Impetus Technologies launched Forward Builders, its enterprise AI practice, on September 23, 2026. The practice merges product, design, and engineering into a single role. Every engagement gets one professional who carries an AI system from idea to production. Impetus says the model formalizes work its engineers have done inside client operations for more than two years.
That work includes flagging fraudulent transactions, approving claims, rerouting shipments, and repricing products.
The bet is organizational, not technological. The old handoff chain — a product manager prioritized, a designer shaped, an engineer built — worked for software that followed deterministic rules. It works less well, Impetus argues, for systems that reason, decide, and act with their own agency. Forward Builders puts one owner on every system. Each builder arrives knowing which decisions the AI can make alone, which need a human, and which a regulator will ask to audit.
CEO Nachiket Deshpande’s framing is blunt: “An architect never hands a blueprint down a line mid-build.” The practice launched with 100-plus certified builders, qualified through a five-stage internal standard, and plans to double by the end of 2027.
The timing is easy to read. Impetus named Deshpande, an AI-business president from LTI Mindtree, as CEO in November 2025. It launched its Leap AI platform family in May 2026 and became an OpenAI Select Partner on September 17. Two weeks ago, Deloitte launched a strikingly similar talent model in the same markets. The enterprise AI practice space has suddenly become crowded with much larger players pushing the same thesis.
The Competitive Picture for the Enterprise AI Practice Model
Forward Builders enters a market where every large integrator already has an agentic AI story. The fight is over talent models and context, not models.
The giants are converging on the same thesis
HFS Horizons’ 2026 agentic services assessment names Accenture, Cognizant, IBM, Infosys, and Wipro as its leaders. Accenture reports more than 100 agents in production and 600 more in pilots, anchored by its Amethyst platform.
Its “Intelligent Digital Brain” architecture argues that “the next moat is not just data; it is proprietary context” — the same thesis Impetus markets as Context Fabric. Cognizant sells Agent Foundry with context engineering at its center. Infosys counters with Agentic Foundry, a library of pre-built agents across finance, healthcare, and manufacturing. None of them need Impetus to teach them context engineering.
The forward-deployed talent race
On September 9, two weeks before this launch, Deloitte announced its Open Model Engineering practice in India and other global markets. The plan: hire, train, and certify forward deployed engineers through fiscal 2027. The language matches Impetus’s almost line for line — a model made famous by Palantir, now adopted across the consulting industry. Impetus’s twist is folding product, design, and engineering into one certified role rather than just embedding engineers on site.
The context-engineering consensus
Impetus has been genuinely early on context engineering. Its CEDL methodology and Context Fabric platform predate most rivals’ messaging. But consensus cuts both ways.
Gartner told AI leaders that “context engineering is in, and prompt engineering is out” back in July 2025. At its 2026 data and analytics conference, Gartner analysts called the context layer critical infrastructure, with 44 percent of data leaders already running semantic layers and 48 percent more planning them by 2027. Being right about the trend no longer differentiates anyone.
What the Data Shows
Impetus is private, so there are no financial filings to consult. Public profiles put the company at 2,000–3,000 employees, founded in 1996. It is headquartered in Los Gatos, California, with staff across 22 countries and roughly $350 million in total funding from backers including Kedaara Capital.
Its website claims 3,000-plus data and AI engineers and more than $200 billion in enterprise data assets migrated. Named Fortune 100 clients include ADP, American Express, Bank of America, Lilly, United Airlines, and Verizon.
The OpenAI partnership announcement credits Impetus with more than 30 production-grade AI use cases across Fortune 500 enterprises. Its own practice numbers are more modest: 100-plus certified builders is roughly 3 percent of the claimed engineering bench. Doubling to 200 by end-2027 keeps it a small fraction.
