{"id":3901,"date":"2026-09-10T12:02:17","date_gmt":"2026-09-10T12:02:17","guid":{"rendered":"https:\/\/www.infobip.com\/developers\/?p=3901"},"modified":"2026-09-10T12:35:49","modified_gmt":"2026-09-10T12:35:49","slug":"the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip","status":"publish","type":"post","link":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip","title":{"rendered":"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\"><strong>At Infobip, more than 90% of employees are monthly active AI users. Adoption turned out to be the easy part.&nbsp;The much harder question is what all that usage actually delivers.&nbsp;AI spend is centrally visible; the value it creates is not.<\/strong>&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Value can show up in very different ways: a cost might disappear from the budget, a team might free up capacity and use it to take on more work, or AI might surface a commercial opportunity that eventually turns into revenue. These are all meaningful outcomes, but each one requires different evidence.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>We started grouping AI impact into three categories: savings,&nbsp;productivity&nbsp;and revenue.<\/strong>&nbsp;The categories gave us a practical way to ask what changed, what evidence supports the change, and how confidently we can attribute it to AI.&nbsp;<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">The distinction matters because the same AI use can lead to very different outcomes. A faster process might reduce external spending, free up capacity, or simply shift work elsewhere. An AI-supported commercial lead might turn into revenue &#8211; or it might lead nowhere. And a manager may reasonably feel that a team is more productive, even when the delivery data tells a more complicated story.<\/p>\n<\/blockquote>\n\n\n\n<h2 class=\"wp-block-heading\">AI changes where the work happens\u00a0<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Our first real lesson came from engineering:\u00a0<strong>getting faster at one part of the job\u00a0doesn&#8217;t\u00a0necessarily mean getting more done overall<\/strong>.\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In the teams and&nbsp;period&nbsp;we looked at, code was coming together faster in&nbsp;a number of&nbsp;cases. But that was only part of the story. More effort was also going into reviewing AI-generated code, understanding it, testing&nbsp;it&nbsp;and fixing it. The quality measures we tracked remained stable or improved slightly.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The work had changed shape. Some tasks disappeared, some became faster, and others moved further down the process. Delivery metrics could show us that something had changed, but they&nbsp;couldn&#8217;t&nbsp;always tell us what that change meant for productivity.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That distinction matters\u00a0for\u00a0the framework.\u00a0<strong>If AI saves money, we can usually point\u00a0to\u00a0a financial decision. If it frees up capacity, we need to understand what happened to that capacity.\u00a0If it supports revenue, we need to follow the outcome far enough to establish what actually happened<\/strong>.\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Writing code has never been the only thing between an idea and a product in customers&#8217; hands. Engineers still need to understand the problem, decide what to build, choose the right&nbsp;architecture&nbsp;and make sure the finished product works as intended.&nbsp;<strong>AI can accelerate one of those steps without accelerating all of them<\/strong>. In some cases, it simply moves the bottleneck.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We also saw the role of&nbsp;the developer&nbsp;expanding. As AI took on more of the implementation work, people were able to contribute to problems that previously&nbsp;required&nbsp;more specialized help. Teams could take on more within the same footprint.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Managers across Infobip reported increases in capacity.&nbsp;Almost all&nbsp;said they would feel the absence&nbsp;immediately&nbsp;if their AI tools disappeared tomorrow. But there was another side to that gain.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Some of the remaining work required more context switching and cognitive effort. People were reviewing generated output, moving between a wider range of&nbsp;tasks&nbsp;and keeping more context in mind at once. AI took some work off people&#8217;s plates, but it also changed what they had to pay attention to while doing the rest.&nbsp;<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">When manager estimates and delivery data disagree&nbsp;<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Where we could compare managers&#8217; estimates with delivery telemetry, the picture became more complicated. Some measures supported the confidence behind those&nbsp;estimates,&nbsp;others&nbsp;didn&#8217;t.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That&nbsp;doesn&#8217;t&nbsp;necessarily mean the managers were wrong. They see things that a dashboard cannot: work that quietly stopped happening, capacity redirected to a more valuable task, or problems that were solved before they ever became tickets or metrics.