A refinery mixer is a useful antidote to an airy conversation about artificial intelligence. It turns. It wears. If it fails, a software recommendation suddenly has consequences measured in equipment, production and money. In a May 2026 account, Enterprise Minds specialist Dr. PG Madhavan describes applying the company’s TwinARC technology to just such an asset. The interesting part is what happens before anyone touches the controls.
- The work: custom software, SAP systems and AI automation.
- The distinctive bet: causal digital twins that let engineers rehearse an intervention.
- The lesson: test the proposed fix, then let people validate it in the real operation.
The company’s proposed sequence is straightforward: collect industrial sensor histories, investigate what is driving instability, and simulate a change. Plant specialists then implement and physically validate the selected fix. Madhavan calls the approach “Fix Before Fail”. That short phrase carries a bigger ambition than sending a warning. It asks software to help choose an action.
Knowing what happens next is only half the job
A prediction can leave an operator with an awkward second problem. Suppose the software says a machine is approaching failure. Which setting should change? What else will that change disturb? A busy dashboard can describe trouble beautifully while leaving the hard decision untouched.
TwinARC’s CASO initials spell out its intended job: Causal Analysis, Simulation and Optimization. Enterprise Minds describes a digital twin as a working virtual representation of an asset, process or plant. Its version adds causal modeling to rules and statistical methods. The ambition is to examine how a system behaves when someone intervenes, rather than merely extend a trend line.
- 01ObserveIndustrial time-series data
- 02InvestigateModel possible causal drivers
- 03RehearseSimulate an intervention
- 04ValidateEngineers check the physical result
The refinery account makes the division of labor important. Software proposes and tests; customer experts bring knowledge of the equipment and verify the physical result. For a buyer, that suggests a practical evaluation: ask whether the model improves a particular maintenance decision, and how the team will check its answer against what actually happens.
The unglamorous work is the business
Enterprise Minds, founded in 2017 by Bharath Mundlapudi, is an engineering services company with software of its own. Its work includes product development, application modernization, design, cloud systems and automation. AutoX addresses agentic process automation. TwinARC addresses digital twins and causal investigation. The company also offers industrial IoT engineering and an over-the-air platform.
Follow the industrial offering downward and the vocabulary becomes refreshingly physical: programmable logic controllers, MTConnect agents and MQTT brokers. These are ways of accessing equipment, collecting machine data and moving messages. Before an algorithm can offer useful advice, somebody has to connect it to the machine. Enterprise Minds sells that work too.
The same pattern appears in enterprise applications. SAP lists Enterprise Minds as a Silver service partner. Its practice covers S/4HANA implementation and migration, analytics, integration and support. An automation project can therefore involve the underlying business system as well as the agent that operates across it. That makes finance and procurement workflows part of the opportunity, alongside factory equipment.
Its customers include startups and larger enterprises across manufacturing, healthcare, energy, banking and telecom. The company advertises more than 45 clients on its homepage. The breadth matters because its offer is partly about joining disciplines: software development, operational data, enterprise applications and AI have to meet inside the same assignment.
A consulting engagement with reusable parts
The commercial model combines engineering services with proprietary accelerators and licensing. Enterprise Minds also describes digital twins as a service and offers several delivery arrangements, including onshore, offshore and Build-Operate-Transform-Transfer. The economic idea is familiar: reuse useful engineering so each customer does not fund a fresh start.
HFS named the business a Challenger in September 2025 and described it as bootstrapped. Its profile reported two separate client outcomes: $5 million in labor-hour savings from engineering automation, and a threefold productivity gain in production tickets. These are reported case results, with different measures. They are useful prompts for a buyer’s questions, rather than a forecast for the next project.
“Simplify life through automation.”Enterprise Minds’ stated vision
My reading of its market position is an engineering partner between specialist software vendors and large systems integrators. A buyer could assemble tools and an internal team, hire an SAP specialist, or commission a larger consulting firm. Enterprise Minds’ proposition is to bring several of those capabilities together. The relevant comparison is the scope a customer needs covered.
Teaching the rehearsal
In July 2026, Enterprise Minds and the Lean Six Sigma Institute announced a joint digital-twin lab. The proposed training includes a two-hour sandbox inside Black Belt lessons and an eight-hour workshop. The partnership announcement connects TwinARC simulations with process-improvement methods, giving practitioners somewhere to test changes before deployment.
There is a cultural counterpart inside the company. Enterprise Minds describes delivery pods, mentoring, code reviews and rapid iteration. Its culture page even promotes a “20 Percent Project” for exploring interests outside ordinary work. These are stated practices, but they fit a business whose software needs people willing to investigate why a system behaves unexpectedly.
The transferable idea is modest enough to use: choose one operational decision, establish its current performance, simulate a proposed change, and assign a person to validate the result. That approach depends on usable data and a model that represents the relevant behavior. Without those, rehearsal can create confidence faster than understanding.
The mixer brings the story back to earth. An attractive prediction is the beginning of a maintenance conversation. Enterprise Minds wants to supply more of what comes after it: the integration, the investigation, the simulated fix and the engineering judgment that turns a suggestion into something a plant can use.