V0.3
Andy Weeger / Neu-Ulm University of Applied Sciences Neu-Ulm / Digital Innovation in Industry

IBM as Client Zero

DI in Industry (DIiI)

Andy Weeger

Neu-Ulm University of Applied Sciences

May 18, 2026

Learning objectives

After working through this case, you will be able to:

  1. Distinguish the architectural and behavioural properties of agentic AI from traditional, rule-based automation using a real enterprise case.
  2. Classify an operational AI workflow against the five agentic workflow patterns of Anthrophic (2024).
  3. Analyse containment-rate limits and escalation cases through the lens of complementary strengths (Dellermann et al., 2019; Hemmer et al., 2025).
  4. Interpret the automation-augmentation paradox (Raisch & Krakowski, 2021) for entry-level job design.
  5. Evaluate how an organisation can govern autonomy, learning, and inscrutability in HR-related agentic systems (Berente et al., 2021).
  6. Reflect on the strategic value of the “Client Zero” living lab approach for a B2B technology vendor.

Prologue

Discussion

A CHRO triples entry-level hiring two years after announcing an AI-driven hiring freeze.

What does this tell you about how agentic AI is reshaping work?

05:00

The hook

Before we sell it, we run it. IBM Client Zero programme principle

The case

Client Zero in one paragraph

IBM applies its own watsonx platform to its own back office before commercialisation. The internal deployment serves three purposes:

  • Operational: unlock productivity in HR, IT, finance, and supply chain
  • Empirical: generate validated metrics that condition B2B sales claims
  • Strategic: train an “AI-first” operating model that becomes a reference architecture

Inside the deployment

  • AskHR orchestrates Workday, SAP SuccessFactors, and Concur for HR workflows
  • IT support uses autonomous agent architectures for standard ticket resolution
  • May 2023: hiring freeze announcement for ~7,800 administrative roles
  • Early 2026: tripling of entry-level hiring, redefined as AI-orchestration roles

Same company.
Same technology trajectory.
Opposite hiring signal.

What changed?

Headline KPIs

Business unit Deployed system Reported effect
Human resources AskHR via watsonx Orchestrate (Workday, SuccessFactors, Concur) 11.5M employee interactions p.a.; 94% containment rate; 40% reduction in HR operating costs
IT support watsonx agent architectures with automated ticket routing 86% of standard IT tickets resolved autonomously; substantial MTTR reduction
Enterprise-wide watsonx Governance and Living Lab architecture $4.5B productivity gains over 3 years (finance, supply chain)
Table 1: Reported outcomes of IBM’s internal AI transformation (synthesised from IBM Client Zero disclosures)

A historical detail to keep in mind

At the onset of the AskHR automation drive, IBM’s internal HR Net Promoter Score reportedly fell from +19 to -35 before recovering.

Evolution

From traditional to agentic AI

Compare the characteristics of traditional, rule-based chatbots with the autonomous architecture of IBM’s AskHR.

Which of Acharya et al. (2025)’s core characteristics cleanly distinguish AskHR from traditional process automation, and which are blurred?

Frameworks to apply

  • Acharya et al. (2025) on agentic core characteristics
  • Russel & Norvig (2022) on the agent taxonomy (simple-reflex, model-based, goal-based, learning)
  • Berente et al. (2021) on autonomy, learning, and inscrutability

Agentic workflow patterns

When watsonx Orchestrate invokes specialised sub-agents to execute a cross-system workflow, which of the five agentic workflow patterns of Anthrophic (2024) is primarily utilised?

The five candidates:

  1. Prompt chaining
  2. Routing
  3. Parallelisation
  4. Orchestrator-workers
  5. Evaluator-optimizer

Justify your classification using case evidence, and identify the strongest opposing classification.

Human-machine interaction

The 94% and the 6%

AskHR records a 94% containment rate. What does the remaining 6% look like?

Using complementary-strengths frameworks (Dellermann et al., 2019; Hemmer et al., 2025), construct a typology of the escalation cases.

Incorporate the historical NPS plunge (+19 to -35) into your analysis: what does it tell you about the relationship between containment and complementarity?

Frameworks to apply

  • Hemmer et al. (2025) on information and capability asymmetry
  • Dellermann et al. (2019) on hybrid intelligence and role distribution
  • Raisch & Krakowski (2021) on the automation-augmentation logic

The entry-level employment paradox

IBM moved from a planned hiring freeze in 2023 to a tripling of entry-level hiring in 2026.

How does agentic AI reconstruct the classic automation-augmentation paradox (Raisch & Krakowski, 2021) specifically for junior professionals and entry-level career tracks?

Is the 2026 pivot a reversal of strategy, or a confirmation of the paradox? Defend your answer with reference to task content versus role content.

Governance

Inscrutability and algorithmic bias

HR decisions are governed by tight legal, ethical, and organisational constraints.

