4 Questions To Help You Translate Ai Theory Into Practice

“AI for nan Enterprise: The Playbook for Developing and Scaling Your AI Strategy” is simply a caller ebook I wrote for The New Stack pinch sponsorship from Red Hat and Intel. It walks organizational leaders done really to found and pass a cross-organizational AI strategy, and past really to enact and measurement it all.

The book balances experimenting pinch AI for nan biggest wins and getting nan proven — and yes, often boring — AI usage cases successful place. It helps engineering teams group up nan guardrails that tin yet break down information silos and spot up information holes earlier you see your agentic AI future. And it teaches you really to measurement it all.

This book is intended to thief you commencement activity conversations. And, astatine only astir 50 pages — and free, truthful on-budget! — It makes a cleanable agency book nine subject, because these are topics that everyone successful your statement should beryllium talking about.

Get nan Conversation Started

Since AI will soon touch each roles successful an organization, you should purpose to make your book nine arsenic cross-functional and cross-departmental arsenic possible.

I had a speech pinch Gemini astir nan book, and together we developed a bid of questions that tin thief guideline you done nan 4 chapters. These questions are designed to thief anyone construe nan book’s concepts into reflections, discussions and past actual soul plans.

The book offers a model for building an intentional, top-down (read it earlier you conflict maine connected it) program. But if you don’t deliberation your C-suite is considering these things, these questions tin thief empower you and your squad to not only do AI successful an intentional measurement that enhances your work, but to thief you commencement nan correct conversations pinch leadership.

Thank you for reference and sharing my book! Don’t hide to fto maine cognize what you deliberation and what questions it sparks for you and your team.

Chapter 1: Set Your Organization Up for Success

This section is each astir establishing clear direction, cross-organizational activity buy-in — not conscionable tech! — and a coherent firm imagination that each unit and stakeholders understand.

Have you publication your organization’s AI strategy? Do you moreover cognize if your executive activity has one?

Activity: Write a one-paragraph soul connection that intelligibly outlines your company’s AI priorities successful a measurement each labor tin understand. If you’re not successful nan C-suite, draught a paragraph you’d for illustration your company’s activity to constitute — and stock it pinch them, if you dare!

Discuss your paragraph pinch different members of your book club. What are nan differences? Can you harvester your ideas into 1 moreover much powerful paragraph? (Spoiler alert: Corporate messaging must position AI arsenic an augmenter of work, an automator of nan tedious — not an excuse for layoffs.)

Chapter 1 besides discusses nan disconnect betwixt engineering leaders and developers regarding AI’s effect connected productivity. What concrete, cross-functional mechanisms — for illustration a dedicated experimentation week, a formalized feedback loop and breaking down information silos — could your statement create to adjacent this spread and promote safe, bottom-up experimentation?

A top-down AI strategy to guarantee reliability and information is recommended. Beyond appointing a chief AI serviceman (CAIO), what are nan 3 circumstantial action items your existing executive activity squad will/should perpetrate to and instrumentality successful nan adjacent 4th to visibly thrust nan AI strategy?

Be judge everyone tin reply your company’s charismatic stance connected utilizing customer information to train AI models and really you will guarantee afloat customer power and transparency successful each AI-enhanced products.

Chapter 2: Take nan Incremental Approach

The 2nd section of nan book focuses connected placing a large stake connected AI solving pervasive, long-standing problems while (most of nan time) enacting nan proven AI wins.

In this chapter, Red Hat’s Marty Wesley suggests a two-speed AI finance by pursuing nan 80/20 rule.

Activity: Each personnel of your book nine identifies and reflects connected 1 proven usage lawsuit — for illustration soul hunt documentation, paying down method indebtedness aliases customer support consequence — that you tin use to your ain occupation next week to make it much businesslike and little stressful. About 80% of your AI fund should provender specified proven usage cases.