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We found a similar tension in external research. METR&#8217;s 2025 randomized controlled trial found that&nbsp;<strong>experienced open-source developers working in repositories they knew well took about 19% longer to complete tasks with AI&nbsp;assistance<\/strong>, even though they estimated afterward that they had been&nbsp;roughly 20%&nbsp;faster.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A February 2026 follow-up complicated that\u00a0result\u00a0further as more developers began opting out of working without AI,\u00a0introducing a self-selection effect into the measurement.\u00a0And as AI moves from a\u00a0tool\u00a0we prompt to agents that run on their own, that gap only widens &#8211; value gets harder to attribute, not easier.\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Activity, perceived\u00a0productivity\u00a0and business impact are different things<\/strong>.\u00a0The framework needed to keep those distinctions visible rather than compressing them into one AI productivity number.\u00a0<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What counts as AI impact?&nbsp;<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">AI use at Infobip goes well beyond engineering. In go-to-market, for example, AI supports account research, proposal generation, and the identification and prioritization of commercial opportunities across the portfolio.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That means the same three categories&nbsp;have to&nbsp;work across&nbsp;very different&nbsp;kinds of work. Engineering and product teams give us delivery telemetry, while business functions may need evidence from process times, customer response, risk, pipeline,&nbsp;conversion&nbsp;or revenue.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Across Infobip, we ask the same questions of every AI use case. What changed? What did it cost? Why do we believe AI contributed? Who owns the outcome? And what evidence would convince someone outside the team?\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The answers help us test each claim against the three categories: savings,&nbsp;productivity&nbsp;and revenue. Each&nbsp;requires&nbsp;different evidence and a different standard for what counts as a credible result.&nbsp;<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Savings:&nbsp;did&nbsp;a spending decision actually change?&nbsp;<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Savings are the most straightforward category, because they tie back to a real financial decision:&nbsp;a license cancelled, a vendor replaced, external spending reduced, a planned cost avoided in a way you could point to.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">One example held up well: a team built a tool during an internal hackathon that replaced an external regulatory-monitoring subscription the company had been paying for.&nbsp;When the tool went live, the subscription was cancelled and the cost left the budget.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Not every freed-up hour belongs here, even when&nbsp;it&#8217;s&nbsp;genuinely valuable. Time saved that never actually leaves a budget line belongs&nbsp;in&nbsp;productivity.&nbsp;<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Productivity:&nbsp;what&nbsp;happened to the capacity that got freed up?&nbsp;<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Productivity claims&nbsp;almost always&nbsp;start with &#8220;we saved time.&#8221; The harder question is what happened to that time:&nbsp;did the team get more done, deliver faster, improve quality, reduce risk, respond to customers sooner, cut errors, or contribute to revenue downstream?&nbsp;<strong>A team can create real value without its budget&nbsp;shrinking by&nbsp;a dollar<\/strong>, but that value&nbsp;has to&nbsp;show up somewhere observable; it&nbsp;can&#8217;t&nbsp;be asserted from the time-saved estimate alone.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">One pattern\u00a0we&#8217;ve\u00a0seen across our internal hackathons: a team builds a first working tool, puts it into production, then builds a second and third for adjacent problems &#8211; one prototype becoming a small suite addressing\u00a0different parts\u00a0of the same workflow. A team automates a recurring manual process and moves from prototype to recurring operational use within weeks.\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Being in production and being used regularly still\u00a0isn&#8217;t\u00a0a verified productivity return. For the cases here, recurring use is clear; the next measurement step is connecting that use to something measurable &#8211; faster resolution, more completed work, better quality, fewer errors. That step is harder than building the tool, and\u00a0it&#8217;s\u00a0where most of the real measurement work happens.\u00a0<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Revenue:&nbsp;can&nbsp;we complete the chain?&nbsp;<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Revenue is the hardest category to attribute cleanly. A credible claim needs a defensible line from AI-supported activity to a commercial\u00a0outcome\u00a0\u00a0&#8211;\u00a0\u00a0qualified pipeline, conversion, retention, expansion, or recognized revenue &#8211; and\u00a0has to\u00a0reckon honestly with what would have happened anyway.