Shane Gallwey, CEO of Guinness Ventures in
18 May 2026
In February this year, a single AI upgrade sent shockwaves through global markets. When Anthropic launched its Claude Sonnet model with dramatically enhanced AI coding capabilities, investors did not wait to understand the implications. They sold.
The Morningstar US Software Index fell 20% in the first four months of 2026 while the broader market went nowhere. Salesforce lost nearly a third of its value. ServiceNow and Adobe fell further still.
The press coined a name for it: the SaaSpocalypse.
The market has overreacted
The sell-off in Software-as-a-Service (SaaS) companies feels somewhat indiscriminate. The reasoning goes along the following lines: companies can now use AI to build software far faster and cheaper than ever before. Clients will leave or demand ever-increasing discounts. The software-per-seat model will be eroded. Challengers will usurp the incumbents.
The reality is more nuanced, and investors have undoubtedly overreacted. The arrival of AI tools that enable startling productivity will drive seismic change. But having spoken to our investee companies, the picture that emerges is more optimistic: many software businesses will be net beneficiaries of these tools.
Some are better positioned than others. The difference depends on their intellectual property, market position, customer embedment, data assets and the replicability of their offering.
The central question
Can AI replace the software, or does it make the software more valuable?
That distinction matters. AI will make some software easier to replicate. But it can also increase the value of products that own proprietary data, sit inside critical workflows or govern important business processes.
In other words, code alone may become cheaper. But trust, data, distribution, compliance, integration and customer behaviour remain hard.
Which software companies can benefit from AI?
In response to the turmoil, Goldman Sachs Research published its AI Impact Framework in February 2026, identifying the factors that determine whether software companies will benefit or suffer from increasingly powerful AI tools.
The research examines whether a software platform will continue to be needed by clients, or whether it can be bypassed by AI agents. It also looks at monetisation, asking whether revenues are tied simply to providing a service that could be replicated, or whether there is something fundamental to the product, such as data that is not available elsewhere.
It also identifies the importance of system-of-record ownership, where the software platform governs approvals, compliance or operational processes. These are harder to displace because they sit in the operational layer of clients’ businesses.
How AI is changing software moats
Goldman Sachs also examines the concept of the moat. In the past, software development time provided a protective barrier against competition. A company could say: it would take years for a competitor to build software like ours.
That is no longer enough. But data, and the way that data is integrated into the operational records of a business, remains a defensive characteristic that AI cannot simply replicate.
The software businesses best placed to benefit
The companies best positioned to benefit from AI either own the data, or sit at the infrastructure or governance layer of a client’s business and are already demonstrating revenue from AI services, or both.
Conversely, the software businesses that may struggle are those that simply drive productivity without any of the above features, because they are addressing tasks that AI is now beginning to perform directly.
By this logic, a Customer Relationship Management business like HubSpot or Salesforce could be replicated and replaced. It does not own any data. It organises the data owned by clients and presents it back to them. Additionally, as clients adopt AI and employ fewer humans, seat-based revenues may stop growing.
This is potentially true, but it is only part of the story. Salesforce is ahead of the game and has built its own AI workforce, Agentforce, which provides AI agents to complete tasks for clients, such as resolving a customer service ticket or qualifying a sales lead. It is not just about categorising client records. It is built on trust and brand, reliability and security, and the ability to support and upgrade existing software.
What our Guinness Ventures portfolio companies are saying
Turning to some of our own portfolio companies, the picture is similar. These businesses see AI as both a threat and a huge opportunity, and all believe they will be net beneficiaries of the changes underway.
Portfolio perspective
Dragonfly.AI
Dragonfly.AI, in which Guinness Ventures first invested in 2022, uses a patented, biologically inspired algorithm, developed in partnership with Queen Mary University of London, to predict how creative content will perform before it goes to market. Clients including Nestlé, PepsiCo, Unilever and L’Oréal use it to optimise everything from packaging to digital advertising.
Steve King, Dragonfly.AI’s CEO, is unequivocal about the direction of travel: “Absolutely, we benefit. We see it as an arms race, and investing in AI early is giving us an edge in software engineering, operations, finance, customer success and marketing.”
There is a compelling urgency to his thinking: “The risk is that the barrier will be much lower in 12 to 24 months, at which point you’re doing this to catch up rather than get ahead.”
Dragonfly.AI’s Chief Technical Officer, James Harvey, has built AI-powered workflows that distribute content across multiple channels directly. AI is also being used to improve data quality and allow sales teams to be hands-free on data entry.
On the question of defensibility, Steve’s answer reflects the same conclusion to which the Goldman Sachs report points: the code itself is the least important part.
“Our codebase will always be under threat. However, it is the data assets, the patent, and most importantly the implementation parts of the product that sustain value: being part of an existing workflow, training, proving accuracy and ROI.”
Steve King, CEO, Dragonfly.AI, 4 May 2026
James Harvey puts it equally directly: “The unit cost of product design and software engineering will drop significantly over the next 12 to 24 months, so value and defensibility have to come from other places: IP, data, velocity, distribution, brand equity, talent.”
Steve argues that core coding and development represent only around 25% of the total cost of enterprise software. The remainder, including testing, integration, security and compliance, infrastructure, training, change management, ROI measurement, support and ongoing maintenance, is where the real complexity and switching costs sit. AI can accelerate the first 25%. It does not touch the rest.
Dragonfly.AI is also seeing its market positioning shift as customers become more sophisticated about AI. James sees an uptick in API integration requests by brands building internal AI tools as evidence of this.
Portfolio perspective
Fifty
Fifty, backed by Guinness Ventures since 2019, is a technology-led media partner that uses proprietary audience data and AI to help brands understand and reach their customers. Its CEO, Simon Shaw, describes what Fifty has built as a global, network-scale data asset that underpins everything the platform does.
Simon recognises the threat: “AI is fundamentally changing the economics of build versus buy. Historically, most organisations struggled to build effective technology internally, which created a structural advantage for external platforms. That barrier is now eroding.”
But he is equally direct about why Fifty is positioned to benefit: “AI amplifies the value of our core assets, particularly our proprietary data and integrated workflows, rather than commoditising them.”
The most striking evidence of this is operational. Over the past three years, Fifty has reduced headcount while increasing output, moving the business to profitability.
“Build has often been misunderstood as simply writing code. In reality, successful systems require far more: deeply considered product design, structured workflows, high-quality data, and continuous iteration.”
Simon Shaw, CEO, Fifty, 7 May 2026
He expects internally built, AI-enabled platforms to emerge, many of which will be subscale or ineffective, and anticipates that the market will correct over time. That view mirrors the historical pattern of previous technology transitions.
Simon also sees brands deploying their own AI models, where the quality and structure of the data those models are trained on becomes critical. “Generic models without proprietary inputs risk producing undifferentiated outputs.” Fifty’s audience frameworks are therefore becoming more valuable, not less, and Simon reports early demand from clients exploring licensing models to use Fifty’s data to power their own AI initiatives.
The product is evolving from a service into an infrastructure layer.
Portfolio perspective
Aptem
Aptem, a Guinness Ventures portfolio company since 2018, is an end-to-end apprenticeship management platform used by nearly 200 customers including training providers and universities. Its product sits at the compliance and governance heart of one of the most heavily regulated corners of the UK education system.
Rien Sach, Aptem’s CEO, is buoyant about the opportunity: “I think the AI revolution is just a massive benefit for Aptem.”
His logic starts with Aptem’s structural position as the critical system of record at the heart of the learner journey.
Rien says: “As long as we can maintain that moat and expand it, the AI revolution is just a massive benefit.” The platform is embedded in compliance, funding calculations, Ofsted readiness and regulatory reporting.
On the operational side, Rien’s thinking about AI’s impact is ambitious. “The Aptem business has become more profitable through AI efficiencies as well as using our platform to build new revenue opportunities.”
The platform is already delivering measurable gains for customers: review summarisation saving tutors significant time, AI-powered marking aids, a virtual assistant for learner queries, and progress checkpoints to prevent learners falling behind. In aggregate, these represent up to a 50% reduction in admin time for key tutor tasks, freeing people up to focus on what Rien describes as “the things they really add value in.”
On competitive threats, Rien is pragmatic, not defensive. “I think a new entrant could probably rebuild our product if they had the right knowledge or expertise.”
“The same tools they can use to rebuild Aptem, we are using to make Aptem better, so we’ll always be ahead of them. We have the brand, the reputation, the market share, the customers, the expertise.”
Rien Sach, CEO, Aptem, 1 May 2026
He reiterates: “For Aptem, with its structural position and regulatory depth, the AI revolution is simply a massive benefit.”
The SaaS panic looks overblown
The market is driven by fear and greed.
Fear for software’s future has been in the ascendancy, but the panic will pass. The businesses that will struggle are those whose sole purpose was making humans more productive at tasks a machine can now perform for a fraction of the cost.
The businesses that will thrive, and in which Guinness Ventures has chosen to invest, are those built on proprietary data, deep regulatory integration and switching costs that no AI model can dissolve overnight.
We are not complacent about the change underway. But we are confident that the indiscriminate repricing of software stocks tells you more about investor sentiment than it does about the companies we back.
About Guinness Ventures
Guinness Ventures provides growth capital to ambitious companies through EIS and VCT strategies. We typically invest in revenue-generating UK companies with proven products, services or technology, clear customer demand and the potential to scale.
We back businesses across a range of sectors, including software, business services, healthcare and consumer. Alongside capital, we aim to support founders with strategic input, operational insight, introductions, future fundraising support and access to our wider network.
For investors
For investors and advisers, Guinness Ventures offers access to venture capital strategies focused on tax-efficient investment into qualifying growth companies. Please read the relevant product information and risk warnings before investing.
This article may provide information about Guinness Ventures portfolio companies, including recent activity and performance from company management, as well as facts relating to equity markets, and our own interpretation of market developments. Any investment decision should take account of the subjectivity of the comments contained in this article. This article is provided for information only and all the information contained in it is believed to be reliable but may be inaccurate or incomplete. Any opinions stated are honestly held at the time of writing but are not guaranteed. The contents of this article should not therefore be relied upon. It should not be taken as a recommendation to make an investment in the Guinness Ventures portfolio or to buy or sell individual securities, nor does it constitute an offer for sale.