Then, each book nine personnel should place a “20% large bet” — a hard, transformative problem, for illustration a large-scale migration — that, if AI could solve, would make a immense quality to your organization. Did you person nan aforesaid thought aliases different ideas based connected your department? If you are successful executive leadership, really would you prioritize nan ideas?

For each, talk nan results and talk together astir really you will specify nan measurable occurrence of each of these projects.

Chapter 3: Data Security Backed by Engineering Best Practices

This section brings together everything we’ve been recommending for complete a decade. It’s conscionable that nan urgency changes pinch AI. As Harness’s Dewan Ahmed precocious told me, for illustration a slug train going past nan rail, a batch of AI is automating insecurity astatine scale.

Activity: Based connected nan risks outlined successful this section — information leakage, high-risk systems, information compliance, possibly others that are important to your business aliases assemblage — what are nan apical 3 high-risk areas successful your business that your first AI argumentation must instantly consider, vet, specify and put guardrails around?

Don’t hide to induce taxable matter experts to measurement successful connected domain- aliases department-specific AI policies alternatively than relying connected one-size-fits-all rules. For your squad aliases department, sanction 1 circumstantial AI usage lawsuit and nan corresponding human-in-the-loop (HITL) process an master should specify to guarantee value and compliance. If you are an executive leader, brace programme aliases embed connected these teams and inquire them what would thief them do their activity much efficiently. Experiment alongside them — conscionable make judge they cognize it’s not a test.

As codification rolls retired faster than ever, progressively without a quality developer moreover looking astatine it, what worries you nan most? What are nan top risks to your reliability, information and stability? When things spell wrong, really quickly tin you rotation things back?

Chapter 4: Future-Proof Your AI Strategy

Seems for illustration each week we person a caller victor of nan AI GOAT (greatest of each time) award. The reality is that things are changing truthful accelerated that nary statement tin make semipermanent bets connected immoderate AI tooling. Flexibility and measuring everything are nan only ways to future-proof successful nan look of uncertainty. That’s what nan last section of this book is each about.

The closing section introduces nan displacement to an “agentic future” wherever AI agents automate workflows.

Activity: Each book nine subordinate should place 1 simple, repetitive, pattern-following task successful their activity that would beryllium a bully campaigner for an first AI supplier pilot. Once you’ve identified that predictable and boring activity, ask: What observability and context engineering mechanisms must beryllium successful spot to guarantee you tin debug nan agent’s behavior? Brainstorm together astir not only problems but imaginable agentic solutions.

Then, arsenic you look crossed your AI strategy and since you can’t amended what you don’t measure, place 2 metrics — specified arsenic from DORA metrics, nan DX AI Measurement Framework aliases a circumstantial soul costs simplification (please, not lines of code!) — your statement will usage to measurement nan occurrence of its first AI project. Who is nan determination proprietor responsible for determining if nan inaugural is simply a occurrence aliases failure?

This section besides promotes nan worth of unfastened root platforms and nan hybrid unreality for avoiding vendor lock-in and maximizing agility. How does your existent information infrastructure align pinch this vision, and what is nan scheme to guarantee your AI level tin support nan exemplary retraining and redeployment basal erstwhile models inevitably drift?

The last section besides asks which skills are astir successful request successful nan look of AI. Each squad should see what skills aliases aptitudes would thief it respond to this accelerated change. Can that beryllium trained in-house aliases do you person an opening coming up connected your team?

Finally, nan book ends by asking you to clasp nan opportunity of AI, pinch caution. Here, I inquire you to bespeak connected its last words:

“AI is an opportunity to return nan boring of our time to time and to thief everyone do what they do best. AI tin beryllium nan point that emphasizes each organization’s differentiation — but don’t judge that AI is that differentiator. Ground your AI endeavor strategy successful your people, and past your processes and technology, to get nan champion retired of an AI-enriched future. Good luck!”

Get your free transcript of AI for nan Enterprise: The Playbook for Developing and Scaling Your AI Strategy today!

Group Created pinch Sketch.