Welcome to Chat GPT for mining: A conversation with Will Coetzer

Vertically trained AI conceived and trained just for mining and built on a futureproof, multi-lingual platform able to spit out virtually limitless research data points in real-time? That’s the future that could be with us in six months’ time, according to Will Coetzer.

Mining Beacon editor Richard Roberts caught up with the Pulse Intelligence Partner (PIP) founder and director ahead of Resourcing Tomorrow 2024 in London, where PIP will be one of the star attractions in this year’s Tech Pitch Battle.

Richard Roberts: What are you introducing at this Resourcing Tomorrow?

Will Coetzer: We’re launching the MVP [minimum viable product]/Beta version of the Pulse Intelligence platform.

It is a fintech-inspired platform with the user experience at its core, that delivers real-time granular data and analytics specifically tailored to the natural resources sector, with an initial focus on mining, at scale. 

The MVP includes key features such as real-time data aggregation, with all listed companies globally, and highly customisable dashboards. It significantly reduces the time and resources traditionally required for data analysis. Very soon we will unleash our algorithms on private companies.

MVP development focused on research process and data automation. Automated research covers production, capex, discovery, financial, feasibility study etc data and news. Automated, AI-trained document retrieval includes company analysts’ research.

The biggest question we had to answer was and remains about data accuracy at scale. Our AuthentIQ algorithm, probably a boring topic for most, drives everything. We have built and are fine tuning this multi-source, multi-factor algorithm which applies real-time accuracy and probability scoring. 

Richard Roberts: What are you then planning to add to the platform post the MVP stage?

Will Coetzer: Post-MVP, we plan to expand functionality, with extensive vertical AI/ML training, led by mining and tech - human! – subject matter experts on all data, and to enhance the AI interface [through expanded language diversity, etc].

We’re probably six months away from an ML-trained platform launch, hence the beta version is there to see how the data is used, or not used, what people want, etc.

We will also start exploring predictive analytics through our Pulse Radar tool.

Commercially, we’re looking at possible API integrations with data providers and also banks, brokers, funds, etc. We’re also planning to roll out back-office and research automation, and enhanced workflow automation, tools. We already have a paying client, a boutique investment bank, signed up as an early partner.

Our plan is to introduce a subscription and API licensing model as well as bespoke integrations and automation.

Richard Roberts: Why is there a need in the market for this service/product – or is the need emergent?

Will Coetzer: AI will not put businesses out of business. AI will put businesses out of business that do not embrace AI.

To remain at the forefront and build competitive advantage, real-time, actionable data is crucial.

It is no different in mining and natural resources. Traditional workflows often rely on human heavy lifting and static, expensive workload that is theoretically instantly outdated. Industries servicing mining, such as financiers, institutional investors and service providers need dynamic tools to make faster, data-driven decisions. This demand is heightened by digital transformation pressures and resource constraints within these businesses.

It’s the same for mining. Think about even basic inefficiency among corporate development, investor relations and directors having to constantly find competitor data manually. People on hundreds of thousands of dollars salary sifting through websites to find basic data.

Earlier in 2024 PIP uncovered in research that executives and financiers in mining lose between 6-17 weeks per year searching for data. Imagine the efficiency gain if much of this can be automated? What about the improvement in job satisfaction and culture, which is directly related to tenure?

We can automate this data in real time, in perpetuity, and will soon be able to apply Pulse Radar to build your competitor or target base more accurately, automatically.

Richard Roberts: What does the competitive landscape look like from your viewpoint? What makes this stand out, dare I say, like a beacon?

Will Coetzer: While established players like Wood Mackenzie and CRU dominate traditional consulting and static data provision, PIP distinguishes itself through automation, cost-effectiveness, and real-time insights. Our system dramatically reduces – eventually fully eliminating – timeframes for accessing and analysing data while delivering the data at a fraction of the cost.

