AI: what you need to know before setting up a PoC

For our customers' marketing teams, understanding how to seize emerging technologies is a must— in particular, to ensure that their budgets and skills shift toward what will bring the most value to both employees and the company.

So we went to the Gartner Symposium, to get the latest technology recommendations from the analysts. The topic at the heart of the debate? AI.

Here are the 3 key points we picked out for you.

AI integration: a mixed picture for companies

By 2024, almost three-quarters of business leaders (74%) believe AI will disrupt their industry. Chief Information Officers (CIOs) (57%) are therefore expected to develop a strategy and vision in line with AI and generative AI. And large sums of money are being devoted to this objective: in 2023, large companies are expected to spend between $300,000 and $2.9 million to develop PoCs(Proof of Concept).

What's more, according to Gartner, software publishers are riding this wave, and can charge up to 50% more once they have integrated an AI-related feature.

What the analyst tells us, however, is that almost half of the CIOs surveyed (47%) believe that the ROI of their AI initiatives has been disappointing. On the other hand, AI-enabled tools are not yet succeeding in creating real productivity gains. in creating real productivity gains. Or at least not across the board: 72% of IT managers claim that employees are finding it difficult to integrate artificial intelligence into their day-to-day tasks.

What impact will AI have on your market and your business?

Will AI have an impact on your business model? According to Gartner, only marginally (20% of respondents).
Will AI improve your processes? More likely (30%).
Will it increase your employees' productivity? There's a good chance (50%).
In fact, if well thought-out and used, AI-enabled tools could save them an average of 40 minutes a day.

What Gartner also tells us is that, in complex businesses, AI productivity gains are proportional to employee experience. Put another way: AI tools for lower value-added jobs are easier to learn, but cover a narrower range of actions, and are therefore limited in their potential benefits.

In this pivotal period, it is therefore vital for companies to identify the best sources of value from AI, in order to focus their transformation efforts on the roles and people where these tools will deliver the best ROI.

Precautions to be taken to successfully integrate AI in 2025

With AI also come many illusions. While AI is seductive, organizations must not lose sight of the limits still encountered by this technology: high costs of scaling up, security flaws, hallucinations...

For a successful AI implementation, Gartner recommends 4 preliminary steps:

• Framing the idea: what needs will AI meet? Who will use the tools? In what proportion? What are the benefits?
• Choice of technology: identify the tool that will meet the objectives, based on clearly defined specifications;
• Validation by employees: present the tool to — or even pilot it with — the employees who will be using it most
• Scale-up costing: would the tool be viable on a company-wide scale?