Uri Baruchin’s Post

“A pharma company stopped using Microsoft's Copilot AI tool, with an exec citing high cost and low value.” “He compared the slide-generation capability of Microsoft's AI tools to middle school presentations, according to a transcript of a call with the Morgan Stanley analysts that was included in their research note.” “The chief information officer of a pharma company paid extra to have 500 employees use Office 365 Copilot in the fourth quarter of 2023 and first quarter of 2024. This is Microsoft's much-heralded AI upgrade to its popular suite of productivity software.” He paid “about $34 per user a month” for the “E3 version of Microsoft's 365 software suite.” “Adding Copilot AI features costs another $30 per user a month. For 500 employees, that would roughly add an extra $180,000 a year.” “These new Microsoft tools are considered some of the premier examples of powerful artificial intelligence in action in the real world. Investors have bid up Microsoft and other big #tech shares massively, betting that this product and similar offerings will catch on with paying customers.” But the article says: “If a large pharma company can't see $180,000 of value in these tools, that's a problem that should worry the entire tech industry,” a worry that I have written about many times. I have trouble finding any examples of companies claiming productivity improvements from generative text AI, and only a few for generative video AI (marketing videos and helping with movie and TV shows). There was one AI feature that the IT executive found somewhat compelling: that “was the ability to archive and summarize video meetings on Microsoft's Teams app. But he said his legal team was wary of retaining meeting transcripts, so the pharma company didn't use that feature.” Other people have shared their concerns with me about the legal problems from storing their proprietary information like this. Investors are also concerned and thus there are many comments, both internal and external to Microsoft, about “whether its AI services will create enough value to persuade corporate customers to pay more.” “The company is amassing 1.8 million graphics processing units to build and run AI models and related products. It also has a plan to triple data-center capacity, mainly to handle AI workloads. Its capital expenditure hit a record $14 billion in the most recent quarter.” Many investors are concerned that this pharma company isn’t the only company that is finding Microsoft’s Copilot AI services, and those from other companies, not as beneficial as many proclaim. With a big runup in the market capitalizations of the big tech companies over the last 18 months, particularly of the so-called Magnificent 7, those concerns could spark a move in the opposite direction for the share prices of the Magnificent 7.   #technology #innovation #startups #ethics #artificialintelligence https://lnkd.in/gJ_FHBAB

A basic Thursday reflection on: the danger of context-free advice (and the importance of basics). Imagine, if you will, that there exists two companies - Company A and Company B - that appear very similar. Their size is about the same, their revenue is about the same, their brands are about the same, and their products are about the same: they both produce aluminium cans. Now imagine that you were hired by one of the companies to do its marketing. How would you do it? From experience, having asked the question a fair few times over the years to clients and audiences alike, answers tend to end up being variations on a similar theme - the approaches are nearlly identical regardless of which company happened to be the client or employer. Yet, as I am sure you have already surmised, there is a twist. As it turns out, Company A does long runs to a handful of big clients; shipments are few. Marginal costs are low due to economies of scale, enabling a low price. However, the time to market is long. Company B is its opposite. It does short runs to numerous small clients; shipments are many. Marginal costs are consequently high, forcing a higher price, but justified by a short time to market. This information ought to change one’s marketing entirely, particularly in the near-term. If one is employed by Company A and goes after the kind of customer that Company B has, it might be that the size of the order is too small; the company cannot employ economies of scale, costs go up, margins go down, and profitability with them. If one is conversely employed by Company B and goes after the kind of customer that Company A has, it might be that the size of the order is too large; the company has no capacity to fill it. One would have created a demand for which there was no supply. This is a very basic example, yet time and again, I see managers and marketers fail to consider the kind of details mentioned. Instead, the conversation turns into one of context-free solutions, campaigns, clichés about brands and customers, or quotes by Steve Jobs. At the end of the proverbial day, if you do not understand the context in which a firm acts - if you provide the answers before asking the necessary questions - you turn marketing into a game of Jeopardy. There is no such thing as a context-free solution in a context-specific reality, no matter how much business book authors would have you believe otherwise. The basics may be basic, but absorbing the cost of doing them properly usually saves you a whole lot of money later down the line.

Where does the majority of your freelance work come from? Is it your own professional network (i.e. previous clients, colleagues of colleagues), is it via recruiters, via your own cold outreach or sales channels, or via jobs boards and platforms (like YunoJuno)? Or is it something else? Is your own network still the most important factor in finding work (and if so, what does that mean for people trying to break into an industry?)

*Morning everyone!* Don't forget to pop in and see us at our School of Communication Arts 2.0 Open Portfolio Day at The Drum in Shoreditch (25-27 Curtain Rd, EC2A 3LT) on this momentous palindromic day, 24.7.24 (aka Wednesday) from 9–5pm. If you can't make it, please contact Marcia Miller to explain how you missed such an opportunity, and maybe we'll forgive you. Catch us as workable scamps before we blossom into print-ready, copy-checked, client-approved, FINAL FINAL shiny posters.