Autonomous agents are the next phase of enterprise AI, claims Fetch.AI

Intelligent assets can seek out value for themselves

The usefulness of AI as currently deployed in the enterprise lies in automating repetitive manual processes to save money and make them more efficient. Current examples include chatbots, which handle the more straightforward queries flowing into helpdesks and spam filters and fraud detection - tasks that machines can do much more quickly and efficiently than their human equivalents, learning from their mistakes as they go.

But what if AI could take on the more complex tasks of negotiation and unlock unutilised value in assets and data? This is where we are going next, according to Cambridge-based startup Fetch.AI.

Fetch.AI represents real-world entities - be they a hotel room, a seat on a plane or a physical or monetary asset as Autonomous Economic Agents (AEAs) which "represent themselves" on a decentralised platform called the Open Economic Framework (OEF) in which they go about seeking the best deal for themselves according to preconfigured conditions, independent of any human intervention.

This virtual world of interacting agents is coordinated via a "smart ledger" which acts as a collective intelligence for all the AEAs as well as supporting transactions between them.

The Fetch.AI platform is still in beta with an expected release date next year, but the firm recently forged a partnership Clustermarket, a booking platform for scientific equipment.

Lab equipment is often highly specialised and very expensive, with only a small number of each instrument produced. Clustermarket is an online marketplace that allows research institutions and businesses to offer their equipment and services for hire to academics who can book them via the platform.

Together with Fetch.AI, Clustermarket has represented 600 pieces of equipment and schedules as AEAs operating on the Fetch platform, which plugs into its marketplace. The autonomous agents provide personalised recommendations of specific equipment and services to researchers based on previous activities, while on the other side of the equation they help providers of equipment and services to upload them to the platform in a way most likely to achieve success.

Over time, the agents should automatically ensure the optimum availability of the instruments and arrive at a fair pricing structure, dynamically optimising the allocation of the resources to projects that require them based on the providers' preferences.

Johannes Solzbach, CEO of Clustermarket, said: "Users will find it much easier to locate the instruments they need while providers should be able to handle more bookings.

"Scientific equipment is a limited resource and deploying AI that maximises its use provides incremental gains for everyone." he added.

By coupling an autonomous agent with an asset and allowing it to represent itself, hidden value can be realised said Fetch.AI CEO, Humayun Sheikh. "Agents on our network can sell services, handle bookings and solve a range of problems autonomously."

While the first wave of enterprise AI has been mostly concerned with cost-savings, the next will be much more about adding value he believes, with data currently sitting in enterprise data warehouses given a new lease of life by being able to seek out new uses for itself.

Sheikh claims that his company′s platform will enable organisations to share datasets, models and algorithms securely to derive new insights without the need for third-party analytics platforms.

"Asset-owning organisations are already aware that middleman platforms have captured significant value. Therefore we expect those organisations to be very interested in the alternative approach we offer," he said.

Other companies working in this area of decentralised AI include UK-based Cortex, Matrix AI from Hong Kong and SingularityNET which is headquartered in the Netherlands.

Artificial Intelligence and Machine Learning Live 2018 - Monday 19 November 2018

What can artificial intelligence and machine learning do for you and your organisation? If you don't know yet, or want to make sure that you're not missing out, Computing's first AI & Machine Learning Live event is for you. To find out more, check out the Computing AI & Machine Learning Live website. Attendance is FREE to qualifying IT leaders and senior IT pros, but places are going fast