It’s unsurprising to see legacy providers shifting their business models from perpetual software licensing to subscription-based pricing. Some do it with a measure of grace. But lately, some licensees of virtual desktops and applications have been confronted with abrupt changes and even forced to accept and pay for unwanted features. However, there are ways for alienated customers to protect their best interests. This past year has been rife with complaints over a substantial licensing change by Citrix after it was acquired by private equity firms and merged with TIBCO Software. Those changes remarkably parallel a playbook that VMware customers experienced in the wake of their vendor being acquired by Broadcom. For those paying close attention, substantive changes were foreshadowed in 2023 when it was quietly noted that perpetual software maintenance licenses would not be renewed upon expiration. Even those paying attention back then have been hit with what they consider even more egregious changes, and some are citing licensing cost increases of 300% or more. Specifically, some organizations have griped that while they asked for renewal terms six months or more before their deadlines, those requests went into a dark hole until as little as 30 days remained on their current agreement. That’s too late for an organization with substantial numbers of users—or even just a few—to evaluate and prepare for a switch if they so desire. When they receive a replacement, er, “renewal” proposal, surprises abound. First, separately licensed products are now “features” within a universal license. For most, that means paying for shelfware they’ll never use while absorbing sometimes astounding price hikes from what they were accustomed to. Some organizations are coming away from licensing discussions convinced they can only obtain 3-year or 5-year agreements! So much for pay-as-you-go subscription models and winning your customer’s trust every day. Moreover, many are being shifted to new support models while being gaslighted the changes are in their best interests. Only select customers—by invitation only—are being offered platform licenses and support. Most are being shifted toward channel partners. Not only that, but channel partners handling 2,000 or fewer licenses have themselves been shifted to a third-party provider. How to protect your interests If you’ve yet to receive a renewal offer from your vendor or channel partner, there’s no time to lose. If you already agreed to a one-year license, prepare for your next renewal date. First, start planning immediately to ease the transition to desktop-as-a-service (DaaS). The writing is on the wall for terminated support of legacy applications, no matter who the vendor is. So, there’s no time to get ahead of the game like the present. Develop a timeline for your next renewal date. Here are some steps to plan for: Demand your sales representative answer your renewal questions and keep demanding until you get an informed response. Take advantage of modeling and cost estimator tools that can help IT model configurations and better understand resource needs and budget impacts for a shift from legacy VDI to DaaS. Determine what virtualized applications to modernize, migrate, or retire–this is something all organizations should be doing no matter their circumstances, as it likely will save money now and ease a transition whenever it comes. Determine the cloud service that can best suit your DaaS needs for the long term, including ease of integration and collaboration with third-party tools and services. Prepare users for upcoming changes to interfaces and processes. Involve them early in the process to both gain feedback and demonstrate the expected benefits. Find the best consolidated management tools that can help enterprises and MSPs manage scaling, security, and updates. Large vendors are continuously looking to cut costs, and that’s not likely to change. Support is costly, whether it’s continuing maintenance of older products or hand-holding customers that vendors no longer view as premium. The more you can control your situation, the better off you’ll be. For more insights into DaaS, read Gartner’s Magic Quadrant report, notably its overview of Microsoft Azure Desktop. source

Security and infrastructure are two of the top concerns for organisations rolling out generative AI, according to a recent report from IT company NTT Data. However, most companies are optimistic about its future potential. The Global GenAI Report, based on responses from 2,307 generative AI decision-makers and influencers, primarily from large organisations globally, found that CISOs are the executives most pessimistic about the technology. Many CISOs (45%) held negative sentiments about generative AI because they were “feeling pressured, threatened and overwhelmed” by it. Only 19% of total respondents from various roles shared the same sentiment as CISOs. “CISOs are uniquely positioned to anticipate risk — and it is clear they see it [in Generative AI],” the report stated. CISOs are more negative about generative AI than other executives including CIOs. Source: NTT Data Executives outline generative AI concerns Among CISOs, one in three said they are uncomfortable with the “black box” nature of some generative AI models and its “unclear decision-making algorithms,” the report found. However, CISOs aren’t the only group harboring concerns about AI: 90% of all executives said that legacy infrastructure greatly affected business agility and their ability to use generative AI. Nearly 8 in 10 respondents still remain unsure of the actual benefits of generative AI to their operations. Three in four respondents said their organisation’s generative AI ambitions conflicted with — or could be negatively affecting — its sustainability goals. 44% of total respondents agreed that their organisation had already established the optimal infrastructure to efficiently and cost-effectively scale generative AI in a cloud environment. Chief data officers are particularly hesitant about GenAI. Nearly half of them viewed a lack of transparency and the difficulty in explaining the reasoning behind complex generative AI models as issues affecting their own adoption. Additionally, 86% of all respondents and 96% of chief data officers agreed that algorithm bias remained pervasive in current models. SEE: Nearly half of security professionals believe AI is risky More must-read AI coverage Most generative AI decision-makers are optimistic about AI Despite some negative attitudes toward AI, the report found that most executives are positive about the technology overall: 60% believe it will be a “game changer” within two years. According to the report: Among COOs, 57% were positive about generative AI. 50% of CIOs and CTOs said they feel “excited” about generative AI, and 21% reported feeling “amazed” by the technology. 44% of all C-suite respondents strongly agree that the promise and potential returns of generative AI outweigh the potential security and legal risks. The C-suite overall is also expecting big impacts from generative AI; 97% of CEOs expect a material impact from the technology, while 99% of organisations are planning to invest more in AI. Only a very small number of organisations are choosing not to invest in AI. Source: NTT Data Organisations looking to 2025, as NTT Data declares “playtime over” NTT Data’s report claims that “playtime is over” for AI, as organisations shift from conducting less experimentation and more identifying tangible successes that can be taken into production. IT is playing the leadership role in generative AI rollouts across organisations. Source: NTT Data Nearly nine in 10 of the respondents surveyed said they were actually experiencing “pilot fatigue” and were shifting their focus to areas where generative AI has had a proven impact on business performance. SEE: 5 generative AI trends in 2025 Looking forward, NTT Data said six in 10 leaders expect a significant transformation from major investment in generative AI in 2025, but 83% accept return on investment will be unclear for the foreseeable future. “Focused spending plans will replace scattered experimentation as organisations look to improve internal operations through more precise experiments. These will aim to transform back-office and middle-office workflows and create new digital products and services with the potential to scale,” the report said. NTT Data predicts that successful experiments will reignite investment in generative AI as CEOs gain clearer evidence of its potential to boost revenue and productivity. source

German tech darling DeepL has (finally) launched a voice-to-text service. It’s called DeepL Voice, and it turns audio from live or video conversations into translated text.  DeepL users can now listen to people speaking a language they don’t understand and automatically translate it to one they do — in real-time. The new feature currently supports English, German, Japanese, Korean, Swedish, Dutch, French, Turkish, Polish, Portuguese, Russian, Spanish, and Italian.  What makes the launch of DeepL Voice exciting is that it runs on the same neural networks as the company’s text-to-text offering, which it claims is the “world’s best” AI translator.   As someone who’s just moved to a foreign country, I’m keen to try a voice-to-text translator that actually might work. All the ones I’ve tried so far aren’t real-time — there’s a lag that renders them pretty useless — and the translation quality is pretty poor.   For face-to-face conversations, you can launch DeepL Voice on your mobile and place it between you and the other speaker. It then displays your conversation so each person can follow translations easily on one device. You can also integrate DeepL Voice into Microsoft Teams and video-conference across language barriers. The translated text appears on a sidebar as captions. It remains to be seen whether DeepL Voice will be available on platforms like Zoom or Google Meet anytime soon.    ‘The next frontier’ While this is DeepL’s first such offering, it’s unlikely to be its last. DeepL’s founder and CEO, Jarek Kutylowski called real-time voice translation the “next frontier” for the business.   “DeepL is already a leader in written translation, but real-time speech translation is an entirely different story,” said DeepL’s founder and CEO, Jarek Kutylowski.  “When translating speech as it happens, you’re dealing with incomplete input, pronunciation issues, latency and more, all of which can lead to inaccurate translations and poor user experience.   “So we built a solution that would take these into account from the offset and enable businesses to break down language barriers by enabling them to communicate in multiple languages as required,” said Kutylowski.  Quality will likely be DeepL Voice’s differentiating factor from the countless other providers of voice-to-text translations. From a technological perspective, DeepL’s success lies in the architecture of its neural networks, the input from human editors, and the training data. But Kutylowski also believes it has a key advantage over its competitors: focus.   “Focus is always an important thing,” Kutylowski previously told TNW. “Translate isn’t the core business of Google — it’s one of the 100 side gigs. The same goes if you consider LLMs and the OpenAIs of this world as our competition; translation is only one thing of what they’re doing and their GPU is doing a tonne of different things. We’re focused on one particular area.”  In May, the DeepL reached a $2bn valuation after securing a new investment of $300mn (€277mn). It covers 32 languages and counts over 100,000 business users.   source

