Enterprise success in AI requires committed leadership

paper-airplane-leadership-metaphor

Artificial intelligence (AI) initiatives ideally should start as pilot projects of limited scope. This allows enterprise IT pros and involved employees to learn quickly through trial and error, thus minimizing the impact of mistakes and enabling the creation of best practices for large-scale AI rollouts.

Alas, problems can persist or arise even after an AI initiative is deployed for live use. Over at CIO, contributing writer Maria Korolov lists a half-dozen reasons why AI projects fail, based on a recent IDC survey of global organizations that already are using AI. Five of the six are data-related — a dearth of data, training data bias, data integration issues, data “drift,” and untouched, unstructured data. (The sixth reason was cultural challenges.)

Yet more than a quarter of IDC survey respondents cite “a lack of staff and unrealistic expectations for the technology” as major challenges, Korolov writes.

AI skills shortages are an ongoing and unfortunate obstacle across multiple sectors, which enterprises can address successfully only through a comprehensive skills acquisition strategy (educate, recruit, and train) or an unlimited budget (good luck with that!). As far as “unrealistic expectations,” I’d call that a failure of leadership. A successful AI initiative must be grounded in realistic and measurable objectives; those start at the top.

If I were to guess, I’d say that some unrealistic expectations around AI projects are rooted in uncertainty over purpose and approach. The more precisely you can define both the strategy and the preferred outcomes of an AI implementation, the more likely it is that you will develop more tangible — and thus realistic — expectations. In the AI-driven digital economy, vagueness is a fatal weakness.

So too is faint-heartedness. AI implementation can be messy, “two steps forward, one step backwards” affairs. Stay the course, learn from your mistakes, and stick it out. Otherwise failure is a self-fulfilling prophecy.

Comments

  1. Anupriya Ramraj says:

    Well said! AI implementations need persistence to see through with investment in both modeling and training data. AIOps for the enterprise is increasingly being specified as a requirement by Managed Service Provider(MSP) audits and DXC continues to invest in this space.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: