When & where is AISWITCH most applicable?
When and how does AISWITCH work best?
Q1. When is the AISWITCH™ framework most relevant- in contexts of development or post deployment, of enterprise AI solutions?
90% organizations have already done some successful POCs on AI and intelligent, integrated automation. But most of them are facing challenges in scaling them up in production, without uniform runtime governance and management best practices and processes frameworks. AISWITCH offers a lean and comprehensive framework, starting with organizational AI strategy & financial management, AI workforce & CoE management, information/ data and technology management and culture change levers and ultimately the outcomes metrics.
10% of organizations that have moved some of their major AI initiatives into production are also still facing major challenges in terms of scaling up usage and RoI, value & outcomes measurements, culture change, information security and quality issues, workforce upskilling & adoption issues. For these organizations, AISWITCH helps them assess current-state realities across people-process-technology and business dimensions, and help them drive specific initiatives to improve the enterprise AI effectiveness scenario, in critical areas.
Net net, AISWITCH offers a 360 framework for AI-automation management, across tools/techstacks and beyond just tech parameters in development to deployment, integrating technologies and data dimensions to other critical aspects of AI-automation, such as strategy, workforce, culture and most importantly- business value metrics. That way, it's relevant across all the lifecycle phases of enterprise AI-automation solutions.
===========================================================================
Q2. If an organization already has platform-specific governance tools and orchestrators, what additional value will AISWITCH enable for them?
The analogy between ITIL and AISWITCH is most relevant in the context of this question.
Before ITIL evolved, every tech-stack across infrastructure and apps management layers had vendor-specific ITSM engines and orchestrators. There were 2 main challenges with that fragmented IT operations approach:
1- There were no uniform processes e.g. to manage IT strategies, financials, service requests, SLAs & OLAs, incidents, changes, capacity, releases, configurations etc. Each vendor was performing each process in their own way. As a result, predictability and controllability were non-existent, and many enterprises had to take a 'Total IT' approach i.e. everything from one/ two vendors, thereby having lock-in's and no flexibility in negotiations, basically falling prey to TSP oligopolies rather than taking best-of-breed advantages from newer, more agile and innovative tech vendors.
2- Consequently, there were no integrated information and control flows between processes and tech-stacks. The entire IT operations landscape across different infra and apps stacks from different vendors, resembled fragmented tech islands that had no common bridges and handshake points.
Current state of AI-automation implementations in organizations resemble pretty much similar situations. Different RPA, IA, AI vendors have created their own islands within enterprises, with their own administrative stacks. It’s next to impossible to ensure and drive uniform process definitions, data definitions, integrated e2e intelligence, controls and visibility, across these islands.
AISWITCH is the common process framework that builds the bridges across the tech-individualistic AI-automation islands, ensuring that all these adjacent tool-stacks function in an integrated, dependency-aware manner.
Detailed FAQs on AISWITCH applications
AISWITCH and pre-existing AI CoEs, DevOps and Agile Practices
Q3. How does AISWITCH™ help when an organization already has an AI COE?
Organizations with an existing AI CoE are obviously better positioned to leverage AISWITCH as they don’t have to start from ground zero. Their AISWITCH journeys will include:·
- Evaluating their progress and AI journey mapping, with self-assessment tools and AISWITCH scorecards, with peer-benchmarking support for best practice discoveries- within domains or cross-pollinated across domains, as required
- enabling the CoE teams to adopt external and internal best practices, in terms of capability and capacity planning, tech/ services partner selections and negotiations, costing, pricing and RoI models, internal consulting on AI-automation opportunities with other SLs and BUs
- Creating, navigating, communicating and updating Technology /Business AI Adoption & Scaling Roadmaps· - increasing top management visibility of strongly differentiated AI-automation capabilities
- In course correction – course conformation – course adoption of AI-automation initiatives
============================================================================
Q4. Who should be looking towards AISWITCH™? The technology folks or the business/corporate leaders?
Both sides need to play the enterprise AI-automation game as ONE TEAM and not against each other. AISWITCH will enable this collaboration by ·
- Translating IT performance and technology attributes to business languages and metrics·
- Defining Corporate guidelines, policies, strategic objectives, and ethical, responsible practices around the AI adoption.·
- Focussing on business-IT combined AI Value realization through shared, collaborative metrics
============================================================================
Q5. I already have an AI team. How and why should I go with AISWITCH™?
Simply for 5 reasons:·
- For tangible, prescriptive, universal, peer-benchmarked process guidelines and best practices·
- Tried and tested approaches to attain effectiveness of AI-automation strategy execution, policy & usage, adoption, risk management and governance processes·
- Focus on the details- nuts and bolts of integrated AI-automation solutions in production·
- Apply the lessons learned from real-time hands-on, grounded experiences·
- Manage high capex and opex of AI-automation solutions and tech stacks in production, and effectively plan, negotiate and save vendor expenditure and additional investments
============================================================================
Q6. Will AISWITCH help us frame our AI policies and strategies with example templates?
