Cognizant has agreed to buy an Australian SAP services firm with more than 300 employees, and the quality of the mining and utility records that firm maintains will constrain the artificial intelligence built on top.
Records that make an SAP system useful at a mine were typed in by maintenance planners long before anyone said the word agent. Cognizant announced on October 22 that it has agreed to acquire DyFlex Solutions, a Perth-based SAP Platinum Partner, and those records are where the AI promise in the announcement will be tested.
Cognizant's analyst relations team sent me the announcement, and DyFlex was a new name to me. PiLog and SAP HANA, the High-performance Analytic Appliance database under SAP S/4HANA, are familiar from my own experience, which is why the master data angle caught my eye first. Congratulations to the DyFlex team: Consultancy.com.au counted about 150 staff in February 2025, and Cognizant's announcement counts more than 300.
Cognizant grew up running and rebuilding other companies' applications
Cognizant's own company snapshot says it originated from expertise in data and applications, and the model has held for three decades: consultants near the client with a large delivery workforce behind them, and acquisitions to fill a gap in a platform or an industry. Second-quarter 2026 revenue was $5.5 billion on a headcount of 356,700 (Cognizant, 2026).
Cognizant now describes itself as an AI builder. In practice that means its own platforms, including Neuro AI and an SAP agentic platform, and partnerships with model and cloud vendors. I covered the hiring plan and the Odyssey Logistics rebuild earlier this month in Cognizant to Hire 1,500 U.S. Graduates; Odyssey Logs 37 Percent Rebuild Cut, so this post stays on the SAP side.
In the second quarter, Snohomish County Public Utility District in Washington State selected Cognizant to move its on-premises SAP to SAP's cloud and to add Cognizant's asset lifecycle accounting solution. DyFlex does comparable work for Australian resource companies: according to a DyFlex case study published on Mastering SAP, it implemented SAP S/4HANA Public Cloud for Kimberley Mineral Sands. Cognizant says it wants to move clients beyond isolated experiments to production-ready agentic systems across their SAP environments, which is the goal this pairing is meant to serve.
Cognizant's reporting shows little of the region it is buying into
Consultancy.com.au reported in May that Cognizant's regional business has more than 1,500 professionals across eight locations, including a delivery center in Adelaide. Carla Ramchand took over as managing director after Rob Marchiori left for Westpac. DyFlex's more than 300 employees include a delivery presence in the Philippines, so the figure does not translate into Australian and New Zealand headcount one for one.
Cognizant does not report Australia and New Zealand on their own. North America supplied 75.3 percent of second-quarter revenue, while Rest of World, the line that holds the region, supplied 6.0 percent and shrank 1.2 percent (Cognizant, 2026). In my read, a company that earns three-quarters of its revenue in one region has reason to look for growth elsewhere.
SAP's 2027 maintenance deadline concentrates demand for experienced people
SAP's mainstream maintenance for ERP Central Component 6.0, its older enterprise resource planning product, ends on December 31, 2027 for the later enhancement packages, with paid extended maintenance available through 2030. That is a support deadline, not a shutdown date, and each customer chooses how to respond.
Basis Technologies analysis reported by iStart in July 2024 found that only 57 percent of ERP Central Component customers were on track to finish migrating to S/4HANA by the 2027 date. Local sources told iStart that Australian and New Zealand buyers had resisted S/4HANA more than buyers in North America and Europe. If that hesitation persisted, regional demand would concentrate closer to the deadline.
Cognizant calls DyFlex one of the leading employers of SAP talent across the region. In my read, that makes the deal as much about capacity as capability.
Mining and utility master data describe different things and break in similar ways
Start with mining, resources, and energy. In SAP Plant Maintenance, a site becomes a tree of functional locations, such as a crusher or a conveyor. Planners install equipment records at those locations and remove them when a unit is swapped or rebuilt. Bills of materials connect each machine to its spare parts, and every spare part is a material master record.
Mobile equipment adds churn: a haul truck or a drill can change position, borrow parts from a sister unit, or return from a rebuild with a different configuration, and each change needs a record update. A 2017 SAP Community post by Dean Fitt ranks enterprise asset management master data among the highest concerns of customers in asset-intensive industries, citing multiple systems of record and asset information that changes over a long period. PiLog lists data quality and governance tooling for equipment and maintenance plan records on its SAP partner page, so tooling for the problem exists.
Picture an agent that orders a replacement bearing from the records it finds. If three records describe that bearing, it could choose the one with no stock or a stale price, and unless approval or exception controls catch it, the choice surfaces only when the part fails to arrive. That is a hypothetical risk, not a reported incident, and it is the reason the controls around an agent matter as much as the agent.
Utilities organize master data in two families. SAP's utilities model separates business master data, such as the business partner and the contract, from technical master data, such as the installation and the meter, and links the two through the installation and the contract. Device locations correspond to functional locations, the same object a mine uses for a crusher. In a utility, a mislinked installation becomes a wrong bill and a customer call. At a mine the cost takes another form: a machine waits in the workshop while someone hunts for a seal.
Migration alone does not resolve these data-quality problems. SAPinsider's guidance on the 2027 deadline says duplicate or incomplete records in the old system move into S/4HANA and amplify problems downstream, and that teams struggle to map legacy data structures to the newer data model.
DyFlex has built for this work. It acquired Bluetree in August 2025, which added SAP Datasphere and SAP Analytics Cloud practices to its S/4HANA and SuccessFactors work, and it markets a Cloud ERP Mining package on RISE with SAP S/4HANA Cloud. Each of those engagements starts with equipment and material records that someone has to map, clean, and own.
Migration alone does not resolve these data-quality problems.
DyFlex sold independence, and customers should ask what the new structure changes
Five V Capital's February 2025 investment came with a stated goal of building Australia's leading independent full-service SAP partner, according to TechPartner.news. In August 2025 DyFlex managing director Peter Lander called the firm a "dynamic and flexible alternative to the global systems integrators," CRN Australia reported.
Cognizant brings delivery scale and an AI platform portfolio. Customers lose a mid-size independent to price against the global integrators, which matters to anyone running a 2027 migration tender.
Cognizant's announcement says clients will keep working with the people they know. It is silent on integration plans and on DyFlex's SAP partner status after closing, which is common at the signing stage. DyFlex chief executive Evan Mantis spent 18 years at SAP in Australia and the United Kingdom, according to ARN, and that relationship is worth asking Cognizant to preserve.
Financial terms were not disclosed.
If DyFlex or Cognizant staffs your SAP program, who will own your equipment and material records after the first quarter of 2027, and what is the duplicate rate on them today?
Before closing, ask your account lead for two things in writing: the named person accountable for master data quality on your account, and the current duplicate count for material and equipment records. Utility buyers should ask the same about installation and contract records. An AI roadmap that arrives without both numbers is built on records nobody has measured.
