independently verifiable through Google's partner program.

### Who Fits This Vendor

Improving is not a staff augmentation shop or a low-cost offshore body shop. Their model is closer to a long-engagement technology partner for mid-to-large enterprises with complex AI, data, or application modernization problems. The client examples skew toward regulated industries (healthcare devices, energy, financial services) and large consumer businesses (cruise lines, learning companies). If you are a startup or early-stage company, this is probably not the right fit. If you are an enterprise with a multi-year modernization program or an AI initiative that has stalled at pilot stage, they are worth a serious conversation.

### Delivery and Engagement Model

With 21 offices across 7 countries and 3 continents, Improving has a global delivery footprint. Specific city locations are not detailed in available public information, so confirm delivery locations and timezone coverage for your team during the evaluation process. The engagement model appears to favor long-term dedicated partnerships over short discrete projects, which has practical implications for procurement: expect a retainer or multi-year contract structure rather than a simple project bid.

### Compliance and Certifications

No specific compliance certifications (SOC 2, ISO 27001, HIPAA, etc.) are called out in currently available public information. For buyers in regulated industries like healthcare or financial services, this is a direct question to put on the intake call. Given the Medtronic and McKesson relationships, it is reasonable to assume some compliance posture exists, but I would not assume anything not explicitly confirmed.

### Editor's Take

The combination of a 90+ NPS, 7.5-year average engagement, and named Fortune 50 clients is a credible signal that Improving delivers real value to a specific type of buyer: large enterprises with complex, long-duration technology programs. The trust-first positioning is easy to say and hard to fake across a 7.5-year average, which gives it more weight than the usual vendor pitch.

That said, I would verify three things before shortlisting. First, ask for references from engagements that had rough patches, not just success stories. Any vendor with 7.5-year partnerships has had hard quarters. How they handled those is the real signal. Second, clarify pricing structure and what all-in costs look like across a full engagement, including staff transitions and knowledge transfer if you ever want to bring work in-house. Third, confirm the compliance and security posture in writing before any regulated data touches their systems. The client names suggest the posture is real, but in regulated industries, "it is reasonable to assume" is not a sufficient answer.