A cabinet manufacturer had two software histories to reconcile. After ACProducts and Masco Cabinetry combined into Cabinetworks Group, one side used Microsoft 365; the other still had SharePoint 2013 on premises. There were also Nintex forms and workflows to carry forward. Keeping them meant keeping another license bill. SoHo Dragon saw an opening in tools the customer already owned.
- Microsoft consulting, custom applications, managed support, and technical staffing.
- Named clients include ADP, Cabinetworks Group, Emerald Built Environments, and 1010data.
- The recurring idea: connect existing tools, remove duplication, and help employees use the result.
A license bill with a second life
SoHo rebuilt the manufacturer’s forms and workflows with Power Apps and Power Automate. According to its case study, the change eliminated a $20,000 annual Nintex license expense. That figure describes a recurring bill removed. The work itself involved a migration, a redesign underneath the forms, and several pieces of custom functionality.
The interesting choice was what the consultants kept. Tabs stayed familiar. Fields appeared in the same order. A person filling out a long form would recognize the route. Behind that familiar surface, XML and JSON differences required translation, approval reminders needed a separate flow, and delegation required custom components. Familiarity was part of the implementation.
Annual Nintex license expense removed after rebuilding workflows in Power Platform.
This is a useful way into SoHo Dragon. Its business sits between the software a company buys and the work employees actually finish. SharePoint, Azure, Power BI, Teams, and Microsoft 365 provide the ingredients. SoHo supplies assessments, application development, integrations, migrations, and support. The recurring problem is organizational friction expressed as a technical task.
The connector is the product
Consider Emerald Built Environments, a construction and sustainability management business. Its deal information lived in HubSpot. Its conversations happened in Teams. SharePoint held important documents. Employees needed current deal notes across all three, including when they were working away from the office. Buying another destination for information would have added another place to check.
SoHo initially planned to use a webhook to open a Teams channel when a deal was created. The company’s account says that approach did not work in HubSpot. Custom code replaced it. Over five weeks, the team connected the deal information across the applications, with weekly stakeholder meetings to keep the implementation tied to what Emerald needed.
The company also sells a HubSpot/MS Teams Sync Engine, publicly listed at $100 per tenant per month for unlimited users, deals, and deal notes. Its defined job is two-way note synchronization. A bespoke three-system integration involves a broader scope. The attraction is easy to understand: people can contribute to the record from the application they already have open.
A consultancy you can call back
Founder Peter Ward started SoHo Dragon in a spare bedroom in Long Island City in 2010, initially with a SharePoint practice. The company’s timeline tracks an India office in 2015, London in 2017, Azure in 2018, and Murex and Archer practices in 2021. In 2024, it records a move to Hoboken and a focus on Azure and AI.
Today, its partners page identifies Microsoft’s Cloud Solutions Partner - Modern Work designation, alongside relationships with Orchestry, Murex, Archer, and Snowflake. Its leadership page names Ward as CEO and Tosha Desai as CTO. The breadth matters: a bank’s risk system, a document library, and a cloud deployment can require different specialists within the same engagement.
“The people you meet at the start will be the same people who will deliver the solution.”
SoHo Dragon / About Us
That is the firm’s stated distinction in a crowded consulting market: access to experienced practitioners and continuity through delivery. Its revenue model combines project work, recurring support, staffing and direct placement, training, and software subscriptions. Buyers can hire a team to finish a defined job, bring in individual talent, or retain operational help.
The data has to survive the move
ADP’s PDF project shows how product and consulting meet. The payroll and human-capital-management company wanted multiple bookmarks, direct imports into SharePoint, and the ability to save pieces of a PDF as separate documents. SoHo customized its markup tool, hosted it in ADP’s Azure environment, and implemented single sign-on between SharePoint and Azure. The published engagement lasted three months.
Another case involved approximately 70 terabytes in Azure Data Lake Storage Gen1. Moving the files would leave a second problem: existing U-SQL workloads could not simply move with them. SoHo used Azure Data Factory for migration, converted queries to Databricks SQL, and checked the destination against the source using row counts and checksums. The deliverable included continued access to usable data.
There is a practical limit to the savings argument. Existing Microsoft subscriptions only help when the required features and permissions fit the work. Replacing a license can create custom components that someone must maintain. A buyer should compare that continuing responsibility with the bill being removed, and agree how both functionality and data will be checked.
The Archer work at an unnamed international bank makes the maintenance question tangible. SoHo’s audit found unused applications and nearly 100 unused data feeds slowing the risk platform. The bank accepted a cleanup and cloud roadmap, then brought SoHo back to lead implementation after encountering roadblocks. It is an instructive sequence: diagnosis can establish a direction while execution still requires help. For a buyer, the useful discipline is to name who owns the changes, how dependencies will be retired, and what evidence will count as a working system.
Teach the people who will teach everyone else
At market-intelligence company 1010data, SoHo trained eleven power users through ten two-hour Power BI sessions over three months. Participants came from product, customer engagement, and engineering. They finished by working with the company’s own data; recordings and guides supported further learning. The plan was to equip people who could help colleagues after the formal sessions ended.
Its current AI work extends into readiness workshops, custom Copilot agents, and agent management. The video library lists AI meetups through September 2026. For a prospective client, the older projects offer a concrete starting point: inventory what you own, identify where work gets stuck, preserve useful habits, and test the result with the people who will depend on it.