CampaignCampaign(opens in a new tab)
Case Film
Demo Film
Presentation Image




Why is this work relevant for Innovation?

Millions live in unwanted loneliness. Scientific studies show that isolation, with no one to talk to or share memories with, is as dangerous to the health as smoking or alcohol. It can trigger depression, early stage dementia and even lead to suicide.

We were briefed to help Stockholm Exergi change this and to make unwanted loneliness part of the cultural conversation.

Memory Lane is the world’s first reverse engineered voice assistant that can drive meaningful conversations an ask personal questions. It’s an artificial intelligence innovation that will contribute to making society more socially sustainable and inclusive.


Patent pending Memory Lane is the result of 2 years of research on Natural Language Understanding. It can be deployed on any smart speakers, collect users’ stories and save them safely.

We overcome challenges with deep creative technical solutions.

1.We inverted the smart speakers’ native engineered capabilities. It generates personalised decision trees, becoming a proactive biograph which can trigger relevant follow-up questions.

2.We programmed speakers to drive personalised conversations. We also enabled them to save, safely store and produce conversation summaries with the latest advances in Generative Artificial Intelligence. Combining extractive and abstractive Artificial Intelligence models, we’ve integrated human talent for a hybrid-intelligence powered solution.

3.That allows us to structure and record ¬content using the Memory Graph. It encodes conversation history in a graph structure, which is based on classified taxonomy where information is represented in the graph as nodes and the relationships between them as edges.

Describe the idea

250.000 Stockholmers live alone in Stockholm many of who suffer from unwanted loneliness. This has a profound impact on their mental health and wellbeing, and might lead to depression, early stage dementia, and even suicide. We wanted to help Stockholm Exergi change this situation, as they’re known for its social sustainability work “Making Stockholm Warmer, For Everyone”.

After 2 years of research on Natural Language Understanding, we created the world’s first tool and innovation to collect and save stories via smart speakers – The Memory Lane.

We overcome challenges with deep creative technical solutions as we turned Google Home assistants into Memory Lane agents. We reversed their native engineered capabilities — a unique mechanism of a decision tree – enabling Memory Lane to generate instantly personalised decision trees and becoming a proactive conversationalist.

We installed Memory Lane voice assistants in the participants’ homes and it instantly sparked self-perpetuating, personal conversations.

What were the key dates in the development process?

1st September 2017 – Natural Language Processing research and creation of first Knowledge graph structure

1st March 2018 – Research on new Abstractive and Extractive Neural Networks

1st July 2018 – Design of Memory Lane solution

20th October 2018 – Design of the Memory Lane campaign

10th April 2019 – Patent application for the overall Memory Lane technology

Describe the innovation/technology

We created a solution powered by Artificial Intelligence and human interventions to interview a person and have a deep meaningful conversation.

The solution is composed to 3 key technical components:

- Automatic Interviewer: The automatic interviewer works thanks to a feedback loop and a global database of interview questions. The Automatic Interviewer doesn’t simply ask predefined questions but creates on the fly the most appropriate question to drive a meaningful conversation.

- Memory Graph: The Memory Graph encodes information about voice conversations history in a graph structure. This structure is based on classified taxonomy where information is represented in the graph as nodes, and the relationship between them as edges.

- Advanced Summariser: The advanced summariser is a solution powered by Artificial Intelligence and human intervention, which aims at creating stories.

The solution is running on the cloud, in a secured and encrypted private environment.

Describe the expectations/outcome

Raising awareness of Project Memory Lane and putting the dire consequences of unwanted loneliness on the agenda and part of cultural conversations, and thus making Stockholm a warmer city, for everyone.

The Memory Lane technology is scalable at the pace and growth of voice assistants which is a rapidly growing home device across the Nordics. The tech solution is platform agnostic and works across both Google Home and Amazon Alexa.

We anticipated that for a biographer to question all potential participating people, collect their answers, process the information and write a book, would take between 2-3 months to execute manually.

Instead, we created a built-in feature in the Memory Lane solution to scale the questioning aspect, the transcriptions of the answers and to semi-automate the story writing.

More Entries from Product Innovation in Innovation

24 items

Grand Prix

Early Stage Technology



(opens in a new tab)


24 items


Good Health and Well-being



(opens in a new tab)