Customer relationship management systems can provide extensive analytics and reporting capabilities for businesses to better understand their sales processes on a more comprehensive level. Secure data mining, data recording, and report generation translate this information into visual representations of KPIs. Popular metrics businesses generate and track include their sales cycle length, marketing ROI, sales rep performance, and deal forecasting. 1 HubSpot CRM Employees per Company Size Micro (0-49), Small (50-249), Medium (250-999), Large (1,000-4,999), Enterprise (5,000+) Micro (0-49 Employees), Medium (250-999 Employees), Large (1,000-4,999 Employees), Small (50-249 Employees) Micro, Medium, Large, Small 2 Zoho CRM Employees per Company Size Micro (0-49), Small (50-249), Medium (250-999), Large (1,000-4,999), Enterprise (5,000+) Any Company Size Any Company Size Features Calendar, Collaboration Tools, Contact Management, and more 3 Pipedrive CRM Employees per Company Size Micro (0-49), Small (50-249), Medium (250-999), Large (1,000-4,999), Enterprise (5,000+) Any Company Size Any Company Size Features Calendar, Collaboration Tools, Contact Management, and more What is CRM analytics? CRM analytics is internal programming that collects, organizes, and analyzes data around customers, sales, and revenue. Generating detailed reports like this is a top CRM feature that businesses of any size and market can benefit from. This data provides actionable insights for leaders, managers, and individual users that assists in making timely and data-based decisions. CRM analytics has also recently been integrated with AI functionality, allowing for more niche reports and suggestions to be generated on command. 5 key CRM metrics Sales cycle length A sales cycle length is the average time it takes a deal to be closed from beginning to end. This starts with a prospect converting to a lead, and that lead becoming a confirmed customer. The goal is to have the sales cycle length be as short as possible, so that leads aren’t lost to competitors, or their need for your solution dissipates. Measuring this metric will help identify opportunities where the sales cycle could be condensed. Customer churn rate A churn rate is a percentage metric that shows how many customers leave your sales pipeline or stop doing business with your company entirely. A high churn rate is something businesses try to avoid, and typically it can also identify where in your sales process the churn is happening. It’s always good to know where you’re succeeding, but equally important to understand where you are losing business in order to make immediate adjustments. Marketing ROI Marketing ROI directly tracks how resources that are allocated to marketing contribute to overall revenue growth. While CRM solutions are sales-oriented, marketing features are included in that. That includes social media communications, email marketing campaigns, and even web and landing page building. This is especially important for small to mid-sized businesses looking to invest more into marketing campaigns, because you need concrete data that supports such an investment. Rep performance Rep performance is tracked by monitoring their overall sales activities and close rates. Common activities that are monitored are sales calls, email performance, referral rates, and lead conversion rates. This tracking can be done on an individual scale, by team, according to location, or for an entire department—and can help with quarterly reviews. This transparency is important especially in a commission-driven environment. Business forecasting The best CRM solutions offer business and revenue forecasting as an advanced CRM metric. These reports take into account customer activity, sales length, churn rate, sales history, and more to come up with predictive analytics around projected revenue wins. While this information is always subject to change depending on any number of factors, having a projected idea of revenue QoQ or YoY helps immensely when planning bigger initiatives for a company. Top 3 benefits of CRM data analysis Increased productivity I recommend using a CRM software as a way to increase productivity around customer interactions, and data analysis helps do exactly that. Since there isn’t time being wasted on manually pulling and sorting data, users can focus on nurturing client relationships and closing deals. Trusted analytics make sure efforts are being put into the right place at the right time and optimize resource allocation. This way, reps, agents, or administrators are spending time doing tasks that are proven to be beneficial to the business. Improved customer satisfaction When you have detailed data analysis built around successful—and less successful—marketing and lead nurturing campaigns, it’s easier to understand what customers consider as positive interaction. With the help of personalized engagement and detailed analysis of metrics like email click rates, conversions, and more, businesses won’t need to guess which strategies actually move customers through their pipeline. This will build customer satisfaction