Using Berente et al. (2021)’s three dimensions (autonomy, learning, inscrutability), evaluate how IBM can ensure that autonomous agents do not replicate or systematise historical biases during promotional screenings.

Which organisational and technical guardrails are indispensable?

Frameworks to apply

  • Berente et al. (2021) on the three management dimensions
  • Shavit et al. (2023) on safe-operation practices
  • Herath et al. (2024) on accountability anchoring
  • Jarrahi & Ritala (2025) on principal-agent framing
  • Papagiannidis et al. (2025) on operational governance phases
  • EU AI Act risk classification

The strategic value of Client Zero

Elaborate on the economic, strategic, and trust-building value of the “Client Zero” model for a global B2B technology vendor.

In what ways does using one’s own organisation as a living lab validate and condition market-facing B2B value-creation claims?

Name two failure modes of the Client Zero approach that would invalidate the strategic argument.

Frameworks to apply

  • Soh & Markus (1995) on the IT value-creation process and the missing link
  • Schryen (2013) on internal/external × tangible/intangible IS business value

Sources

Mandatory primary sources

  • IBM Client Zero Hub (public case page)
  • Wharton interview with Radha Plumb (VP of AI Transformation)
  • IMD Business School analysis of AskHR (incl. the NPS history)
  • 2023 hiring-freeze coverage (Al Jazeera, May 2023)
  • 2026 hiring-pivot coverage (Times of India, 2026)
  • IBM Institute for Business Value (IBV) thought leadership on Agentic Operations

Required theory

  • Acharya et al. (2025) (core characteristics and architectures)
  • Hemmer et al. (2025) (sources of complementarity and design implications)
  • Berente et al. (2021) (three management dimensions)
  • Raisch & Krakowski (2021) (the automation-augmentation paradox)

Guiding questions for your reading

  1. Which of IBM’s published metrics are directly verifiable, and which depend on definitions IBM controls?
  2. Where in your own project do you currently expect autonomous agent action, and what is its consequence class?
  3. What is the equivalent of IBM’s NPS for your project: the complementarity metric that should be tracked alongside the productivity metric?
  4. Sketch the minimum governance design that would let your project deploy at one-tenth of IBM’s internal scale.
  5. Could your organisation plausibly become a “Client Zero” for the solution you are designing? Why or why not?

Q&A

Literature

Acharya, D. B., Kuppan, K., & Divya, B. (2025). Agentic AI: Autonomous intelligence for complex goals–a comprehensive survey. IEEE Access.
Anthrophic. (2024). Building effective agents. Anthropic Research Team; https://www.anthropic.com/engineering/building-effective-agents.
Berente, N., Gu, B., Recker, J., & Santhanam, R. (2021). Managing artificial intelligence. MIS Quarterly, 45(3), 1433–1450. https://doi.org/10.25300/MISQ/2021/16274
Dellermann, D., Ebel, P., Söllner, M., & Leimeister, J. M. (2019). Hybrid intelligence. Business & Information Systems Engineering, 61, 637–643.
Hemmer, P., Schemmer, M., Kühl, N., Vössing, M., & Satzger, G. (2025). Complementarity in Human–AI collaboration: Concept, sources, and evidence. European Journal of Information Systems, 34(6), 979–1002. https://doi.org/10.1080/0960085X.2025.2475962
Herath, S., Shrestha, Y. R., & Krogh, G. von. (2024). Design principles for artificial intelligence-augmented decision making: An action design research study. European Journal of Information Systems, 34(2), 207–229. https://doi.org/10.1080/0960085X.2024.2330402
Jarrahi, M. H., & Ritala, P. (2025). Rethinking AI agents: A principal–agent perspective. California Management Review. https://doi.org/10.1177/00081256251320040
Papagiannidis, E., Mikalef, P., & Conboy, K. (2025). Responsible artificial intelligence governance: A review and research framework. The Journal of Strategic Information Systems, 34(2), 101885. https://doi.org/10.1016/j.jsis.2024.101885
Raisch, S., & Krakowski, S. (2021). Artificial intelligence and management: The automation–augmentation paradox. Academy of Management Review, 46(1), 192–210. https://doi.org/10.5465/amr.2018.0072
Russel, S., & Norvig, P. (2022). Artificial intelligence: A modern approach. Pearson Education.
Schryen, G. (2013). Revisiting IS business value research: What we already know, what we still need to know, and how we can get there. European Journal of Information Systems, 22(2), 139–169.
Shavit, Y., Agarwal, S., Brundage, M., Adler, S., O’Keefe, C., Campbell, R., Lee, T., Mishkin, P., Eloundou, T., Hickey, A., et al. (2023). Practices for governing agentic AI systems. Research Paper, OpenAI.
Soh, C., & Markus, M. L. (1995). How IT creates business value: A process theory synthesis. ICIS 1995 Proceedings, 4.