Mark T. Boyer: The Integration Edge – Leading Fintech & AI Innovation

AI rarely fails in enterprises because the models are weak. It fails because the workflow is. A recent MIT-backed study reported that 95% of generative AI pilots stall with little measurable P&L impact—most often because teams never integrate the tool into how work actually gets done.¹
The companies seeing real value are treating business process reengineering with AI as an operating-model change: redesigning decisions, handoffs, controls, and data flows so AI can clear routine work end-to-end—while people move up the stack to exceptions, judgment, and client conversations.

Business process reengineering with AI is not “automation plus”

Automation improves a step. Reengineering improves an outcome.

When leaders automate a single activity (summarize a document, draft an email) but leave the broader journey untouched, the bottleneck simply shifts. That’s why the best programs start with the process objective (cycle time, error rate, abandonment, cost-to-serve) and redesign the journey around what AI can now do well: extract, compare, classify, draft, and route—at speed and scale.

McKinsey’s 2025 global AI survey found that high-performing companies are about three times more likely than others to redesign workflows while implementing genAI, and they’re more likely to invest materially—often allocating more than 20% of digital budgets to AI.² Put differently: winners don’t bolt a chatbot onto yesterday’s process maps.

Banking: routine work is being redesigned around AI

nCino: instrumenting onboarding, validation, and decisioning

nCino’s May 2025 nSight release is a crisp example of workflow-level AI in banking software. CEO Sean Desmond argued that “AI is the catalyst driving a broader transformation” and said nCino is “pivoting” R&D toward agents that “reimagine and automate every task.”³

More important than the framing are the metrics nCino published across customer segments:

• Commercial banking: onboarding time reduced “from months to days,” and one institution cut document processing time by 74%.³
• Small business banking: loan decisions accelerated 62%, and application abandonment fell 41%.³
• Mortgage lending: real-time document validation reduced documentation completion time 47% and inquiries 68%.³

These are reengineering outcomes—cycle time, abandonment, and inquiry volume—driven by AI that converts documents into data, validates completeness, and routes work with fewer back-and-forth loops.

Bank of America: AskGPS makes expert lookup near-instant

Bank of America’s Global Payments Solutions team built a genAI assistant (AskGPS) trained on 3,200+ internal documents to answer employee questions “within seconds.”⁴ The bank targets a specific workflow pain point: complex inquiries that previously took “an hour” across regions can now be handled “almost instantly,” with “the potential to save tens of thousands of hours annually.”⁴

Mark Monaco, head of GPS, described the shift from search to workflow support: “AskGPS turns institutional knowledge into real-time intelligence… It’s more than a search tool—it’s a strategic engine.”⁴

JPMorgan Chase: an internal “AI hub” and a push into agents

JPMorgan’s internal genAI environment, LLM Suite, reached 200,000 onboarded users in eight months, and the bank is now integrating the tool more deeply into internal data and workflows to support agents.⁵ Derek Waldron, JPMorgan’s Chief Analytics Officer, said: “The North Star for LLM Suite is to position it as an AI hub for employees.”⁵

In a McKinsey interview, Waldron quantified adoption at scale: “Today we have nearly a quarter-million people with access,” and “a little under half of JPMorgan employees use gen AI tools every single day.”⁶ He also made the key reengineering point: if you want real cost takeout, you must “prioritize the journey and reimagine it end to end.”⁶

Citi, Goldman Sachs, and HSBC: toward agentic execution (with guardrails)

• Citi began rolling out “agentic” capabilities to a pilot group of 5,000 employees in Citi Stylus Workspaces, aiming to streamline “multi-stage workflows into a single, automated process.”⁷
• Goldman Sachs is building autonomous agents with Anthropic for internal functions including “trade and transaction accounting” and “client due diligence and onboarding.”⁸
• HSBC signed a multi-year deal to deploy Mistral models (self-hosted) for analysis, translation, risk assessment, and client communication, with an explicit emphasis on responsible-AI governance and data protection.⁹

The pattern is consistent: banks are standardizing on controlled internal platforms, then extending from “assist” to “execute” in targeted, auditable workflows.

Retail and CPG: compressing cycle times in product and content

Retail and consumer goods organizations are applying AI to workflows where speed and creative iteration drive revenue—but brand control still matters.

