Will Ashford has the kind of career potentially described as ‘LinkedIn fantasy’ - 4 years in the Australian Army followed by Private Equity in London, machine learning engineering at QuantumBlack (McKinsey’s AI consulting arm) and now at the helm of VC-backed AI startup TrueState.
To VCs everywhere, he is essentially ‘6’5, man in finance’. Working with corporate clients enabled Will to get on-the-ground experience to see the impact that AI can have, as well as spotting consistent themes across clients. He told me in an interview,
“[I] quickly realised there is far too much work required to implement modern AI algorithms in a private and secure way. (Businesses) are either investing in a huge tangle of third-party AI toys that are more hype than impact, or they are taking months or even years to build their own AI tools, often relying on costly external consultancy firms.”
The problem with old-school AI
“We would build bespoke old-school AI models to solve the exact business problem - for example, forecasting sales accurately to make better predictions and predicting customer churn to enable accounting teams.”
An admirable outcome, but ultimately, “it took a really long time, because underlying tech is complex and immature. This was before AI really took off.”
Even now, there are 2 major problems with AI implementation for enterprise customers
- Too lengthy: companies get caught up with the traditional ‘army of consultants’ building tonnes of supporting software to support their dream use cases. “It’s like they wanted to conduct some financial analysis but were forced to build their own Microsoft Excel along the way.”
- Technically complex: Adoption is not possible when organisations struggle to build internal teams required to manage and scale these systems over time.
The inception of TrueState
Will says “I had a view that a lot of the problems in the world came from the bias of decision-making.” TrueState started as an MVP to help non-tech users fine tune their own data. Now, the platform includes a suite of AI templates that can be integrated with customer operational data to solve their biggest and common challenges - eg. customer churn, operational supply and demand, inventory and cash flow.
What would take consultants and engineering teams up to 12 months to do, they can automate. “(We) want to make things more transparent through an algorithm approach, using frameworks to evolve companies into a better, more transparent way of working.” This means anyone in the organisation can use the tailorable AI platform, which includes non-technical staff, and an AI application can be built in as quick as half an hour. It’s like having McKinsey’s AI consultancy in a box.
Australia’s Appetite for Funding AI Startups
How is the Australian investment scene? “Launching a startup in Australia has been a blessing - if we launched in the US / San Fran / EU, the level of competition would have been huge”
TrueState was able to cut through the noise. It isn’t without its challenges, though, “we have an amazing team with high-quality code in our team, but it takes a really long time to find engineering talent” - something I hear time and time again from AI startups.
In what will come as a massive shock to literally no one, being a founder with deep domain expertise in an AI startup didn’t scare away investors. Will’s knowledge of enterprise customer pain points plus a track record of success in AI product delivery meant TrueState resonated with Aussie VCs. They have landed AUD $1.5m backing from some of the best in the country, including Airtree, Side Stage Ventures, Galileo Ventures, and Imprint Capital Partners.
For Ben Grabiner, partner at Side Stage Ventures: “TrueState (is) a new breed of AI platform enabling enterprises to iterate and execute on AI solutions fast, leverage multiple models and deliver integrated solutions that create real cost and productivity efficiencies.”
How was the team at True State able to pitch such a technical product? “It’s a huge challenge - for investors and customers. We approach this top-down sitting down and finding out how businesses work in a granular way, defining the problems given their lived experience before focusing on big picture solutions.”
Why SMBs now?
TrueState is currently focusing on businesses with 50-100 staff. They are the ones that are most susceptible or turned off by the prices from a top-tier consulting firm. “We are a small team with a compute-based pricing structure. We want to approach smaller businesses with smaller budgets but still have outsized impact.”
“We are focused on solving the high-impact problems, matching the best tool to the high-impact solution. For customers - what is your key issue and, therefore, the key algorithm for you. "Most startups look for a single algorithm for a single use case, which is slow to build, and so TrueState speeds this process up in a more accessible way. We seldom see AI startups working the same way we do.”
How does an AI Founder Staying Sharp in AI Tech
AI is a fast-moving industry, so how does the team at TrueState stay up to date on the latest AI trends and developments? “We have limited time, but we try to stay abreast of ISO standards - a lot is in its infancy. We keep on top of industry trends through reading papers.”
In terms of bespoke builds - “there are new algorithms and a bulk of work. If we built our own foundation models - such as competing with OpenAI, we would be reading a different part of the literature. Our job is to stay up to date on open-source methods.”
Corporate Confidentiality
It goes without saying that privacy is of critical concern, and TrueState’s infrastructure is built in silos where data inherently can't be shared between organisations. “Companies’ own data and models are completely segregated. We are in the last stages of getting compliance certification SOC 2.” (voluntary compliance standard that specifies how organisations should manage customer data including data security, processing integrity, confidentiality and privacy).
How does this differ from other AI companies dealing with privacy for example? “OpenAI uses consumer data, which has its own privacy laws - anyone dealing with B2C models has completely different models and standards. As we enable customers with their own internal applications, we focus on model governance and do so based on international standards.”
What’s next for TrueState?
The funding that TrueState has obtained will help it further build out the tech platform, grow the team and expand to the US and UK later this year.
“Every industry is being reshaped by AI. Winners are defined by their ability to harness the right AI tools to deliver real commercial value before their competitors.”