From our position AI companies in mining are trying to solve the really big problems, such as the future of humanity! They are each focusing on their micro niche. We are an enabler, improving efficiencies exponentially so money and time can be reinvested more productively.

The cost aspect is a game-changer for our target market.

Financiers and service providers no longer need to invest in expensive, static reports or maintain large in-house teams for analysis. PIP’s customisable dashboards and API’s will empower users to extract the data they need and then automate the upkeep, improving ROI by streamlining decision-making and reducing headcount requirements.

Richard Roberts: You founded mining executive search and management advisory firm Stratum International in 2012 When/why did you hatch the concept for PIP?

Will Coetzer: The idea for PIP originated last December whilst deep-diving and educating myself on possible application of AI and automation at Stratum.

I noticed a consistent gap in the availability of timely, actionable data for decision-makers in mining and natural resources – me included! I could not find granular data at scale, even through the super-expensive platforms, not to mention any automation and AI-trained, or contextual, distillation at scale for things I actually wanted to know about. I solved the problem for myself on a smaller scale and realised there is a bigger problem to solve as AI and real-time analytics technology advances, making such a platform feasible.

Being in an international industry and having delivered mandates in more than 60 countries globally through my 20-year career, I also put my investor hat on and visualised a world where anybody has access to the right data in any language.

Think about it. How much more money and talent could mining attract if presented fintech-style, with an Apple-esque user experience, in any language in real time?’

That is a massive driver for me.

Richard Roberts: What is technology enabling you to do now that you couldn’t do when you first thought of PIP?

Will Coetzer: We are fortunate in that we started in this era, so our tech stack is brand new, latest tech, from the ground up.

Definitely the advancements and democratisation in AI and Machine Learning, cloud computing and API integrations now allow us to aggregate, analyse and present data in real-time. You could argue that our incredible algorithm could have theoretically been available if we wrote it earlier, but it would have been a seedling, waiting for AI’s computing power to make it blossom. That time has now arrived.

The crucial thing for us is accuracy at scale. As I’ve learnt the hard way through a few motorcycle accidents on race tracks in my earlier days, speed is nothing without accuracy and control!

Our AuthentIQ multi-source, multi-factor algorithm is core. Once we are satisfied with the results at scale in real-time we’ll start training our models vertically, which then also informs Pulse Radar.

These tools enable us to automate processes that previously required significant human intervention, such as generating investment-grade insights or integrating external datasets dynamically.

We’re also constantly building for what’s coming in 5 years, not just for now. We recognise there will be advancements and tech, both what we know and don’t know about yet, that our tech stack needs to be prepared for.

Richard Roberts: What is Pulse Radar/what does it enable users to do?

Will Coetzer: Predictive analytics. This is a little while away. It will learn and predict organically, or act upon your instruction. Because it’s vertically trained in-house on validated data it does not hallucinate.

Richard Roberts: Are you looking for backers? Who is already on board?

Will Coetzer: We’re open to strategic partnerships with industry players, though we are theoretically still self-funded. We also have a paying client for whom we’re about to commence a larger project, we have – albeit small – recurring revenue and are at the early stages of discussing with several large data providers about research automation.

We’re also close to several other proof-of-concepts [POCs].

We are realising that our major client base initially will likely not be S&P-type subscribers but data companies or financial institutions and investors.

Richard Roberts: What are the connectors, if any, between this venture and Stratum?

Will Coetzer: While there’s no direct operational overlap my 20 years of industry experience, reputation and network-building at Stratum naturally serves as an advantage. As a side note, I do not run Stratum operationally. The company is now led by Nichole McCulloch, who merged her firm Clifton Hill Associates with Stratum in 2023. Most of my focus is now on PIP.

Richard Roberts: How many PIP registrations pre-launch?

Will Coetzer: We’ve secured confirmed beta interest from more than 60 people in leading organisations before we’ve sent a single beta invite. This reflects early confidence in the platform’s potential and ensures a strong feedback loop during the MVP phase.