Walmart: from trend to shelf in six weeks

Walmart’s “Trend-to-Product” system uses AI to accelerate product development. Axios reported that it can reduce the timeline from six months to six weeks.¹⁰ Walmart’s Andrea Albright added a constraint many executives should copy: “We don’t need speed for the sake of speed.”¹⁰ The goal is faster iteration with quality gates, not acceleration at any cost.

Zalando and Mondelez: AI content factories, human review gates

Two widely reported examples show measurable impact in marketing operations:

• Zalando told Reuters it cut image production costs by 90% and reduced campaign production time from six to eight weeks to three to four days using AI; it also said 70% of its editorial images for Q4 were AI-generated.¹¹
• Mondelez is using generative AI to reduce marketing content production costs by 30% to 50% and has invested more than $40 million in digital and AI transformation, Reuters reported.¹²

Both cases illustrate “creative operations” reengineering: brief → generate → review → localize → publish, with humans shifting toward governance and final approvals.

Manufacturing and logistics: inspection, digital twins, and exception management

General Motors: AI inside factory feedback loops

General Motors describes using AI and machine learning in manufacturing to support “digital twin” simulations and quality inspection—tightening feedback loops so issues are detected earlier in the build process.¹³ Even without publicly quantified savings, it reflects a core reengineering move: connect sensing, analytics, and decision rules so defects are prevented upstream.

UPS: centralizing exceptions yields hard-dollar savings

UPS offers a quantified operations example. CIO reports that UPS’s Gateway Technology Automation Platform (GTAP) helped identify $13.5 million in savings in 2024, with a projection of about $24 million in 2025.¹⁴ GTAP centralizes exception management and uses AI and “machine decision-making” to automate manual tasks and optimize resources.¹⁴ In high-throughput environments, exceptions are not edge cases—they are the business.

A manufacturing analogue: generative AI for industrial commercial workflows

A 2025 C3 AI case study of a global steel producer reported a generative AI assistant delivering an estimated $9 million in annual economic value, 87% response accuracy, and a 90% acceleration in “detailing time.”¹⁵ The pattern mirrors banking: document-heavy, expertise-driven work shifts from “search and assemble” to “verify and decide.”

Life sciences: discovery workflows, with validation as the constraint

Reuters reported that Takeda partnered with Iambic Therapeutics to apply AI to drug development and cited the claim that AI could reduce a preclinical stage that typically takes about six years to less than two.¹⁶ Even if that timeline proves optimistic, it highlights the discovery version of workflow reengineering: AI is inserted into iterative loops (hypothesis → candidate generation → experiment planning → results interpretation), where validation gates matter as much as throughput.

Challenges and pitfalls leaders should plan for

The risk profile of AI reengineering is largely governance and people, not “model selection.”

• Workflow integration risk. The MIT research cited by Fortune argues most pilots fail because enterprises don’t integrate tools into workflows; it also reported buying tools/partnering succeeds about 67% of the time versus internal builds that succeed only about one-third as often.¹⁸
• Workforce anxiety and adoption drag. BCG found employee positivity about genAI rises from 15% to 55% with strong leadership support—and that comprehensive AI-driven redesign correlates with higher job-security anxiety (46% vs. 34%).¹⁷
• Explainability, auditability, and validation. HSBC explicitly frames its rollout under a responsible-AI governance model to preserve transparency and data protection.⁹ In regulated domains, the workflow must include human review gates, sampling, and logging—not just generated text.
• Over-automation and creativity loss. AI can accelerate a bad decision faster than a human can. And in creative workflows, speed can homogenize output if humans don’t remain accountable for brand and differentiation—echoing Walmart’s “not speed for the sake of speed.”¹⁹

A pragmatic executive playbook

1. Start with a measurable workflow outcome (cycle time, abandonment, rework, exceptions, cost-to-serve).
2. Redesign the journey before choosing tools—remove steps, don’t just automate them.
3. Define execution vs. recommendation vs. approval points so humans retain judgment where it matters.
4. Standardize a governed AI platform layer (identity, permissions, logging, monitoring) to reuse controls across workflows.
5. Ship in narrow slices, then scale once the new workflow is stable and measured.

AI is steadily becoming a workflow substrate—especially as agentic systems mature. The leaders who win won’t be the ones with the flashiest demos. They’ll be the ones who redesign how work moves, how decisions are made, and how risk is controlled—then measure outcomes like any other operational transformation.

