Skillcast: Fast AI Adoption, Slowing ARR — and £13.6m Burning a Hole
The compliance-training SaaS grew revenue 10% and EBITDA 48%, yet the whole call circled back to a cash pile investors want back and an ARR trend management keeps promising to fix.
SKL.L · Earnings Call · 2026-10-02
An AI story that is finally shipping product
Skillcast Group plc is a £37.7m-market-cap UK compliance-training SaaS business. On the surface, its H1 2026 print (reported October 2) is more of the same steady compounding this company has been grinding out for years: subscription revenue up 15% to £7.4m, total revenue up 10% to £8.2m, gross margin up 1.6 points to 77.1%, EBITDA up 48% to £1.0m, and £13.6m of net cash with zero debt. What is genuinely new is the AI product cycle. In 2025 the company shipped its first AI tool, the Ask Aida compliance assistant; in July, just after the half closed, it launched Edit with Aida, an AI product that lets clients customise courses themselves. CFO Richard Steele flagged the take-up: “within two months of launching, by the end of September, 15% of all clients were using this tool” — Richard Steele, Investor Relations or similar, presenting the results · 2026-10-02. More is queued — Learn with Aida and Translate with Aida later this year, and Create with Aida on the 2027 roadmap, framed as agentic AI built inside a walled, secure environment. The AI pitch is doing double duty. Externally it is differentiation; internally it is margin. Management held headcount flat (+1% year on year) while revenue grew, and pointed to that discipline as the source of operational gearing — cost of sales and overheads rising slower than sales. This is the cleanest part of the story, and the market theme around it is real: AI adoption at scale keeps appearing across the curated global tape, with AI at scale and open-weight model names among the recent 90-day advancers.But the growth engine is decelerating
Here is the tension the headline hides. Annualised recurring revenue rose 14% to £14.5m, but that is a step down, and management said so plainly. The Rule of 40 — ARR growth plus EBITDA margin — came in at 26%, and Steele conceded it was “6% below the H1 last year due to lower ARR growth” — Richard Steele, Investor Relations or similar, presenting the results · 2026-10-02. ARR growth is the first keyword on the company's own trajectory this quarter for a reason: it is the number everyone is watching, and it is softening. The customer mechanics underneath it are worth slowing down for. Net retention held at 100%, flat year on year — but only because price rises of 5% (up from 3%) offset churn that rose to 9% from 7%. Steele was candid: “We did have slightly higher churn of 9% rather than 7% the year before. While slightly disappointing, there is no one factor that is particularly driving this” — Richard Steele, Investor Relations or similar, presenting the results · 2026-10-02. When you need a fatter price increase just to stand still on net retention, the underlying upsell engine is doing less heavy lifting than the narrative implies. Client count rose 6% to 1,250 and average ARR per client rose 7% to just over £11,402 — solid, but the mix is shifting toward simply charging the existing base more. The second soft spot is professional services, which fell 22% (after a 7% drop the year before). Management's explanation is unusually blunt and, frankly, the most macro-relevant line in the call: “There are still some people that want to spend lots of money on people like us to actually write bespoke courses from scratch. But it is understandably reducing” — Richard Steele, Investor Relations or similar, presenting the results · 2026-10-02. That is a company telling you AI is eating its own bespoke-services line — the same disruption the market has been pricing into software broadly, where software exposure and AI risk have been prominent curated themes. Skillcast's answer is to lean into the disruption rather than defend the old revenue.The cash pile and the question management keeps deferring
The Q&A was dominated not by AI but by capital allocation. With £13.6m of surplus cash sitting against a £37.7m market cap — more than a third of the equity — investors came loaded for a buyback. Steele's answer was a masterclass in deferral:So the buyback is real but not soon; the timeline is pushed to "by the end of 2027," with the stated priority being content acquisitions — buying ARR to leverage the existing platform and cost base, potentially into adjacent sectors like GDPR, public sector, or health and safety. That is a coherent strategy for a business that already runs 90% recurring revenue and needs new logos to reaccelerate ARR. The risk is that "confident progress on M&A" becomes a standing line that never converts, while the cash drags. There is a subtle signal in the company's own keyword list: retail investor appears as a top theme, born from the Q&A exchange about Mello events and attracting new shareholders. A £37.7m micro-cap whose investor dialogue is as much about liquidity and share price as about product is telling you something about the shareholder base it is trying to court. Management referenced a perceived fair value around £0.90 versus a deeply discounted share price — and, tellingly, acknowledged that a buyback could hurt the thin AIM liquidity. It is a genuinely awkward spot: the cheapest use of cash might also be the one that damages the stock's tradability.We are aware that we have got a significantly large cash balance. In the short term, we will be looking at, hopefully, spending this on M&A. If that doesn't happen, I am sure sometime during or by the end of 2027, we will have to come up with other ways. Share buyback may be one option, but also a larger dividend could be an alternative option.