Profisee sells the software equivalent of a stern but useful family archivist. It looks at five versions of the same customer, notices that two addresses are stale and one surname is misspelled, then decides which details deserve to survive. The result is a "golden record" - a trusted version that can feed billing, analytics, supply chains and, increasingly, AI agents that do more than answer questions.
The company arrived at this job by an unusually circular route. Ian Ahern founded Stratature in 2001 to build master-data software. Microsoft acquired the technology in the mid-2000s and turned it into SQL Server Master Data Services, or MDS. Ahern and others from that lineage founded Profisee in 2007. Now Microsoft has removed MDS from SQL Server 2025, and Profisee is pitching a migration utility while operating as the native master-data workload inside Microsoft Fabric. The old road ends almost exactly where the newer company set up shop.
One answer to a very expensive noun
Every company has important nouns: customer, supplier, product, location, patient, employee. Trouble begins when the CRM, ERP and data warehouse disagree about them. Profisee ingests those records, applies validation and matching rules, merges duplicates according to survivorship logic, and lets stewards review ambiguous cases. It then sends governed records back to the systems and teams that need them.
This is multidomain master data management, not merely a customer-data platform. Profisee handles products, suppliers, assets and reference data as well as people and companies. The platform includes data modeling, quality rules, workflows, relationship visualization, address verification and audit history. It can run as Profisee-managed SaaS on Azure or as a containerized platform managed by the customer in a cloud or on premises. Connect supplies low-code links to Fabric, Snowflake, Databricks and Azure Synapse. The Application Edition handles reference data and MDS migration; the Enterprise Edition adds full matching, merging and survivorship.
“There is finally hope.”WSP's Chief Client Officer, after the company gained a global view of customer relationships
The proof is hiding in the duplicates
The clearest Profisee story belongs to WSP Global, an engineering and advisory firm assembled through acquisitions. Regional offices and business lines often served the same client without knowing it. WSP first considered building its own global customer master, then decided the route was too slow and would not scale. It evaluated vendors and chose Profisee for speed, flexibility and scale.
The important part came before launch. WSP invested heavily in preparing real regional datasets and testing matching rules. That work turned 73,000 active accounts into 43,000 unique customers. Ninety-one percent of relationships were resolved without a steward touching them. In 13 weeks, its customer data health score rose from 52.1 percent to 87.6 percent. The company also discovered that more than half its active customers had relationships across multiple regions or business lines - useful intelligence for cross-selling and credit risk, not a tidier database for its own sake.
From account pile to client map
Lexmark offers the commercial version of the lesson. Remote work weakened the old strategy of selling more printers, so the company focused on placing and supplying existing devices better. Its unsupported, highly customized legacy MDM system could not flex without a large new investment. Profisee linked customers, printer locations and third-party service vendors, then sent the cleaned relationships into Databricks forecasting models. Lexmark's business team estimated $2 million in combined North American revenue and savings in one year.
A nonprofit healthcare system shows what failed first. Its SQL and Python scripts for matching patient and payer records were inaccurate and inefficient. With a dedicated Profisee instance, erroneous matches fell 60 percent, staff saved five to ten minutes per case and claims throughput rose from roughly three to 12-14 per hour. Here, a fuzzy duplicate was not an annoyance. It delayed eligibility and prior authorization.
The customer list also shows how the same mechanism travels. Domino's began with a hard assignment: match, merge and enrich about 85 million customer records, then extended MDM to store locations and supply-chain items. Aker BP used Profisee with Fabric, Purview and implementation partner Avanade to deliver 70 consumption-ready data products in less than 12 months; it identified another 70-90 for the queue. YMCA of the USA used it to standardize staff data across 763 branches. The nouns change, but the pattern does not: first resolve identity, then expose relationships, then let an operational or analytical system use the answer.
What it costs - and what buyers forget to count
Profisee does not post a dollar price. Its quote has three levers: software edition, total mastered-record volume and deployment. Domains and attributes are unlimited; volume is counted across domains. SaaS bundles hosting and licensing, while a customer-managed deployment shifts infrastructure and cloud expertise back to the buyer. Profisee says its customers manage more than 100 million golden records and promotes implementations completed in under 90 days.
The license is only part of the bill. Buyers must fund data preparation, integration, match-rule testing, governance meetings and ongoing stewardship. WSP's case is instructive because it says the quiet part plainly: the company invested heavily before production. A platform can automate 91 percent of relationship decisions only after people define a trustworthy rule for the other nine percent.
