Shifting Rock

We believe that AI governance is an opportunity, not a burden.

Manage Model Risk

Legislators, regulators, and customers increasingly demand proof that the models behind the products and services they use can be trusted.

Case Study 1

Increase Trust, Adoption, and RoI

The more that customers trust your products, the quicker they adopt. And RoI is increased by reducing the overhead of QA and human review.

Case Study 2

Move Faster

Avoid bureaucratic governance process, or technical constraints that limit the power of your models. Free your teams to focus on performance.

Case Study 3

20 October 2022: Read our submission to the Bank of England consultation exercise on Artificial Intelligence and Machine Learning.
Download Here

What is Agile AI Governance?

AI can deliver great value, but carries a risk of incorrect or biassed decisions hurting customers, data subjects -- or your business. This is why legislators, regulators, and customers increasingly demand proof they can trust the models behind the products and services they use.

Meeting these demands can be a drag on development, as product, engineering, and data science teams implement cumbersome and ineffective post-hoc explanatory layers, limit the models and algorithms they use, spend time on extra documentation, model validation, stress testing, monitoring, and overheads.

But, done right, agile AI governance can accelerate development, enabling squads to focus on delivering value, use the full power of their models, and use ongoing validation to improve performance.

Why Shifting Rock?

A network of experts who can help you deliver safe AI faster.
  • We are product and technology people who have been using AI and agile methods to deliver great products for 20 years. We focus on development velocity and are sceptical of unnecessary process.
  • We have a deep understanding of AI, with PhDs, patents and publications, and have built breakthrough AI products at startups, scaleups, and large tech companies.
  • We have developed governance processes and policies for products reviewed by major regulators, Tier 1 customers, and blue chip investors in the US, UK, and the EU. And Joe Faith's PhD research was the first to show that explanatory transparency is essential to building AI we can trust.

How we help

We help tech companies meet regulatory requirements and use AI governance to drive adoption, by helping them implement agile AI governance:

1. Define We work with your commercial, engineering, product, data science and legal stakeholders to define a policy that fits you: agile, appropriate and meeting all regulatory requirements.
2. Adopt Help the product and engineering teams adopt and own the process, empowering them to make smarter prioritisation and engineering decisions.
3. Audit Provide independent and expert 2LoD and 3LoD audit of your models in production.

Contact

Shifting Rock Ltd is a network of experienced product managers, data scientists and technologists, dedicated to exploiting the potential of AI safely, coordinated by Joe Faith

Joe Faith is an experienced and inclusive tech exec with 10 years product management leadership experience at Google, fintechs, and startups, and 8 years AI research experience with a PhD, original research publications, and patents used in production in Google Cloud.

email joe_at_shiftingrock.com
linkedin.com/in/joefaith/