Achieving Operational Excellence With AI: A Breakdown of the MIT Technology Review Report

In July 2026, MIT Technology Review Insights published an executive briefing paper, Achieving operational excellence with AI, produced in association with the global digital business services company TP. Its argument is refreshingly blunt in a market drowning in AI hype: AI does not create operational excellence — it amplifies whatever discipline (or chaos) already exists.

If your processes are mature, well-governed, and data-driven, AI will accelerate them. If they are ad-hoc and fragmented, AI will accelerate the mess — faster, and at greater scale. As the report puts it, "AI is only as strong as the framework it plugs into."

This is a full breakdown of that report: the thesis, the numbers, the expert voices, and the seven best practices it lays out — plus what it means for any organisation deciding where to invest next.

The one-sentence thesis

Frameworks like Lean Six Sigma and business process management (BPM) earned their place because they brought order to sprawling operations — statistical rigour, quality control, and end-to-end maps of how work should flow. The report’s central claim is that these time-tested playbooks are not being replaced by AI. They are being upgraded by it.

Technology and process, the authors argue, are no longer separate levers you can pull independently. AI can accelerate process excellence, but existing process excellence is what makes AI truly impactful. Only organisations that pull both together realise the full value of either. Companies that already operate with discipline have an edge: they can channel new tools into proven systems rather than bolting them onto shaky foundations.

“Standards are the engine that makes all the technology and all the different processes function. If, as part of your culture, you don’t have high standards and process excellence top of mind, whatever technology you put forth as part of your framework will never work.” — Christian Buschmeier, Global Chief Standards and Process Officer, TP

The numbers that matter

The report anchors its case in a handful of figures worth committing to memory:

• $113 billion — the projected size of the AI-powered process optimisation market by 2034 (MIT Technology Review Insights, with data from Global Trade, 2026).
• 88% of business leaders expect to increase investment in AI-infused process intelligence — most of them within the next 12 to 18 months (with data from the World Economic Forum, 2026).
• Up to 50% efficiency gains from infusing AI into business process management, according to the American Society for Quality.
• ~100% of customer interactions analysed by AI in TP’s quality management — surfacing issues and sales opportunities that human agents then act on.

The takeaway is not “AI is big.” Everyone knows that. It is that the spending is flooding toward process intelligence specifically — and, without the right foundations, much of it may not deliver.

The three key takeaways

The report distils to three points:

1. Process frameworks are getting an AI upgrade

AI is automating routine tasks while simultaneously tracking KPIs, analysing data and interactions, and generating recommendations or assets. Lean principles, in particular, are finding fresh applications in digital environments — the same waste-reduction mindset that once eliminated bottlenecks on the shop floor now trims cost from computing cycles and data pipelines.

2. Without the right foundations, the investment underdelivers

Mature process disciplines and data-driven decision-making are the prerequisite, not the nice-to-have. Enterprises that skip them risk never extracting the full value of AI. This is the line the whole report turns on: AI is only as strong as the framework it plugs into.

3. Best practice is about structure, not just tooling

The organisations pulling ahead link their operational and support standards, institutionalise cascading governance, equip teams with consistent tools, and make sure the right data flows to the right people. We break these down below.

How the classic frameworks are evolving

The report leans on Jayet Moon, Chair of the Delaware Section of the American Society for Quality (ASQ), to explain what the “upgrade” actually looks like.

Intelligent BPM (iBPM)

In a traditional BPM setup — claims processing, say — you define step-by-step handling: approvals, notifications, checks. In intelligent BPM, AI engines handle those tasks automatically while tracking KPIs like time-to-resolution and error rates in real time. Moon puts the efficiency gain at up to 50%.

Lean Six Sigma in digital environments

Lean’s philosophy — waste reduction and evidence-based decision-making — is described as “more relevant than ever.” Moon’s team applied Lean thinking to cut AI deployment costs by switching off tokenisation functionality during periods when it wasn’t needed. Value stream mapping, meanwhile, gives organisations clear visibility into where value is created and where waste accumulates as workflows span multiple platforms.

Hyperautomation

The larger movement is toward hyperautomation, where AI, process mining, robotic process automation, and orchestration combine into self-improving workflows — systems that get more agile and compliant over time.

Process mining: know what actually happens before you deploy AI

One of the report’s sharpest sections features Wil van der Aalst — professor at RWTH Aachen University, chief scientist at the process intelligence company Celonis, and often called the “godfather of process mining.” Process mining reconstructs how work actually flows by reading the digital event logs your systems already generate.

