A presentation for Straker’s August 20 annual meeting made no bones about why it is repositioning itself as an AI firm.
Staker’s shares surged during the pandemic tech boom that peaked in early 2022 (when its shares topped A$2). Since then, they’ve been in a long decline. A table in its AGM report shows its peers have suffered a similar sharemarket fate.
The LLMs (large language models) like OpenAI’s ChatGPT, first released in November 2022, accelerated the rise of free language translation tools.
“We know investors have seen market caps for traditional language service providers decline,” Staker’s AGM presentation said.
“Straker has been caught up in this industry-wide sentiment.”
Ensuring Straker is seen as an AI company has become a key goal for management, investors were told.
There have been marketing tweaks. In 2021, the firm’s name was changed from Straker Translations to Straker – or, often, in a nod to its new website address, Straker.AI.
But Straker says the changes are more than skin deep.
While many “AI” startups are “ChatGPT wrappers”, selling services based on OpenAI’s chatbot, Straker says his firm has a deep technical and commercial relationship with IBM. Straker is building its own AI models, and his staff are working with IBM’s to jointly pursue revenue opportunities.
He says the new $28m deal with Big Blue “is a pretty unique thing for a New Zealand company to achieve, and it expands our relationship with IBM at a strategic level where our AI models will run inside of the Watson X infrastructure. IBM teams globally can sell our technology”.
“It expands our relationship beyond the translation services space and into the general AI space.
“Given our long history in building machine learning models, this has naturally led us to become experts in building small language models. We would be one of the only companies in New Zealand with extensive AI model-building capability.”
Revenue down, more red ink
On the face of things, Straker’s FY2025 results look challenging – and certainly didn’t get an enthusiastic reception from investors as its stock stayed in the doldrums.
Revenue fell 10% from $50m to $44.8m as net profit widened from $2.2m to $10.2m and the firm shed 40 staff as it reduced its headcount to 170.
But Straker says most of the red ink was non-cash impairments on previous acquisitions (which totalled $6.8m). Accounting changes to the way software is amortised.
Underlying numbers stronger
Straker says the global economic slump was a factor in the revenue dip, but the staff cuts were largely because of AI automation and would have happened regardless.
However, he focuses on adjusted earnings before interest, tax, depreciation and amortisation (ebitda), which increased 18% to a record $4.8m.
“We finished the year with $12.9m in cash [from $12.1m in FY2024] and no debt. This consistent financial strength means Straker has not had to raise equity funding to invest in new AI-driven product offerings,” Straker says.
He says the percentage of higher-margin revenue from AI-driven products is on the rise and could account for 60% of turnover within six months. His firm is forecasting AI will account for 70% of its revenue within 18 months.
“The pivot into AI has had its moments, but the winners in this space are going to be big winners,” Straker says.
“And it’s great for us to be starting to build out an AI ecosystem from New Zealand that can sell to the world, rather than NZ just being consumers of AI.”
AI, with a touch of human checking
AI’s reputation is mixed in its early stages. One example: A new study from the University of Auckland says for every minute saved by artificial intelligence in the legal profession, an equal amount of time is lost to fact-checking.
Straker is honing its AI models, but its secret sauce ingredients continue to include fact-checking humans on its staff.
Straker says there are now thousands of IBM employees using his firm’s “Verify”, which the firm describes as an “AI content quality verification tool”, which has AI doing the heavy lifting but also “humans doing the final touches”. It can be embedded in platforms like Slack.
Straker says the benefits of hundreds of billions being spent on AI infrastructure today won’t fully flow through to firms like his for 18 to 24 months.
The firm’s share price shows many investors are yet to be convinced. But Straker says his firm’s largest shareholders are in it for the long haul. Beyond his own holding, David Kirk’s Bailador Investments has been a longtime cornerstone backer.
Chris Keall is an Auckland-based member of the Herald’s business team. He joined the Herald in 2018 and is the technology editor and a senior business writer.