The company’s richest public proof point is a sponsored theCUBE interview from May 2026. Deshpande described an airline lost-baggage agent that improved from ten to eight-and-a-half mishandled bags per unit — once it learned which airports use partner staff rather than their own. The number is company-reported, the segment was sponsored, and the improvement is honest about being gradual. That candor is worth more than most vendor case studies.
One more detail: the release carries an Indore, India dateline, though Impetus is headquartered in California. The launch is being pushed hardest in the Indian market, where the AI services talent war is fiercest.
What’s New vs. Repackaged in Impetus’ Enterprise AI Practice
Genuinely new
The single-role org design is the real product. Collapsing product, design, and engineering into one accountable owner is a differentiating talent model. Certifying that person through a five-stage standard sharpens the claim, though it remains untested at scale. The practice brand itself is new, as is the explicit commitment to being accountable for traceability and cost-per-decision on every engagement.
Improved
The Leap platform integration tightens what existed before. Context Fabric now feeds governed context to agents at sub-250-millisecond latency through MCP servers, with pre-built ontologies for finance and healthcare. Agent traceability — which agent acted, on what information, with which tools — is a real operational capability, not just messaging.
Repackaged
The delivery work is more than two years old by the company’s own admission. Forward Builders is the third AI-services packaging Impetus has launched in roughly two years. GenAI Innovation Labs arrived in August 2024, and the Leap AI family followed in May 2026. Each relaunch re-covers the same underlying engineers and platforms with sharper branding.
Unclear
What the five-stage certification covers, and who audits it, is not disclosed — it is an internal standard. How “accountable for both” translates into contracts is also unstated. The release does not say whether outcome guarantees, penalty clauses, or shared-risk pricing back the accountability claim.
The Questions That Weren’t Answered
Accountability comes first. The practice says builders are accountable for traceability and cost-per-decision on every engagement. What does that mean in the contract? Is pricing tied to outcomes, or is accountability a service-level promise? No vendor in this category has answered that cleanly, and the release does not try.
Key-person risk comes second. One owner per system eliminates handoffs — and creates a single point of failure. What happens when a builder rotates off an account? Does the engineered context stay with the client in transferable form, or does it depend on the person? The CEDL methodology suggests documentation and runbooks, but the release is silent on handover.
Scale comes third. Can 200 certified builders by 2027 move transformations at Fortune 500 clients whose programs run for years? The model concentrates expertise in scarce individuals; that is its pitch and its bottleneck. Utilization plans and leverage mechanisms are not disclosed.
Last, platform lock-in. Forward Builders runs on the Leap family. Ask what happens to the context layer, ontologies, and agent traceability data if you leave — the same exit question analysts now tell every buyer to put in RFPs.
What This Enterprise AI Practice Means for You
If you are evaluating AI system integrators, the delivery model behind any enterprise AI practice now matters as much as its platform. Ask any vendor — Impetus, Deloitte, Accenture, Cognizant — three questions. Who exactly owns your system end to end? What does their certification cover? What contractually happens at handover?
The vendors with real answers to those three questions are the ones betting on delivery, not just branding.
If you are building in-house, the one-owner idea is cheap to copy. Analysts observing successful deployments keep landing on the same role: data product managers. They own the agent, its data, its policies, and its evaluation metrics end to end. You do not need an external practice to install that accountability structure internally.
If you are an Impetus customer or prospect, the foundation underneath the branding is real: the OpenAI Select partnership, 30-plus production use cases, and a context-engineering platform that predates the current hype. Treat Forward Builders as packaging around that foundation — and price the engagement on the foundation, not the label.
Editor’s Note
This article draws on Impetus’ September 23, 2026 press release, company announcements from November 2025 and May 2026, the company’s OpenAI partnership announcement of September 17, 2026, and public materials from Deloitte, Accenture, Cognizant, Infosys, HFS Horizons, and Gartner-adjacent reporting. Employee, funding, and client figures are company- or profile-reported and not independently verified. The theCUBE interview was a sponsored segment. No services were evaluated hands-on.