RPA stole the spotlight at Accelerate Vienna, which was themed \u201cAutomation @ The Speed of Change.\u201d From Anil Cheriyan\u2019s keynote on \u201cTackling Digital Transformation Today,\u201d to 25+ RPA sessions highlighting Tricentis RPA customer success stories and best practices, to the \u201cpaparazzi moment\u201d that occurred when we unveiled Tricentis RPA\u2019s disruptive pricing model, RPA was hot from start to finish. This was hardly surprising, given the buzz across the RPA market, paired with Tricentis\u2019 explosive growth and powerful extension from test automation into RPA.\nWhat was surprising, though, were some of the RPA research stats that Forrester\u2019s RPA expert Craig Le Clair shared during his keynote. Here\u2019s a recap of some of his most attention-getting points (for additional context and findings, you can watch Craig\u2019s complete keynote on-demand).\nThe service to license ratio is staggering\nThe service to license ratio is the amount of outside services that you need to make a dollar\u2019s worth of software productive. For RPA today, the service to license ratio is quite high. In 2019, for every dollar spent on RPA software, $3.41 dollars were spent on services to make it work.\nWhy does a fully \u201cnon-invasive\u201d technology\u2014technology that does not alter existing systems in any way\u2014require such a staggering degree of services? Because service providers that \u201cparachute in and never leave\u201d are busy maintaining bots, setting up centers of excellence, attempting to chart a course toward scalability, and addressing the various security aspects that arise when RPA bots gain more intelligence. Pure bot creation accounts for just a fraction of that services spend.\nWorking bots pay off nicely\u2014until you face downtime\nAssume it costs you 25,000 Euros to build a bot that ultimately replaces the equivalent of 2-3.5 FTEs. Even if you have a relatively high services cost associated with maintaining that bot, you will come out far ahead\u2026as long as the bot continues working. If a bot breaks and you can\u2019t revive it immediately, however, your process downtime eats into your ROI very quickly. This is especially problematic if you no longer have ready access to the FTEs that previously completed the task manually.\nIslands of automation are emerging\u2014but it\u2019s rarely paradise\nMost organizations now have licenses for multiple RPA tools, and many are looking to bring in even more RPA tools as their existing solutions fail to scale and meet expectations. Add to that different chatbot stacks (both general and \u201cdomain-ified\u201d), digital process automation, and so forth, and you end up with scattered islands of automation. Unfortunately, there\u2019s usually no coordinating element.\nIdeally, we need a more orchestrated effort driven by strike teams (similar to centers of excellence, but less bureaucratic). The strike team should evangelize automation, manage the guardrails, provide design authority, advise on the technology that\u2019s truly the best for each use case, and\u2014above all\u2014coordinate across the archipelago.\nScaling RPA is a major hurdle for the majority of RPA implementations\nThe majority of enterprise organizations have fewer than 10 bots in production, and only 13% have 100+ bots. That\u2019s dismal. Moreover, production bots are utilized only about 15% of the time. What\u2019s needed to scale? That\u2019s the million-dollar question in this explosive industry which appears to be headed toward a tipping point. See the next section for some ideas.\nBonus: additional RPA insights coming soon\nCraig, Forrester, and Tricentis are in the final stages of some exciting new research on what\u2019s really required to adopt and scale enterprise automation. Together, we\u2019re exploring:\n\nThe industry\u2019s current state of automation maturity\nWhat the top automation leaders do differently\nThe greatest barriers to sustainable automation\nThe business impact of brittle automation\n\nVisit the Tricentis RPA Resource Center in January 2020 to access the complete report as soon as it\u2019s published.