Post-Event Recap

Building the Enterprise AI Factory: From Experimentation to Execution

An executive conversation on what has to change inside the organization — operating models, governance, data context, and team design — to make enterprise AI repeatable, scalable, and operationally useful.

Executive summary

What leaders are working through now

The discussion moved past the familiar conversation about AI pilots and focused on what must change inside the organization for AI to become a repeatable, operational capability.

Enterprise leaders are moving beyond the familiar conversation of AI pilots and focusing on what must change inside the organization to make AI repeatable, scalable, and operationally useful. The discussion emphasized that AI does not scale through technology alone. It requires new operating models, stronger governance, better data context, cross-functional teams, and clear decision-making rhythms that connect AI initiatives to measurable business outcomes.

A recurring theme was that organizations are beginning to treat AI less as an innovation project and more as a core business capability. Successful companies are building reusable systems instead of one-off solutions, forming teams around outcomes rather than isolated experimentation, and embedding AI into existing software development, customer support, privacy, and operational workflows. However, many enterprises still struggle to measure productivity gains, prioritize use cases, and determine which initiatives deserve production-level investment.

The conversation also reinforced the continuing importance of human expertise. AI can accelerate coding, testing, service workflows, and knowledge work, but humans remain essential for judgment, context, architecture, governance, and accountability. As AI becomes more capable, organizations will need to rethink how they train junior talent, preserve institutional knowledge, and prepare employees to supervise increasingly autonomous systems.

Moderator and panel

Perspectives from the room

The panel brought together AI solutions, machine learning, engineering, and privacy leaders from Andela, RingCentral, Kinship, and TikTok.

Headshot of Kennith Jackson Moderator

Kennith Jackson

SVP AI Solutions & Operations

Andela

Headshot of Sushant Hiray Panelist

Sushant Hiray

Senior Director of Machine Learning

RingCentral

Headshot of Sridevi Gouni Panelist

Sridevi Gouni

Head of Engineering

Kinship

Headshot of Jianpeng Mo Panelist

Jianpeng Mo

Director of Engineering, Privacy

TikTok

Key themes

What turns AI into an operating capability

Five ideas consistently shaped the discussion: operating capacity, governance as an enabler, human readiness, value-based prioritization, and the future of human-in-the-loop work.

01

From AI pilots to operating capacity

Organizations making progress are moving away from isolated experiments and toward repeatable AI systems that support multiple use cases across the business.

02

AI governance as an enabler of scale

Governance should not function as a blocker. When designed well, it creates clarity, confidence, and repeatable guardrails that help teams deploy AI faster and more safely.

03

Human readiness and organizational adoption

The limiting factor is often not the model or tooling, but whether teams understand how to use AI, trust its outputs, and adapt their workflows around it.

04

Prioritization based on business value

Enterprises face pressure to pursue many AI use cases at once, but production investment should be reserved for initiatives with clear value, operational feasibility, and repeatable impact.

05

The future of human-in-the-loop work

As AI takes on more execution tasks, organizations must still develop people who understand systems deeply enough to supervise, question, and guide AI outputs.

Actionable takeaways

What enterprise leaders can do next

These actions translate the discussion into practical decisions for business, product, engineering, data, governance, and talent teams.

Build AI platforms, not one-off solutions

Design systems that can support multiple use cases, business units, and workflows rather than building isolated pilots that are difficult to scale.

Organize teams around outcomes

Bring together engineering, business, governance, and domain experts around measurable business problems instead of separating AI into standalone innovation groups.

Define governance before production deployment

Establish clear guardrails, evaluation criteria, risk thresholds, and approval paths so teams know what is allowed and how to move forward.

Measure value beyond tool usage

Track business outcomes, cycle-time improvements, quality gains, deployment velocity, customer impact, and risk reduction rather than relying only on usage metrics.

Avoid misleading productivity metrics

Lines of code, pull requests, or tool logins can create the wrong incentives. Measure end-to-end delivery impact instead.

Prioritize use cases with repeatable value

Invest in AI initiatives that solve meaningful business problems, can be operationalized in real workflows, and can be extended across teams or functions.

Strengthen the organizational context layer

Improve knowledge management, documentation, data quality, and internal context so AI systems can produce more reliable and relevant outputs.

Embed AI into development and delivery workflows

Integrate AI into CI/CD, testing, monitoring, feedback loops, and production processes rather than treating it as a separate layer.

Use AI for governance where appropriate

Apply AI to improve compliance monitoring, privacy protection, anomaly detection, and risk evaluation, while maintaining human accountability.

Train employees to supervise AI, not just use it

Build skills in system design, critical thinking, architecture, evaluation, and domain judgment so teams can manage AI outputs effectively.

Protect junior talent development

Create training paths that allow early-career employees to build foundational expertise, even as AI automates more entry-level tasks.

Maintain human accountability for high-impact decisions

Use AI to accelerate execution, but keep humans responsible for judgment, prioritization, escalation, and final accountability in sensitive workflows.

Event moments

Inside the conversation

The evening paired a moderated panel with a wine tasting and peer discussion, giving attendees room to compare operating models, governance approaches, and implementation realities.