\u00a0<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">One team built a tool during a hackathon that surfaced a commercial opportunity nobody had previously had visibility&nbsp;into.&nbsp;The tool gave the team&nbsp;new information, but it&nbsp;didn&#8217;t&nbsp;produce revenue by itself.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We call this an AI-supported opportunity, and&nbsp;we&#8217;ll&nbsp;only call it a realized revenue result once the financial outcome and the&nbsp;tool&#8217;s&nbsp;contribution to it can both be verified. This is where go-to-market matters&nbsp;most,&nbsp;a shorter path to a deal&nbsp;isn&#8217;t&nbsp;the same as a deal AI closed. The cases that&nbsp;haven&#8217;t&nbsp;completed that chain may still deserve continued investment, so we keep them in the pipeline of things worth&nbsp;watching.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Across all three categories, we start from the changed business outcome&nbsp;because&nbsp;usage tells us&nbsp;only that&nbsp;people used the tool.&nbsp;<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">From success stories to defensible claims&nbsp;<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Once we had a way to categorize AI impact, the next question was harder:&nbsp;<strong>how do we know when a claim is actually backed by evidence?<\/strong>&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We started by looking across the organization, more as&nbsp;a proof&nbsp;of concept than a formal audit. The signals were already there. Managers and teams could point to AI-related outcomes, but most of that information lived in conversations, individual&nbsp;teams&nbsp;and scattered examples.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">What was missing was a consistent way to follow those signals and test them.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>We built a validation loop around that process<\/strong>.&nbsp;A regular survey of managers, which scales better than the ad hoc probing conversations we ran as&nbsp;a pilot.&nbsp;Managers flag potential impact:&nbsp;a metric moved, a tool got used, something changed on the team. An owner is attached to that&nbsp;signal,&nbsp;the outcome&nbsp;we&#8217;d&nbsp;expect to see if the claim holds&nbsp;is&nbsp;defined, and we look for evidence beyond the manager&#8217;s own account. The claim is then confirmed, revised, or dropped.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Finance can&nbsp;validate&nbsp;the financial side of a claim. It&nbsp;can&#8217;t&nbsp;independently judge every quality, risk, or experience outcome.&nbsp;Those need evidence and accountable judgment from the people who understand the work.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Finance and transformation teams have long distinguished forecast benefits from realized ones.&nbsp;We&#8217;re&nbsp;adapting that discipline to AI use spread across far more tools, workflows, and departments than a typical transformation program deals with.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">An estimate stays useful as long as&nbsp;it&#8217;s&nbsp;anchored to something&nbsp;observable,&nbsp;its limits are&nbsp;stated&nbsp;plainly, and&nbsp;there&#8217;s&nbsp;a path toward stronger evidence later. We reserve &#8220;verified&#8221; for the smaller number of savings that have completed independent financial review, and &#8220;verifiable&#8221; for estimates anchored to something observable that&nbsp;haven&#8217;t&nbsp;yet cleared that bar.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The point&nbsp;isn&#8217;t&nbsp;only to justify spending after the fact. Better evidence helps us decide what to scale, where we need sharper instrumentation, which teams need support, and where enthusiasm might be running ahead of the results.&nbsp;<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">We now have a practical framework for thinking about AI value across savings,\u00a0productivity\u00a0and revenue, along with a validation process that is still evolving.\u00a0What we still cannot measure consistently is the value of capacity that has been freed up but redirected into work that has no obvious baseline, as well as the contribution AI makes to outcomes shaped by many other factors. That is where attribution gets\u00a0hardest\u00a0and where the next stage of our measurement work begins.\u00a0The strategic case for AI at Infobip was never really in doubt. The startup instinct that got us here &#8211; people building, fast &#8211; is exactly what we want to keep; our job is to make sure our read on the value keeps pace with it. As\u00a0<a href=\"https:\/\/shiftmag.dev\/ai-first-izabel-jelenic-infobip-10156\/\" target=\"_blank\" rel=\"noopener\">our CTO has put it<\/a>, this is about expanding what people can do, not replacing them. The harder task is the one\u00a0we&#8217;re\u00a0still doing: getting the spend and the value back in the same room and letting the evidence guide where our next investment goes.