AI, business growth and the race to work out what really matters
I’ve been in two interesting rooms recently.
One was the Incredible IT Leaders, a group of CEOs and founders of privately owned and PE-backed technology businesses. The other was our Growth Forum for CROs, CMOs and commercial leaders, many working for exactly the kind of businesses represented in the CEO room.
Different jobs. Different pressures. But increasingly, the same questions. Where are people placing the bets? Where is AI genuinely changing the equation? What does a good investment look like?
And, as one slightly more stick-in-the-mud asked:
“Is AI just good at doing stuff that was never any use anyway?”
It got a laugh. But it might have been one of the best questions of the day, because there is an awful lot of doing going on.
Everybody is experimenting
There was certainly no shortage of AI activity in either room. People are using it for coding, content, analysis, campaign creation, data enrichment, reporting, research and automation. Some businesses are taking an increasingly “AI first” approach, ie, before recruiting another person, asking whether some or all of the job in question could be done by AI.
Others are beginning to turn their own experience into propositions for customers: AI readiness assessments, consultancy, optimisation services and help navigating an increasingly bewildering collection of tools and subscriptions.
One CEO talked about discovering just how much his business was spending every month across numerous AI subscriptions and subsequently cutting that cost roughly in half. Interestingly, that experience has itself helped identify a potential customer proposition: helping clients work out what they need rather than simply accumulating more tools.
So perhaps managing AI sprawl is going to become as important as adopting AI in the first place? There is a lot of experimentation going on. But there is also quite a lot of spend creep.
But is it really driving growth?
And that’s perhaps where the distinction between activity and value starts to matter.
Among the Growth Forum group, there were plenty of examples of AI allowing businesses to do things faster, more cheaply or with fewer people, but much less evidence that AI itself was directly creating revenue. That feels like quite an important distinction.
If a marketing team can produce three times as much content, analyse its market faster or create campaigns that previously required additional people, that’s clearly useful. But back to our stick-in-the-mud… Is it helping us do more useful stuff, or just helping us do more stuff?
One of the best observations in the Growth Forum was that AI is collapsing the capability gap but increasing the prioritisation problem. One person can now do things that would previously have required several different skills or people. Great. But which things should they do?
AI has massively increased our capacity to create. It hasn’t given us any more hours in the day, and it certainly hasn’t removed the need to decide what matters, suggesting that judgement becomes more important rather than less.
No point automating chaos
Perhaps this question of usefulness also explains why one of the less glamourous investments discussed by the CEOs caught my attention.
This business has brought in RevOps expertise to analyse its data, CRM, processes and hand-off points and understand where opportunities were being lost. Nothing particularly futuristic about that. But potentially a much better investment than another piece of technology.
As many have observed in previous meetings, there isn’t much value in sophisticated automation until you’ve got the underlying processes right. Or, put rather better:
“No point automating chaos.”
We’d had almost the same conversation in the Growth Forum. A manual process might take somebody 20 minutes. Automating it properly might take hours because you first need to understand every step, exception and data dependency.
AI doesn’t magically remove that work. However, when it’s done it’s obviously quick to do again if you’ve got it right. Sometimes trying to automate something just exposes the fact that nobody really understands the process, which is quite a useful blind spot to identify!
So perhaps the question isn’t just “what can we automate?” but “what is worth automating?”
Where are the commercial models?
At some point in the CEO discussion, someone asked: “What is your ROI on AI for MRR”
It took me a moment to digest the question. Perhaps three-letter acronyms are simply the price of entry these days.
Underneath it is a serious commercial question. If AI allows us to do something in minutes that previously took hours or days, how do you charge the customer?
A commercial director at the Growth Forum described exactly that conversation with a prospect questioning the traditional day-rate model. If technology has dramatically reduced the effort required, why would the customer continue paying on the old basis?
Fair question, so what happens next? How do we move towards value or outcome-based pricing? What role will there be for subscriptions, tokens, managed AI services, AI wraparounds?
We heard examples of MSPs beginning to package AI-related services and others experimenting with AI readiness or assessment offers as a way into new customers.
But what does a good commercial model look like? Are customers ready to buy in completely different ways? And with token-based models, are we creating a better way to buy technology, or exploiting customers who consume it inefficiently?
Meanwhile, the buyer has AI too
Perhaps we’re looking at this from the wrong side of the table. We’re busy talking about what we can do with AI, but our customers have exactly the same tools.
Sellers are no longer the gatekeepers of knowledge. Buyers can research a problem, interrogate possible solutions, compare suppliers and arrive at the first conversation with a pretty well-developed view of what they think they need.