His warning: a “shocking” number of organisations have no factual picture of how their own processes run — and that is precisely the picture you need before you point AI at them.

“If you have no idea what is really going on in your processes, it’s very naive to think that you can somehow solve your most important problems with AI agents.” — Wil van der Aalst

His analogy is the one to remember: asking a chatbot to diagnose workflow inefficiencies without mining the event data is like asking it to add numbers without a calculator. The real breakthrough, he argues, is marrying fact-based computation with natural-language AI — systems that don’t just generate plausible answers, but ground every recommendation in verifiable reality.

What best-in-class looks like: TP’s TOPS and BEST

The report’s in-depth example is the sponsor’s own model. TP runs two proprietary frameworks side by side: TP Operational Processes & Standards (TOPS) for frontline operations, and Baseline Enterprise Standards for TP (BEST) for support functions such as recruiting, training, and workforce management. Together they “define the rules of the game” so frontline and back-office teams operate in sync — the structure that makes it possible to deploy AI across many teams, functions, and geographies at once.

Layered on top is Lean Six Sigma, with structured yellow-, green-, and black-belt training, biweekly project audits, and an AI-powered “copilot” to guide DMAIC projects (define, measure, analyse, improve, control). Supervisors receive AI-driven next-best coaching recommendations; AI generates fair shift schedules; and AI analyses nearly all customer interactions to surface issues and opportunities that humans then refine.

Crucially, the report stresses balance between AI and emotional intelligence. As ASQ CEO Sid Bhatnagar frames it, AI is not a silver bullet — it still takes change management, root quality principles, continuous improvement, and genuine investment in people’s skill sets to build an infrastructure that can scale.

The seven best practices for the AI era

This is the operational heart of the report — the practices that turn “AI as accelerant” from a slogan into reality.

1. Link operational and support standards. Recruitment standards should be aligned with the queries frontline agents will actually face; workforce-management standards should support service-level goals. Operations and support have to work in tandem.
2. Institutionalise cascading governance. A layered cadence — monthly leadership reviews for strategy, weekly manager sessions to translate plans into action, daily supervisor huddles for immediate issues — creates clarity without heavy-handed oversight.
3. Equip teams with consistent tools. Standardised playbooks, onboarding guides, and user-friendly dashboards reduce variability, make training easier, and create a shared language for what “good” looks like.
4. Treat data like a team sport. Combine operational KPIs, quality metrics, customer feedback, and employee input into a holistic view — and empower people at every level to spot patterns, raise concerns, and propose solutions.
5. Embed change through training and smart scaling. Too many AI pilots fail to scale. Successful organisations treat deployment as an enterprise-wide journey, pairing technical rollouts with robust upskilling. Training builds confidence and helps frontline staff see AI as something to work alongside, not fear.
6. Support organisational shifts. AI demands tighter collaboration between operations, IT, and analytics. Breaking down silos and aligning process improvements with business strategy is what lets insight translate into action.
7. Prepare for the future of frameworks. Experiment with emerging tools — AI agents to support Lean Six Sigma projects, hyperautomation platforms that combine process mining, RPA, and orchestration — to keep process excellence relevant in the next wave.

The Cocoon takeaway: your framework is your people

Read closely, the report is really an argument about capability. A “framework” is not a document — it is a set of habits your people actually practise: data-driven decisions, disciplined processes, continuous improvement. AI amplifies those habits where they exist and exposes their absence where they don’t.

That is why best practice number five — embedding change through training — is not one item among seven. It is the one that makes the other six real. You cannot institutionalise governance, standardise tools, or treat data like a team sport with a workforce that doesn’t understand AI. The report is explicit that most pilots fail to scale precisely because organisations invest in technology and skimp on the human capital around it. That is exactly the gap we explore in why corporate AI training fails and moving AI training from pilot to scale.

If you want to build that muscle deliberately, our AI for Business Leaders and Enterprise programmes are designed around applied, role-specific capability — not awareness sessions that fade in a fortnight. And if you first want a factual read on where your organisation actually stands, the AI Readiness Score is a fast, honest starting point — the organisational equivalent of process mining before you deploy.

“This revolutionary technology, for all its promise, is only as sound as the framework it operates within.” — MIT Technology Review Insights

Or, in one line: buy the AI, but build the discipline. The companies that win the next phase won’t be the ones with the flashiest tools — they’ll be the ones whose people were ready for them.