\u00a0<\/p>\n","protected":false},"excerpt":{"rendered":"<p>At Infobip, more than 90% of employees are monthly [&hellip;]<\/p>\n","protected":false},"author":86,"featured_media":3903,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_import_markdown_pro_load_document_selector":0,"_import_markdown_pro_submit_text_textarea":"","footnotes":""},"categories":[50,28,254],"tags":[261],"coauthors":[334],"class_list":["post-3901","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai","category-blog-post","category-engineering-practices","tag-thought-leadership"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v25.6 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0 - Infobip Developers Hub<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0 - Infobip Developers Hub\" \/>\n<meta property=\"og:description\" content=\"At Infobip, more than 90% of employees are monthly [&hellip;]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip\" \/>\n<meta property=\"og:site_name\" content=\"Infobip Developers Hub\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/infobip\/\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-10T12:02:17+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-09-10T12:35:49+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1200\" \/>\n\t<meta property=\"og:image:height\" content=\"768\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"Andrea Popovi\u0107\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@InfobipDev\" \/>\n<meta name=\"twitter:site\" content=\"@InfobipDev\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Andrea Popovi\u0107\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"9 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#article\",\"isPartOf\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip\"},\"author\":{\"name\":\"Andrea Popovi\u0107\",\"@id\":\"https:\/\/www.infobip.com\/developers\/#\/schema\/person\/90102a76ce84010bc17afec2d3d92ea0\"},\"headline\":\"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0\",\"datePublished\":\"2026-09-10T12:02:17+00:00\",\"dateModified\":\"2026-09-10T12:35:49+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip\"},\"wordCount\":2027,\"publisher\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/#organization\"},\"image\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png\",\"keywords\":[\"thought leadership\"],\"articleSection\":[\"AI\",\"Blog Post\",\"Engineering Practices\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip\",\"url\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip\",\"name\":\"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0 - Infobip Developers Hub\",\"isPartOf\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage\"},\"image\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png\",\"datePublished\":\"2026-09-10T12:02:17+00:00\",\"dateModified\":\"2026-09-10T12:35:49+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage\",\"url\":\"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png\",\"contentUrl\":\"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png\",\"width\":1200,\"height\":768},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/www.infobip.com\/developers\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.infobip.com\/developers\/#website\",\"url\":\"https:\/\/www.infobip.com\/developers\/\",\"name\":\"Infobip Developers Hub\",\"description\":\"Build meaningful customer relationships across any channel\",\"publisher\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.infobip.com\/developers\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/www.infobip.com\/developers\/#organization\",\"name\":\"Infobip Developers Hub\",\"url\":\"https:\/\/www.infobip.com\/developers\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.infobip.com\/developers\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2023\/03\/Infobip_logo_favicon.png\",\"contentUrl\":\"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2023\/03\/Infobip_logo_favicon.png\",\"width\":696,\"height\":696,\"caption\":\"Infobip Developers Hub\"},\"image\":{\"@id\":\"https:\/\/www.infobip.com\/developers\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/infobip\/\",\"https:\/\/x.com\/InfobipDev\",\"https:\/\/www.youtube.com\/channel\/UCUPSTy53VecI5GIir3J3ZbQ\",\"https:\/\/github.com\/infobip-community\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.infobip.com\/developers\/#\/schema\/person\/90102a76ce84010bc17afec2d3d92ea0\",\"name\":\"Andrea Popovi\u0107\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.infobip.com\/developers\/#\/schema\/person\/image\/f75f399190a6d105ec9fdbd61d14eeee\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/c1167efc341af5af939858463700a95a87d6a6a122a9765631788659a44de92a?s=96&d=mm&r=g\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/c1167efc341af5af939858463700a95a87d6a6a122a9765631788659a44de92a?s=96&d=mm&r=g\",\"caption\":\"Andrea Popovi\u0107\"},\"description\":\"Andrea is a Staff Engineer in Engineering Operations at Infobip, working where engineering, product, and outcomes meet. She brings structure to ambiguous problems and uses meaningful metrics to ground decisions in evidence.