The least glamorous software in a company often has the most serious job. It pays people. It closes the books. It remembers which student registered for which class. It approves the purchase order and records the invoice. Nobody gathers around the office coffee machine to admire an ERP, but plenty of mornings can be ruined when one stops behaving.
Arvind Rajan has spent his career in that stubborn, consequential territory. He is the co-founder and CEO of Astute Business Solutions, a California firm that helps organizations run, move and modernize enterprise systems including PeopleSoft, Oracle E-Business Suite and Ellucian Banner. His work has taken him from programming and application hosting to cloud infrastructure, managed services and, lately, a campaign to put practical AI inside the workflows that institutions already use.
The word practical matters. Rajan's public talks are less interested in a machine's ability to write a sonnet than in whether it can read an invoice, find the matching purchase order, flag an exception and pass the result into the right financial system. His version of the future wears sensible shoes. It has permissions, audit trails and a meeting with procurement at four.
01 / OriginsClose enough to hear the servers
Rajan began as a programmer analyst at BPL Infosystems in 1994. He then worked as a systems administrator and application developer at Michigan State University while completing a master's degree in computer engineering. A lead developer and systems analyst role followed at Data Systems Network Corporation. By 1999, he was managing application hosting services at PeopleSoft.
It is an unusually coherent apprenticeship for a future chief executive. The early roles put him close to code, infrastructure and the people waiting for both to work. Hosting also teaches an unromantic truth: the customer experiences technology as reliability. Architecture is elegant only if payroll still runs.
Astute grew from that technical foundation. Rajan's company biography describes him designing highly available, scalable systems for large enterprises and mid-market customers, then carrying responsibility across the customer lifecycle: negotiation, planning, design, development, testing and global rollout. It is a broad span, but the connective tissue is operational. He knows what happens after the proposal becomes a project and the project becomes somebody else's Monday morning.
02 / The cloud yearsModernization as careful renovation
For many enterprise customers, moving to the cloud was never a fashionable reset. It was a logistical escape from aging data centers, slow provisioning and the delicate rituals required to keep complicated software alive. Rajan became a persistent guide to that passage. In webinars and a 2020 PeopleSoft Chat appearance, he explained the case for moving PeopleSoft workloads onto Oracle Cloud Infrastructure: better automation, improved performance and a path out of owning every layer of the machinery.
His approach resembles renovation more than demolition. First, inventory what exists. Then understand security, capacity, compliance, networking and cost. Decide what can move as it is, what needs to be rebuilt and which risks are too important for improvisation. The language is sober because the stakes are. A university does not become more innovative by misplacing its student records.
“There is no substitute for human interaction and live in-person discussions.”Arvind Rajan, reflecting after Alliance 2025
A 2022 customer conversation offers a glimpse of the payoff Rajan likes to emphasize. Astute's team described a global nonprofit upgrade that increased instance deployment by 90 percent through PeopleSoft Cloud Manager and automated the PeopleTools upgrade process. The point was not cloud as a destination. It was time returned to the organization once the repetitive infrastructure work became less demanding.
03 / A new layerWhen AI meets the general ledger
The arrival of generative AI changed Rajan's vocabulary, though not his underlying method. Astute's current demonstrations cover invoice capture, supplier onboarding, contract management, requisitions, HR support, student help desks and natural-language questions over enterprise data. Rajan frames AI as a layer that can make established systems easier to use rather than an excuse to discard them wholesale.
That distinction is particularly useful in ERP. A system of record is designed to preserve structured truth. An AI interface is good at interpreting an untidy request. Put the two together carefully and an employee might ask a finance question in ordinary language, retrieve an answer grounded in approved data, or complete a task without learning the geography of a thirty-year-old menu.
The temptation, of course, is to let the fluent interface disguise uncertain data. Rajan keeps returning to the less cinematic work beneath it: clean information, governance, access control, evaluation and a specific business case. In the podcast he now hosts, the recurring questions are about implementation and return, not merely possibility. Which task is being improved? Who reviews the output? How does a pilot become a dependable service?
Those webinar numbers reveal a market full of specific anxieties. People do not ask 27 live questions when the answer has no consequence. They ask because they have a backlog of invoices, a supplier process that takes too long, or an ERP investment that still works but feels increasingly difficult to extend.
04 / Show the workThe productive danger of a live demo
Rajan has a visible fondness for demonstrations. After an AI-for-finance and procurement session in 2025, he wrote candidly about technical trouble and the value of post-production. He also noted that the presentation contained almost no slideware beyond a welcome and speaker introductions. The rest was demo.
A live demo is a small act of corporate courage. Slides live in a universe where networks are stable, sample invoices are well behaved and every button understands the assignment. Software has other ideas. Rajan appears to value that exposure because it turns an abstract claim into something customers can interrogate. It also lets the awkward questions surface while the builders are still in the room.
“Yes, I am always excited at the prospect of a live demo.”Arvind Rajan
At Alliance 2025, where Astute presented AI sessions for education and government, Rajan found the audience interaction more educational than the formal program. He left with at least half a dozen use cases that might become prototypes. There is a compact operating philosophy in that anecdote: bring working ideas to customers, listen closely to the friction, and allow the next product decision to come from the conversation.
05 / The through-lineA founder who still thinks like an operator
Rajan's public footprint now spans articles, ebooks, webinars, conference sessions, a podcast and a Slideshare archive that reaches back to 2011. The formats keep multiplying because enterprise technology needs translation. A finance leader, database engineer and university administrator may be looking at the same system and seeing three entirely different problems.
Recommendations on his professional profile describe someone able to pair business value with technical expertise, make complex challenges understandable and offer several workable options. One former colleague calls him outgoing, smart and funny. The humor is worth noting. Anyone who voluntarily spends decades around ERP has either developed perspective or an excellent hiding place.
The aspiration stated in Astute's biography is to help customers strategize, plan and execute their journeys to cloud. His more recent work extends the destination. Cloud made the infrastructure more flexible. AI, in his telling, can make the work itself more responsive. Yet the goal remains less about novelty than useful change: reduce a manual step, answer a question sooner, catch an exception and give an employee time back.
There is something reassuringly unfuturistic about the bet. The enterprise of tomorrow will still have invoices, contracts, employees, students, auditors and software with a long memory. Rajan is working on the connective layer between them. If it succeeds, the achievement may not look like magic. It may look like a task completed correctly, a queue shortened, and one less person spending Friday afternoon hunting through a menu built before the iPhone.