📆 onboarding #3 begins AUgust 05 2024
Enterprise-Grade
AI Agents
AI Agents
We build private, personalised AI agents to the specifications required by enterprise so that professionals can fully harness the power of AI.
AI Agents Built For High Performing Professionals
According to Harvard Business School, AI improves highly skilled worker efficiency by +25.1%, productivity by +12.2% and work quality by +40%.1
It is clear that those who leverage AI win - and yet AI hasn’t to date achieved ubiquity in enterprise.
We started Superficial to get AI into the hands of every professional so that more people can benefit from the power of AI in their work.
We do this by solving what is holding back widespread enterprise adoption today.
We started Superficial to get AI into the hands of every professional so that more people can benefit from the power of AI in their work.
We do this by solving what is holding back widespread enterprise adoption today.
1 Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality, Harvard Business School, 2023.
Fully Private & Confidential
The number one concern holding back enterprise adoption today is leakage of confidential data.2
Our agents are built atop multiple Large Language Models (LLMs) providing physical separation from foundational models and their training sets. Confidential personal and organisational data is masked and never shared with any third party or used to train any underlying model.
Users can converse privately with their agent, with data deleted immediately and not used for agent training.
All conversational data is deleted within 24 hours, no compromise.
Individually Tailored
Large Language Models (LLMs) are jacks of all trades, but masters of none, with limited domain specific knowledge.
We build and train custom agents for individual users based on their specific roles and responsibilities, leveraging best-in-class, role-specific training content to deliver contextually aware agents.
Users and their organisations can further train their agents on their proprietary data to create agents that are experts in their organisation.
Agents have memory and learn from conversations with their users, becoming more individualised and personally helpful over time.
We leverage multiple LLMs and choose the best for the job at hand, so you're not limited to any single provider.
Hands On Training
AI, like any new technology, succeeds only when users know how and when to use it best.
We deliver on-site training to all new users before they're provided with their agents.
We provide follow up on-site and ongoing virtual training for all users.
All agents come with access to our Prompt Library - a resource with useful, role-specific prompts added by us and our community to help users get the most from their agents.
Staged Deployment
As with any new technology, adoption will stall without the right internal buy-in, engagement and delivery model.
We start small and scale up, enforcing POCs with small internal groups before wider roll outs.
We require user opt-in. Only those that are interested in participating are invited to join.
We deliver top down. Adoption and usage increase and outcomes improve when staff are supported by their leaders.
2 As AI Advances, Trusted Data and Security Concerns Grow – Salesforce Report, Salesforce, 2023.
Our Deployment Process
Request Call
We work with only a small group of individuals and companies at any given time and onboard new professionals in scheduled cohorts.
Our goals for this step are to:
Our goals for this step are to:
Understand your organisation and goals
Agree on a proof of concept structure and go forward strategy
Deliver Proof of Concept
We recommend an initial fixed price, fixed term proof of concept with a small leadership group of between 5-10, before commencing with a wider roll out.
Our goals for this step are to:
Our goals for this step are to:
Uncover any company-specific requirements to tailor our roll out
Get leadership buy-in to support deployment
Deliver real, tangible benefits to users within the POC
Hands On Training
No user gets access to their agent without completing onboarding training.
Our goals with this step are to:
Our goals with this step are to:
Set a base line of understanding of AI and our agents
Embed agents as part of users' daily routines
Set users up to be champions to support wider company adoption.
Phased Roll Out
Users are onboarded in scheduled cohorts, with each cohort trained and supported throughout.
The goals of this step are to:
The goals of this step are to:
Roll out agents to opted-in users and achieve adoption and usage goals.
Provide ongoing training and hands-on support to ensure continued adoption and usage goals are met.
Uncover learnings to continually improve usage and outcomes.
.png)