Yes, these are available as part of the AISWITCH AI-automation Strategy Management Processes.
============================================================================
Q7. How will AISWITCH help us in scaling up adoption and organizational change management and culture change?
The AISWITCH AI Culture Management Processes cover 5 key elements of organizational change levers and culture shift gears, in terms of
1- measurable awareness building,
2- measurable effective communication,
3- organizational structural changes- new roles-skills maps, training & capability maps, RACI matrix, AI-outcomes focussed KRA-KPIs
4- behavioural metrics along with performance metrics - for AI-automation builders, operators and users within the organization and partner ecosystems
5- measurement of effectiveness of the change and culture shift programs.
Examples, templates and best practices are available under each process covering combinations of these 5 key dimensions of organizational change and culture shifts.
============================================================================
Q8. What is the connection between AISWITCH and agile methodologies?
Agile, DevOps and automated, high-speed release management are de facto practices in AI solutions development lifecycle phases. They are equally relevant post deployment too, e.g. for
- handling data pipelines, incremental training,
- model changes, parametric and algorithmic changes
- version controls
- management of utility-warranty curves of AI assets in terms of performance/ resource-efficiency/ accuracy
- across different phases through their runtime lifecycles.
There are already numerous agile and DevOps based tools, technologies, approaches and platforms available to manage AI solution development phases (Kubernetes, Jenkins). But there aren’t any uniform/ consistent, integrated, universal AI-automation run-time management and governance processes available.
The 30 processes of AISWITCH, across the 6 key dimensions of AI strategy, workforce, information, technology, culture and human-AI augmentation management offer such a framework. These processes are:
- Uniform and consistent (e.g. same processes apply for all AI solutions regardless of their development tech-stacks or tech platforms), ·
- Unified and integrated (e.g. processes that assimilate best practices from all major and evolving standards applicable for AI-automation management e.g. the IEEE standards and ISO/ IEC standards), ·
- Universally applicable (processes are relevant across runtime lifecycles of AI solutions, irrespective of domains/ verticals).
In all of these 30 processes, speed is a major attribute in all the process metrics definitions, e.g.
- AI Strategic agility metrics: Speed of change of AI-automation strategic levers and execution
- AI Workforce agility metrics: Speed of retraining & upskilling % of workforce, speed of re-value-generation by % workforce with now-replaced/ obsolete skills (like coding!)
- AI Information agility metrics: Speed to accuracy in terms of new/ changed datasets/ parameters/ attributes/ feature-sets, and subsequent model retraining- additive/ incrementation ML; speed of adherence to new data security & privacy regulations e.g. changes in GDPR/ HIPAA/ PCI-DSS/ SOX and so on...
- AI Technology agility metrics: Speed of deployment of new algorithms, new functions available, new libraries, open-source/ proprietary tech-stacks from partners; speed of productionizing new AI-optimized green infra sandboxes and runtime environments; speed of upgrading to infra-secured/ hard-gated AI; speed of provisioning; speed, elasticity & resilience of AI infra capacity management- cloud/ on-prem/ hybrid; speed of model change execution on infra/ base tech stacks
- AI Culture agility metrics: Speed of culture change-correlated to effectiveness of communication and training programs, R&R programs, HR & OB initiatives; speed of behavioural change levers e.g. on proactiveness, collaboration, learnability, openness, fairness, reverse mentoring, knowledge contributions, training & e-learning consumptions outside compulsory role-based programs, third party certifications uptake
- AI Human Augmentation (Business Impact) metrics: Speed to Value metrics, TTM, TTV, TimeToChange, TimeToNewProducts, TimeToDevNewServices, production release cycle frequencies, time-to-targeted-accuracy-in-production, etc.
All these metrics are agility-focussed. Hence, agile techniques are the most effective approaches towards running all the AISWITCH processes.
============================================================================
Q9. Will AISWITCH make an organization future-ready in terms of AI policies and evolving standards of ethical enterprise AI governance, regulations and compliance audits?
Yes, it will improve preparedness and maturity of an organization in terms of ethical and risk-managed AI usage, as the framework predictively factors in key considerations of evolving AI-IA standards and major government policies across the world.
============================================================================
Q10. Does the AI CoE of an organization have to implement all the 30 processes across 6 dimensions in AISWITCH?
NOOOOO!!!! This is another key difference between AISWITCH and ITIL that, although prescriptive, AISWITCH processes have interrelations not interdependencies, with some key exceptions (e.g. Data & Tech governance mechanisms & processes MUST be there, for the assets and lifecycle management processes to work). Otherwise, how the processes will be prioritized and adopted will entirely depend on the current realities, priorities and challenges of the AI-adopting organization.
