1、Optimizing governance to accelerate GenAI deploymentAuthor:Dawn Bushaus,Contributing AnalystEditor:Ian Kemp,Managing EditorSeptember 2025 inform.tmforum.orgContents3 The big picture 6 Key findings7 Section 1:Are CSPs deploying GenAI in production,at scale?11 Section 2:Data and AI model governance ar

2、e inextricably linked17 Section 3:Explainable AI provides critical guardrails21 Section 4:AI centers of excellence are a steppingstone to becoming AI native27 Section 5:How ODA is evolving to optimize AI governance33 Additional sponsor feature:Accelerate L4 autonomy:Scale agentic AI solutions into p

3、roduction 37 Additional resourcesWe hope you enjoy the report and,most importantly,fjnd ways to use the ideas,concepts and recommendations detailed within.You can send your feedback to the editorial team at TM Forum via editortmforum.orgThe big picture3inform.tmforum.org“If you look at projects that

4、 do not have governance,you will see a lot of failures,because they wont get past legal;they wont get past a risk assessment;and they wont go into production,”says Eoin Coughlan,CTO and Industry Lead for Telecommunications,Media and Entertainment at IBM.Indeed,two recent studies illustrate how diffi

5、cult it is to scale GenAI in production.A February 2025 report commissioned by Lenovo and authored by IDC found that 88%of generative AI(GenAI)pilots never reach production,while a more recent study from MIT,published in July,highlighted that a staggering 95%of GenAI tools fail to deliver expected r

6、eturn on investment(ROI).“And here in the telco industry,we know that over half of telco AI initiatives never really get off the starting blocks,”said TM Forum CEO Nik Willetts during his keynote at Innovate Americas earlier this month.“Thats not yet scale that is just frustration.”Governance is imp