Footnotes

1. Sheryl Estrada, “MIT report: 95% of generative AI pilots at companies are failing,” Fortune, August 18, 2025, https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/ (accessed February 11, 2026).
2. McKinsey & Company, “The State of AI: Global Survey 2025,” November 5, 2025, https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai (accessed February 11, 2026).
3. nCino, “nCino Unveils Transformative AI-Powered Banking Solutions at nSight,” May 20, 2025, https://www.ncino.com/news/ncino-new-ai-powered-banking-solutions-nsight (accessed February 11, 2026).
4. Bank of America, “BofA’s New GenAI Assistant Transforms Global Payments Solutions,” press release, September 30, 2025, https://newsroom.bankofamerica.com/content/newsroom/press-releases/2025/09/bofa-s-new-genai-assistant-transforms-global-payments-solutions.html (accessed February 11, 2026).
5. JPMorganChase, “LLM Suite named 2025 ‘Innovation of the Year’ by American Banker,” June 3, 2025, https://www.jpmorganchase.com/about/technology/news/llmsuite-ab-award (accessed February 11, 2026).
6. McKinsey & Company, “JPMorgan Chase’s Derek Waldron on AI and banking,” October 29, 2025, https://www.mckinsey.com/industries/financial-services/our-insights/jpmorgan-chases-derek-waldron-on-building-an-ai-first-bank-culture (accessed February 11, 2026).
7. Roberto Torres, “Citi deploys agentic tools to in-house AI platform,” Banking Dive, September 23, 2025, https://www.bankingdive.com/news/citi-agentic-AI-tools-stylus-workspaces/760868/ (accessed February 11, 2026).
8. Reuters, “Goldman Sachs teams up with Anthropic to automate banking tasks with AI agents, CNBC reports,” February 6, 2026, https://www.reuters.com/business/finance/goldman-sachs-teams-up-with-anthropic-automate-banking-tasks-with-ai-agents-cnbc-2026-02-06/ (accessed February 11, 2026).
9. Reuters, “HSBC taps French start-up Mistral to supercharge generative-AI rollout,” December 1, 2025, https://www.reuters.com/business/finance/hsbc-taps-french-start-up-mistral-supercharge-generative-ai-rollout-2025-12-01/ (accessed February 11, 2026).
10. Axios, “Walmart is using AI to help it spot and respond to trends faster,” April 9, 2025, https://www.axios.com/2025/04/09/walmart-ai-trends-fashion (accessed February 11, 2026).
11. Reuters, “Zalando uses AI to speed up marketing campaigns, cut costs,” May 7, 2025, https://www.reuters.com/business/retail-consumer/zalando-uses-ai-speed-up-marketing-campaigns-cut-costs-2025-05-07/ (accessed February 11, 2026).
12. Reuters, “Mondelez uses GenAI tool to slash marketing content costs by up to 50%,” October 24, 2025, https://www.reuters.com/business/retail-consumer/mondelez-uses-genai-tool-slash-marketing-content-costs-by-up-50-2025-10-24/ (accessed February 11, 2026).
13. General Motors, “How GM is using AI in manufacturing,” March 11, 2025, https://www.gm.com/stories/how-gm-using-ai-manufacturing (accessed February 11, 2026).
14. Thor Olavsrud, “UPS transforms air cargo operations with data, AI,” CIO, July 25, 2025, https://www.cio.com/article/4026273/ups-transforms-air-cargo-operations-with-data-ai.html (accessed February 11, 2026).
15. C3 AI, “Generative AI Assistant for a Global Steel Company,” case study PDF, May 14, 2025, https://c3.ai/resources/generative-ai-assistant-for-a-global-steel-company/ (accessed February 11, 2026).
16. Reuters, “Takeda, Iambic to use AI to speed up drug development,” February 9, 2026, https://www.reuters.com/business/healthcare-pharmaceuticals/takeda-iambic-use-ai-speed-up-drug-development-2026-02-09/ (accessed February 11, 2026).
17. Vinciane Beauchene, Sylvain Duranton, Nipun Kalra, and David Martin, “AI at Work: Momentum Builds, but Gaps Remain,” Boston Consulting Group, June 26, 2025, https://www.bcg.com/publications/2025/ai-at-work-momentum-builds-but-gaps-remain (accessed February 11, 2026).
18. Sheryl Estrada, “MIT report: 95% of generative AI pilots at companies are failing,” Fortune, August 18, 2025, https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/ (accessed February 11, 2026).
19. Axios, “Walmart is using AI to help it spot and respond to trends faster,” April 9, 2025, https://www.axios.com/2025/04/09/walmart-ai-trends-fashion (accessed February 11, 2026).