with your business and improve customer retention. Competitive advantage Having strong analytics in a CRM software allows for businesses to keep a constant eye on market changes or trends. This way, they can anticipate those changes and plan accordingly. For example, these trackable trends help make informed decisions about where to invest more or invest less. If a company anticipates a seasonal increase in business, they should invest in marketing campaigns in advance to get ahead of their competitors. It also helps increase overall efficiency and allows for a flexible and agile business plan. CRM analytics use cases When choosing a CRM solution, opting for an analytical CRM tool means there’s an emphasis on data warehousing and mining, plus advanced forecasting. These are great for medium to large teams or businesses that need detailed, secure, and constant data monitoring. Identifying consumer trends In industries that are completely driven by consumer trends, having analytical CRM helps businesses stay ahead and anticipate major changes. Whether it’s financial forecasting, product and service trends, or mapping seasonal spikes, an analytical CRM will take complex data and turn it into understandable insights. For example, a retail business might use purchase history and trends from their consumers to help identify when in the fiscal year their products are in demand. They can then plan to invest in more marketing campaigns leading up to that timeframe in the new year. Scalability When your goal is to grow

In today’s fast-paced digital landscape, AI platforms are playing a pivotal role in reshaping industries and driving business transformation. As businesses across the UAE embark on their digital journeys, AI has emerged as a key enabler, streamlining operations, enhancing decision-making, and fostering innovation. In a recent fireside chat featuring leading industry experts from the tech sector, attendees gained valuable insights into how AI platforms can be leveraged to create new growth opportunities and build a sustainable future in the era of digital transformation. One of the key themes discussed during the session was the growing importance of GenAI as a transformative tool in business operations. Alexander Knigge, Chief Digital & Technology Officer at Modon, a key speaker at the event, emphasized that “Generative AI is our priority,” highlighting the shift towards AI-powered platforms that can enhance everything from customer engagement to operational efficiency. As businesses continue to explore the potential of AI, the integration of generative technologies is seen as essential to remaining competitive in a rapidly evolving market. AI’s ability to generate content, automate complex tasks, and facilitate personalized interactions is reshaping how companies operate and deliver value to their customers. Andrew Murphy, CIO at Abu Dhabi Airports, another expert at the fireside chat, underscored the importance of AI in his company’s long-term strategy. “Generative AI is very important in our journey,” he explained. “In the last few years, we opened more international offices, and as part of our five-year strategy, AI is one of our five key focus areas, with a very strong profile and significant investment.” source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More One-bit large language models (LLMs) have emerged as a promising approach to making generative AI more accessible and affordable. By representing model weights with a very limited number of bits, 1-bit LLMs dramatically reduce the memory and computational resources required to run them. Microsoft Research has been pushing the boundaries of 1-bit LLMs with its BitNet architecture. In a new paper, the researchers introduce BitNet a4.8, a new technique that further improves the efficiency of 1-bit LLMs without sacrificing their performance. The rise of 1-bit LLMs Traditional LLMs use 16-bit floating-point numbers (FP16) to represent their parameters. This requires a lot of memory and compute resources, which limits the accessibility and deployment options for LLMs. One-bit LLMs address this challenge by drastically reducing the precision of model weights while matching the performance of full-precision models. Previous BitNet models used 1.58-bit values (-1, 0, 1) to represent model weights and 8-bit values for activations. This approach significantly reduced memory and I/O costs, but the computational cost of matrix multiplications remained a bottleneck, and optimizing neural networks with extremely low-bit parameters is challenging.  Two techniques help to address this problem. Sparsification reduces the number of computations by pruning activations with smaller magnitudes. This is particularly useful in LLMs because activation values tend to have a long-tailed distribution, with a few very large values and many small ones.   Quantization, on the other hand, uses a smaller number of bits to represent activations, reducing the computational and memory cost of processing them. However, simply lowering the precision of activations can lead to significant quantization errors and performance degradation. Furthermore, combining sparsification and quantization is challenging, and presents special problems when training 1-bit LLMs.  “Both quantization and sparsification introduce non-differentiable operations, making gradient computation during training particularly challenging,” Furu Wei, Partner Research Manager at Microsoft Research, told VentureBeat. Gradient computation is essential for calculating errors and updating parameters when training neural networks. The researchers also had to ensure that their techniques could be implemented efficiently on existing hardware while maintaining the benefits of both sparsification and quantization. BitNet a4.8 BitNet a4.8 transformer architecture (source: arXiv) BitNet a4.8 addresses the challenges of optimizing 1-bit LLMs through what the researchers describe as “hybrid quantization and sparsification.” They achieved this by designing an architecture that selectively applies quantization or sparsification to different components of the model based on the specific distribution pattern of activations. The architecture uses 4-bit activations for inputs to attention and feed-forward network (FFN) layers. It uses sparsification with 8 bits for intermediate states, keeping only the top 55% of the parameters. The architecture is also optimized to take advantage of existing hardware. “With BitNet b1.58, the inference bottleneck of 1-bit LLMs switches from memory/IO to computation, which is constrained by the activation bits (i.e., 8-bit in BitNet b1.58),” Wei said. “In BitNet a4.8, we push the activation bits to 4-bit so that we can leverage 4-bit kernels (e.g., INT4/FP4) to bring 2x speed up for LLM inference on the GPU devices. The combination of 1-bit model weights from BitNet b1.58 and 4-bit activations from BitNet a4.8 effectively addresses both memory/IO and computational constraints in LLM inference.” BitNet a4.8 also uses 3-bit values to represent the key (K) and value (V) states in the attention mechanism. The KV cache is a crucial component of transformer models. It stores the representations of previous tokens in the sequence. By lowering the precision of KV cache values, BitNet a4.8 further reduces memory requirements, especially when dealing with long sequences.  The promise of BitNet a4.8 Experimental results show that BitNet a4.8 delivers performance comparable to its predecessor BitNet b1.58 while using less compute and memory. Compared to full-precision Llama models, BitNet a4.8 reduces memory usage by a factor of 10 and achieves 4x speedup. Compared to BitNet b1.58, it achieves a 2x speedup through 4-bit activation kernels. But the design can deliver much more. “The estimated computation improvement is based on the existing hardware (GPU),” Wei said. “With hardware specifically optimized for 1-bit LLMs, the computation improvements can be significantly enhanced. BitNet introduces a new computation paradigm that minimizes the need for matrix multiplication, a primary focus in current hardware design optimization.” The efficiency of BitNet a4.8 makes it particularly suited for deploying LLMs at the edge and on resource-constrained devices. This can have important implications for privacy and security. By enabling on-device LLMs, users can benefit from the power of these models without needing to send their data to the cloud. Wei and his team are continuing their work on 1-bit LLMs. “We continue to advance our research and vision for the era of 1-bit LLMs,” Wei said. “While our current focus is on model architecture and software support (i.e., bitnet.cpp), we aim to explore the co-design and co-evolution of model architecture and hardware to fully unlock the potential of 1-bit LLMs.” source

At the risk of dating myself, I’m a proud member of Gen X. So, perhaps I can be forgiven for channeling Jerry Maguire and George Constanza when I think about the key takeaways from the latest Forrester Wave™ evaluation of customer feedback management (CFM) solutions. My conversations with the customer references for the nine vendors included in this Forrester Wave reinforced my evaluation and what I have been hearing from our clients: The “how” is often more important than the “what” when buying CFM solutions. In a market where vendors’ technical offerings are incredibly similar, CFM buyers need to consider the wise words of some 1990s icons: “Help me help you.” CFM vendors offer a range of services, but buyers often focus on the tech and don’t budget for adequate services. The need for services does not necessarily decrease with organizational maturity. For example, some reference customers for this research use strategic vendor services for advanced predictive and what-if analyses as well as for simpler survey design and deployments — freeing up their team for higher-value work. Buyers should assess their own maturity, resources, and speed requirements when considering the role of services. “It’s not you, its me.” Nearly half of the customer references I spoke with cited internal reasons for not using some of the features offered by their CFM solution. For example, most reference customers underuse the analytics tools offered by their CFM. For some organizations, internal policies limit their ability to bring data into the CFM to perform these analyses. Other organizations prefer to pull data out of the CFM because they do analysis in other tools. Either way, CX leaders need to understand their organization’s policies and processes when selecting a CFM to ensure contextual fit. And because no blog in 2024 is complete with requisite mention of generative AI (genAI), CFM buyers are still looking for vendors to “Show me the money!” GenAI-enabled features, like natural language interfaces to query data, are top of mind for buyers. Many vendors are rushing to respond, but expectations still exceed reality. As we’ve written recently, CX teams should keep expectations in check and focus on employee-facing use cases for genAI first. Forrester clients: Check out the latest new report for more insights, and book a session with me if you’d like to go deeper. source