Why we partnered with Tasklet

Some of the clearest investment convictions come from products we already depend on.

At Oleka, we were among Tasklet’s earliest adopters. What began as experimentation quickly became part of how our firm operates. Within months, our team was using Tasklet agents across investment research, diligence, reporting, data analysis and the flow of information across the firm.

We experienced the product through repeated daily use. We saw what worked, where the product needed to improve and how quickly the team responded to feedback. Most importantly, we saw Tasklet begin to change how our own team worked. That experience sits at the centre of why we invested.

The Product Layer Is Still Under Construction

The capabilities of artificial intelligence models have advanced extraordinarily quickly. Yet enterprise adoption has moved more slowly than the underlying technology.

The bottleneck is increasingly not the intelligence of the models. It is the product layer around them.

Most businesses do not want to build their own agent infrastructure, choose between models, maintain integrations, manage cloud environments and continually redesign workflows as the underlying technology changes. They simply want to describe an objective, connect the relevant information and applications, and allow an agent to complete the work reliably.

Tasklet makes this process unusually simple. Users can create agents by explaining what they want in everyday language. These agents can connect to company software and data sources, conduct research, process information, prepare reports, update systems and manage recurring workflows. Because the agents operate in the cloud, they can continue working without requiring a user to keep a browser window open or supervise every step.

This matters because deploying an agent inside a business should eventually feel less like developing software and more like delegating work to a capable colleague. Tasklet is helping close that gap.

We Were Customers Before We Were Investors

At Oleka, information is distributed across emails, documents, company presentations, financial models, meeting notes and portfolio reporting systems. Much of our work is context-heavy and repetitive, while still requiring judgement. Tasklet became a practical way to connect these fragmented workflows.

We have used agents to conduct market research, analyse potential investments, prepare recurring reports, summarise CRM and meeting activity, monitor portfolio developments and organise information for our team. Agents that began with narrowly defined tasks gradually assumed responsibility for more persistent workflows. Tasklet became part of our everyday operating system.

Using the product so intensively also gave us a close view of the company’s execution. We watched the team release improvements rapidly, turn user requests into product features and continually reduce the amount of effort required to build useful agents. That pace of iteration gave us confidence not only in the product as it exists today, but in the team’s ability to keep rebuilding it as the AI landscape changes.

No Single Model Will Rule the World

We do not believe that one AI model will be best at everything. Different models already perform differently across reasoning, coding, research, creative work, speed and cost. Their relative positions are changing constantly. A model that leads today may be overtaken tomorrow, while a smaller and cheaper model may be more than sufficient for most routine enterprise workloads.

Businesses should not have to reorganise their workflows every time that hierarchy changes. Tasklet can become a neutral layer above the model providers: selecting the right intelligence for each task while giving the customer a consistent product and workflow. This allows enterprises to benefit from competition between model developers instead of becoming locked into a single provider.

It also creates the potential for what we think of as a “Costco model for intelligence”. Most enterprise tasks will not require the most expensive frontier model. They will be completed by the cheapest intelligence capable of producing a sufficiently reliable result. A platform that aggregates demand, routes work efficiently and gives businesses access to multiple models can make AI usage materially easier and more economical.

We have encountered other companies pursuing versions of this aggregation thesis. Many remain developer tools: powerful, but difficult for normal teams to adopt. Tasklet’s distinction is that it combines access to intelligence with an approachable product. The value is not simply procuring model capacity. It is packaging that capacity into agents that people can deploy, manage and collaborate with. That product layer is what can turn declining intelligence costs into widespread enterprise adoption.

Traction That Reflects Real Product Pull

Tasklet’s early growth has been exceptional. During the first four months of 2026, the company grew by more than 1,200%.

At such an early stage, growth rates will inevitably fluctuate. What matters to us is what sits behind the headline: users are not only experimenting with Tasklet but building agents into recurring workflows, expanding their usage and introducing the product to their teams.

The company achieved this momentum before constructing a traditional enterprise sales organisation. Much of the growth came from users discovering the product, finding valuable use cases and increasing their reliance on it. That kind of product-led adoption is particularly meaningful in a category where many enterprises remain stuck between impressive AI demonstrations and limited production deployment.

A Team That Has Built Foundational Products Before

Tasklet is led by a team with an exceptional record of building important technology products.

Co-founder and CEO Andrew Lee previously co-founded Firebase, the application development platform acquired by Google. Firebase succeeded by taking complicated infrastructure and making it dramatically easier for developers to build modern applications.

Co-founder Jonny Dimond was an early member of the Firebase team and later served as a technical lead for Google Cloud Firestore. The wider Tasklet team brings experience from Firebase, Google, Amazon and OpenAI.

That background is especially relevant to the problem Tasklet is solving. The team understands complex infrastructure, but it also understands that infrastructure becomes truly valuable only when it is abstracted into a product people can use. Tasklet applies the same underlying instinct to AI agents: preserve their capability while removing the technical burden of deploying them.

The Road Ahead

The final shape of enterprise AI is not yet settled. Models will improve, prices will fall and the boundary between software and labour will continue to move.

Enterprises will need a product that absorbs these changes on their behalf: connecting models, information, applications and people through a consistent agent layer. We believe Tasklet can become that layer—making powerful intelligence easy to deploy and increasingly embedded in how teams operate.

We began as early adopters, became daily users and watched the product improve alongside our own usage. We are now proud to support Andrew, Jonny and the Tasklet team as investors in the next stage of their journey.