A month after Raj joined Simwood and took on our ambitious sales and marketing team from my temporary tenure in that field, I’ve had a lot of time to think, rehearse, challenge and generally noodle on the direction of our industry.
Following a week at ITW, networking, debating and learning, I am now pretty confident that in years to come, our current era of AI strategy will be taught in MBA programmes. Those that watched the recent SimCron will have a sense of where this is going!
The eager business school students of the future will be learning about OpenAI, Anthropic et al with the same 20/20 vision of hindsight I did, when learning about Southwest Airlines, IBM, Henry Ford and the whole myriad of innovations in the 1900s.
I always caveat my predictions with a warning that they will be wrong, but I am increasingly sure in this case, I might at least be in the ball park.
In those business school classrooms, in 2040, there will be a lecture about 2026’s economy having four distinct concurrent AI strategies, with one being a victor. With all such good backward looking business academia, there will be some colourful descriptors of each group, so here follows my crystal-ball gazing.
The Hares
Promises of milk and honey, valuations inflated on such promises, adoptions driven by expectations of the same. Then reality bites. Klarna found this out the hard way, and Microsoft just found out in a different, but still relevant, way.
These examples sit in two distinct categories. Firstly, false hope that AI can replace swathes of customer-service humans as we saw with Klarna, or, secondly, Microsoft’s use of Claude turning out to be materially more expensive than anticipated – one can only assume the expectation is driven by a benchmark of some keen junior developers. Claude is a good tool, and one that has several roles within Simwood, but it is not a replacement for a human here.
While distinct, these categories have a common lineage – a market bubble frothing expectations of CEOs and CFOs to realise AI “savings”, and queues of shiny-suited salespeople lining up to help those pressured C-suite execs deliver those “savings” with various degrees of promise.
On the one side, we have press releases saying X company has adopted Y AI to achieve savings. Then a while later, X company is ripping out Y AI for not performing to expectations.
We recently advised Ofcom to dig into this effect. From a BT/EE press release a figure was quoted of customer facing AI interactions, and our regulator had placed some weight on that. But the press release was weasel worded – AI being involved in a customer interaction, be it starting one off before a frustrated customer eventually had to rehash their conversation with a human, is very different to a satisfied customer having resolved an issue in an AI only journey. Informed by conversations with companies who work with large scale (as in very large scale – think international airlines) contact centre operations, I would wager BT/EE fall into the former category.
In an early draft, and on SimCron, I called this category the “Snake Oil Salesmen”. These exist; but motive matters. There are differences between the Microsoft/Anthropic story and the Klarna story, not least that a good tool was badly applied, and a bad tool was very badly applied.
My naming rationale also comes from not tarring every Hare with the same pejorative brush – not least because OpenAI and Anthropic et al are placing a multi-year, even decades, long bet that their technology will evolve to scale alongside their ambition and provide the fabric others build upon – even if it’s not obvious to us (and especially the CEO of IBM) how that ambition will generate a return. But again – motive matters.
That said, whichever angle we view this from today, there are substantial elements of overpromise and underdelivery and a too rapid implementation of technology, with a rationale of appeasing market expectations over substance.
The Ostriches
Fun fact: Ostriches don’t actually bury their heads in the sand. But that’s what I am talking about here. The bracket of “AI” vendors whose AI is a rapidly rebadged crummy chatbot or search. Slack (yes, Salesforce!) are in this category. They launched an AI feature and expected to monetise it. Fast forward a while… and all of a sudden they are bundling it in with existing subscriptions, but [jacking those up]. Having tried said feature, I suspect the change in approach was because the feature in question had been adopted poorly, as a direct consequence of it being, well, poor.
Rinse and repeat among a lot of the existing players in any space: in ours, I assume they got their marketing departments to just do a find and replace on their website and collateral of last week’s buzzword with “AI’. No-one is talking about 5G any more are they? Yet, the same people that were, are suddenly in AI? Every 5G-enabled feature is now somehow magically AI-powered instead? Even our own superhuman development team hasn’t turned our speedboat on a dime, and these Ostriches are bloated supertankers.
Caught napping, this group is scrambling to keep up and not be made redundant as CEOs and CFOs rip out old systems for the new to meet market expectations for savings.
Valuable blue-chip customers are at risk because the Hares are wining and dining their customers’ executives. AI is the hook by which one ERP vendor can persuade a prospect to undergo the cost and inconvenience of displacing the existing ERP vendor – and we know from the excellent Cavell conferences that this is a visible threat and trend to the legacy software vendors.
The need for the Ostriches to have a story – any story – is the driver here. Minimum Viable Product is no longer a deliberate development strategy, it’s a life raft.
The Dinosaurs
While Ostriches are doing something, in this category we have those that are saying nothing, and doing nothing, because either they don’t believe in it, see the relevance of it, or it’s just too hard and difficult.
History is littered with these examples. Blockbuster video, who failed to see the importance of streaming. Polaroid and Kodak, who dismissed digital cameras. Nokia and Blackberry, who missed the shift to modern smartphones.
Here, the strategy is to not have a strategy. These will largely be large incumbents whose vision is limited by some back-office savings, and not embracing established routes to market to grow value with new offerings and technology.
I did think long and hard about whether these are a form of Ostrich, but, unlike the Ostriches, whose failing is being caught short and playing catch up (either with poor AI like Slack, or by trying to make out they have a story when they don’t), the Dinosaurs aren’t doing anything.
Some Dinosaurs and Ostriches may eventually move. Salesforce, for example, has significant resources and is unlikely to be left completely behind. However, many smaller or slower incumbents do not have that luxury and risk remaining dinosaurs for longer than they realise.
Others are, frankly, already extinct — they just don’t know it yet. In telecoms particularly, the threats are significant. WhatsApp, Teams, and AI are just some of the asteroids heading toward traditional players. If I ran this analysis on Teams or WhatsApp, I suspect many of the same names would appear in the Dinosaur bucket.
The Tortoises
Hang on Pete? You’re talking about AI? An industry characterised by product development lifecycles in the weeks not months and years? The bleeding edge of the very best of human ingenuity and innovation? And you come out with a derogatory term for the (spoiler!) group you think will succeed?
Yes, the industry is full of lean, mean, sprinting hares, but as we know from Aesop’s Fables, the Tortoise wins.
These are the ones with niche specialities that don’t overpromise, but make humans more efficient. Our CTO’s current mantra is “augment, not replace”. We’d like to think our own [Conversational AI] is in this bracket, but in all seriousness, we will not be alone.
An AI that listens into a 999 call and prompts the call handling agent they miscategorised the urgency. A tool that whispers into the air traffic controller’s ear that he’s just cleared a fire truck to cross a runway a CRJ900 is on short final to. An AI that tells a pit boss which blackjack table he should focus his attention on.
This is evolution, not revolution. This is where analysts are already talking about Human+ Intelligence, over AI. All of these applications have a defined, and readily measurable, purpose, output, and return on investment. None require the laying off 700 people to just be rehired (yep, Klarna again), but they will give their users a competitive edge in their industry.
And that is precisely what has driven human entrepreneurship for centuries – good old fashioned corporate oneupmanship as companies innovate and advance beyond the laggards and incumbents who are in turn left to wither on the vine.
I am genuinely excited that we have an AI story that doesn’t promise to lay off half of our customers’ staff, but we also actually have a good one. Long live the tortoise.