In one Sales Director’s experience customers are already using AI to effectively fact-check his supplier recommendations. Which creates another slightly uncomfortable question: What happens when a prospect asks AI to compare six MSPs, software companies or consultancies and it concludes that they all look basically the same?
AI can make weak differentiation much easier to expose, which takes us back to some very traditional questions. What are you good at? How are you different? Why would somebody choose you? What evidence is there to support it? Who else says you’re good?
If producing information becomes almost free, perhaps trusted expertise, opinion, reputation and proof become more valuable rather than less.
Are we going to end up craving human beings?
Both groups are made up of tech businesses, so guess what, there is a lot of technology chat in both. Which made it more interesting that some of the strongest growth bets discussed in Growth Forum were remarkably human.
Relationships, referrals, small, carefully curated events, practitioner-led content, getting closer to existing customers and building relationships with a relatively small number of influential partners all came up.
One business is deliberately recruiting a salesperson despite taking an AI-first approach elsewhere. Why? Because they don’t want to automate away the personality and there is a suggestion that the pendulum is swinging back towards authenticity.
If everybody can produce unlimited quantities of perfectly competent AI-generated emails, posts, articles and presentations, will our prospects start placing a premium on things that quite obviously came from an actual person? Will the slightly imperfect email become more credible than the beautifully structured one? And how much of selling can really be automated anyway?
As one sales director said, a client texting him about football on a Saturday night might tell you more about the strength of a commercial relationship than the percentage probability sitting against the opportunity in the CRM.
Not terribly scientific, but still quite important.
So where would you put £100k?
The Growth Forum is made up of CROs, Sales Directors and Marketing Directors from small tech businesses. We asked them, if somebody gave you another £100,000 tomorrow to spend on growth, where would you put it?
The answers were reassuringly different.
Paid digital. Content and credibility. Events and relationship-building. Making better use of contacts already sitting in the CRM. Conversion. Smaller gateway propositions that allow a customer to buy something, establish trust and expand from there.
Most said they’d invest a bit in tooling and automation, but nobody said: “I’d spend the whole lot on AI.”
Perhaps that the answer to the question in the title. There isn’t a universal AI strategy, marketing channel or growth tactic that everybody should suddenly be betting on. The better question seems to be:
Where is growth getting stuck?
If you have plenty of leads but poor conversion, buying more leads isn’t going to fix it. If nobody knows you exist, endlessly optimising your sales process isn’t going to fix that either.
If your proposition looks exactly like everybody else’s, automating your outbound just allows you to tell more people, more efficiently, that you’re the same as your competitors. And if you’ve got thousands of contacts sitting untouched in a CRM, perhaps you don’t need another shiny new lead-generation tool.
Gut feel and guard rails
How do you operate when the technology is changing this quickly?
One CEO used an expression I really liked: “Gut feel and guard rails.” Experiment, try things, accept that some won’t work, but set the boundaries within which people are free to experiment.
There was an interesting observation from Sean O’Shea, a culture specialist who joined the IT Leaders Forum. He works with firms to help them engineer their culture and environment for the future. His suggestion was that allowing a new employee to fail at something during their first few weeks can be one of the healthiest things an organisation does. Make a mistake. Own it. Learn from it. Move on. If we’re genuinely in a period of experimentation, we must give people enough space to try new things.
Which makes one wonder whether the businesses getting furthest with AI won’t necessarily be the ones with the biggest budgets, the most subscriptions or the greatest number of agents. Maybe they’ll be the ones that are prepared to try enough things to learn quickly but are disciplined enough to stop doing the things that aren’t useful.
Back to basics (but don’t sit on the sidelines)
For all the discussion about enormous technological change, one of my favourite comments came from a CEO of a rapidly growing cyber security business:
“Every time we focus on the basics, things just get better.”
Maybe that’s rather unexciting but it strikes a chord. Get the proposition right, understand where growth is getting stuck, get closer to customers, build better relationships, understand the numbers, fix the processes. And only then work out where AI can make those things faster, better or more scalable.
Sitting on the sidelines is not an option either. Another takeaway from the CEO discussion was simply: “You’ve got to be in it to win it.”
The models are changing quickly. The use cases are changing. The economics are changing. We are all learning.
So perhaps those two slightly contradictory thoughts aren’t contradictory at all. Experiment, learn and keep moving, but don’t mistake experimentation for strategy.
The dust will settle. Some of today’s tools will disappear. Some of today’s exciting AI propositions will become completely ordinary. And presumably we’ll eventually get rather better at understanding what a sensible level of AI investment looks like.
For now, I’m left with more questions than answers. Where are you placing the bets? What are you already doing differently? And when you strop away the excitement and novelty, is AI helping you make money, save money or do something genuinely useful?
Or are we just getting extraordinarily good at doing more stuff?