\",\"url\":\"https:\/\/www.infobip.com\/developers\/blog\/author\/andrea-popovic\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0 - Infobip Developers Hub","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip","og_locale":"en_US","og_type":"article","og_title":"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0 - Infobip Developers Hub","og_description":"At Infobip, more than 90% of employees are monthly [&hellip;]","og_url":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip","og_site_name":"Infobip Developers Hub","article_publisher":"https:\/\/www.facebook.com\/infobip\/","article_published_time":"2026-09-10T12:02:17+00:00","article_modified_time":"2026-09-10T12:35:49+00:00","og_image":[{"width":1200,"height":768,"url":"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png","type":"image\/png"}],"author":"Andrea Popovi\u0107","twitter_card":"summary_large_image","twitter_creator":"@InfobipDev","twitter_site":"@InfobipDev","twitter_misc":{"Written by":"Andrea Popovi\u0107","Est. reading time":"9 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#article","isPartOf":{"@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip"},"author":{"name":"Andrea Popovi\u0107","@id":"https:\/\/www.infobip.com\/developers\/#\/schema\/person\/90102a76ce84010bc17afec2d3d92ea0"},"headline":"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0","datePublished":"2026-09-10T12:02:17+00:00","dateModified":"2026-09-10T12:35:49+00:00","mainEntityOfPage":{"@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip"},"wordCount":2027,"publisher":{"@id":"https:\/\/www.infobip.com\/developers\/#organization"},"image":{"@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage"},"thumbnailUrl":"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png","keywords":["thought leadership"],"articleSection":["AI","Blog Post","Engineering Practices"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip","url":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip","name":"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0 - Infobip Developers Hub","isPartOf":{"@id":"https:\/\/www.infobip.com\/developers\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage"},"image":{"@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage"},"thumbnailUrl":"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png","datePublished":"2026-09-10T12:02:17+00:00","dateModified":"2026-09-10T12:35:49+00:00","breadcrumb":{"@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#primaryimage","url":"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png","contentUrl":"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2026\/09\/Milena.png","width":1200,"height":768},{"@type":"BreadcrumbList","@id":"https:\/\/www.infobip.com\/developers\/blog\/the-hard-part-after-ai-adoption-proving-ai-roi-at-infobip#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.infobip.com\/developers\/"},{"@type":"ListItem","position":2,"name":"The Hard Part After AI Adoption: Proving AI ROI at Infobip\u00a0\u00a0"}]},{"@type":"WebSite","@id":"https:\/\/www.infobip.com\/developers\/#website","url":"https:\/\/www.infobip.com\/developers\/","name":"Infobip Developers Hub","description":"Build meaningful customer relationships across any channel","publisher":{"@id":"https:\/\/www.infobip.com\/developers\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.infobip.com\/developers\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/www.infobip.com\/developers\/#organization","name":"Infobip Developers Hub","url":"https:\/\/www.infobip.com\/developers\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.infobip.com\/developers\/#\/schema\/logo\/image\/","url":"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2023\/03\/Infobip_logo_favicon.png","contentUrl":"https:\/\/www.infobip.com\/developers\/wp-content\/uploads\/2023\/03\/Infobip_logo_favicon.png","width":696,"height":696,"caption":"Infobip Developers Hub"},"image":{"@id":"https:\/\/www.infobip.com\/developers\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/infobip\/","https:\/\/x.com\/InfobipDev","https:\/\/www.youtube.com\/channel\/UCUPSTy53VecI5GIir3J3ZbQ","https:\/\/github.com\/infobip-community"]},{"@type":"Person","@id":"https:\/\/www.infobip.com\/developers\/#\/schema\/person\/90102a76ce84010bc17afec2d3d92ea0","name":"Andrea Popovi\u0107","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.infobip.com\/developers\/#\/schema\/person\/image\/f75f399190a6d105ec9fdbd61d14eeee","url":"https:\/\/secure.gravatar.com\/avatar\/c1167efc341af5af939858463700a95a87d6a6a122a9765631788659a44de92a?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/c1167efc341af5af939858463700a95a87d6a6a122a9765631788659a44de92a?s=96&d=mm&r=g","caption":"Andrea Popovi\u0107"},"description":"Andrea is a Staff Engineer in Engineering Operations at Infobip, working where engineering, product, and outcomes meet. She brings structure to ambiguous problems and uses meaningful metrics to ground decisions in evidence.","url":"https:\/\/www.infobip.com\/developers\/blog\/author\/andrea-popovic"}]}},"_links":{"self":[{"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/posts\/3901","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/users\/86"}],"replies":[{"embeddable":true,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/comments?post=3901"}],"version-history":[{"count":7,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/posts\/3901\/revisions"}],"predecessor-version":[{"id":3914,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/posts\/3901\/revisions\/3914"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/media\/3903"}],"wp:attachment":[{"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/media?parent=3901"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/categories?post=3901"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/tags?post=3901"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.infobip.com\/developers\/wp-json\/wp\/v2\/coauthors?post=3901"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}