There is also a timing question. Profisee markets quick implementations, but "available in minutes" describes a hosted environment, not a finished operating model. Connecting Salesforce is not the same as deciding whether Salesforce wins when its address conflicts with SAP. A useful business case therefore counts the first release, the people who own exceptions and the downstream systems that must accept the result. The easiest budget to defend is tied to a measurable leak - rejected claims, late supplies or unseen client relationships - rather than a general ambition to improve data quality.
Microsoft-native, but not Microsoft-only
Profisee's sharpest difference is distribution, not a checkbox. Its workload appears inside Microsoft Fabric's interface, where teams can model and steward data in the environment already holding their analytics. It integrates with OneLake and Purview, runs on Azure and reaches Microsoft 365 Copilot and Power Platform. The company's ancestry makes the fit less like a recent alliance and more like a long family argument that finally produced sensible architecture.
That position is valuable but specific. Informatica sells a broader data-management suite. Reltio emphasizes a cloud-native data platform. Semarchy, SAP, Oracle, Ataccama, Boomi and TIBCO each bring their own ecosystem or usability case. A Microsoft-centered enterprise can treat Profisee's Fabric embedding as a shortcut. A company organized around SAP or Salesforce should compare the native gravity of those alternatives. Profisee can connect across stacks; its greatest advantage still appears when Microsoft is already the center of the room.
The likely buyer is not a startup with one database. It is a mid-market or large enterprise with acquisitions, regional systems, regulated records or enough scale that duplicates now affect revenue and risk. Data stewards work in the platform, but the beneficiaries are broader: finance gets cleaner hierarchies, marketers get a customer view, supply teams get reliable locations, and data scientists spend less time wrangling identity. Profisee sits between integration and use. It does not replace the source applications or the warehouse; it makes their shared entities agree.
AI changed the budget, not the underlying chore
In 2025 and 2026, Profisee put an AI face on its long-running plumbing. Aisey, its assistant, now orchestrates specialist agents for modeling, integration, matching, quality and stewardship. Profisee says an administrator can upload a requirements document and receive a configured MDM environment in about ten minutes, with roughly 90 percent of setup performed automatically and human review at judgment points. Its MCP Server lets Copilot, Claude, ChatGPT, Gemini and custom agents read or write governed records through authenticated, audited access.
Those claims matter because an agent acting on bad data creates a different class of mistake. A chatbot that confuses two suppliers is embarrassing. An autonomous workflow that reprices the wrong contract or routes a shipment to a stale address costs money. Profisee reported 45 percent year-over-year new-sales growth in the first half of fiscal 2027, after a 37 percent annual recurring revenue compound growth rate at the end of fiscal 2026. The figures are company-reported, but the buying logic is plausible: AI budgets are absorbing a job once filed under cleanup.
The limit is worth stating. Generated models and semantic matching can accelerate configuration, but they cannot invent a company's definition of a household, choose which legal entity owns a contract or decide when two near-identical patients must remain separate. Public user reviews generally praise Profisee's configurable workflows and quality rules, while some users describe setup, navigation or large match clusters as difficult. Automation reduces mechanical work. It does not remove the need for experienced modeling, current training and a support plan.
What teams can copy
Choose one costly decision, one dirty data domain and a named owner. Test matching on real records. Measure the before state. Automate only after stewards trust the exceptions.
When it will not work
No software can settle ownership nobody accepts. A vague "clean everything" mandate, absent business sponsor or no stewardship capacity will turn MDM into an expensive holding pen.
The playbook is smaller than the sales pitch
A reader does not need to copy Profisee's product to copy its method. Start where a bad record causes visible pain: denied claims, missed shipments, duplicate outreach or cross-sell blindness. Limit the first release to the sources required for that decision. Put business owners beside data engineers. Test on the ugliest data available. Track the cost or time before and after. Only then add domains.
It will not work under every condition. A small organization with one clean operational system may not need enterprise MDM. A company unwilling to name authoritative owners will merely centralize its disagreements. Match automation also deserves caution when false merges carry clinical, legal or financial consequences. Even Profisee's newest agentic release keeps human review where judgment matters.
The company itself is privately held, backed by ParkerGale Capital since 2017 and majority investor Pamlico Capital since 2021; transaction values were not disclosed. Len Finkle is chief executive. Profisee describes a distributed "Village" culture built around growth mindset, constructive candor, helpfulness, getting things done and making fun. Employee surveys put it on the 2026 Top Workplaces for Remote Work list. The supplied company data places headcount around 190, while Profisee reported 20 percent workforce growth in fiscal 2026.
Profisee's wager is that corporate truth should be a maintained product, not a heroic spreadsheet repair. That idea was easy to postpone when bad data only made a dashboard suspicious. It becomes harder to ignore when software agents are allowed to act. The comeback is not really about replacing Microsoft MDS. It is about making sure the next machine in the chain knows which customer everyone means.