Identifing the right use cases
AI Use cases should be identified based on detailed quantitative analysis of data pertaining to the business problem with an objective assessment of expected benefits to the customer. From a solution provider’s standpoint, it should address a common problem for an industry/ across industry verticals to ensure high re-usability and ROI.
In the future state, the following changes are predicted:
- Use cases for solution building with greater collaborative co-development involving industry domain experts/analysts, AI Solution Engineering and the customer.
- Good market intelligence to assess applicability of the solution across industries, competition landscape and defining the USP.
- Rapid fail fast prototyping approach to accelerate conceptualization of the solution.
Key Success Measures: Cross industry applicability matrix, Reusability, ROI, client topline impact and NPS.
Solution building blocks
Time to market is of essence for an impactful solution. The key decision point is to choose between in-house development using available software components/ APIs from the engineering arm or to build the solution components grounds up. The key challenges are time commitments and talent availability.
In future, leveraging a larger ecosystem will become crucial. Hence:
- Right choice of technology partners and boutique vendors will be a key CSF to complementing each other’s strengths.
- This approach will address the challenge of building the best in the shortest possible time.
Key Success Measures: “Leverage to Build” ratio, TTM, TTV, TATs for processes augmented by AI solutions.
Time and Talent
To an AI solution delivery leader, time and talent availability is critical for successful production rollouts. Availability of experienced AI/ML specialist in the market with Deep Learning, NLP skills is limited which results in schedule overruns and quality issues.
In future, the talent demand-supply landscape will change drastically when the millennial workforce joins in larger numbers to skew tomorrow’s skill-pool. This will be facilitated by:
- Federated availability of niche skilled talent pool.
- Crowdsourcing platforms- global pool of talent with deep specialization in specific topics.
- What will be tricky is how atomically the solution components and specifications could be defined and detailed and how seamlessly they get integrated.
- Microservices-based design paradigm will enable integration and interoperation of components built by crowdsourced workforce.
Key Success Measures: Number of function points delivered through crowdsourcing platform/ total function points.
AI solution development is experimental with algorithms, ML models incrementally refined over time. Requirements are sketchy during design phase, making the build cycles iterative. This iterative, incremental mode of execution makes Agile process-model ideal for AI projects.
- Engaging a crowdsourced global team would mandate distributed Agile process-model.
- Delivery governance processes should address driving commitment on timeline, quality and accountability from a crowdsourced workforce.
- Component integration and DevOps will be crucial to successfully build and deploy the end to end solution.
- Governance will no longer focus on operational cost control or risk management only, but will shift emphasis on topline indicators and business outcomes.
Key Success Measures: Code Quality Metrics, topline measures, customer mindshare and market share.
With wider acceptance of Cloud, deployments of AI solutions and services would shift to Cloud platforms.
In that future scenario:
- Portability across Cloud platforms and provisioning using SaaS model will become the preferred deployment modes.
- Traditional Fixed Price/ Managed Services commercial models will give way to outcome-based pricing models, gain share models, usage-based pricing etc.
Key Success Measures: Higher Business Value realization, ROI.
In a world where “customer defines everything”, the end consumers would best endorse a solution.
In that spirit:
- Industry forums, social media, customer’s professional network will influence perceptions around any software solution.
- Engaging Analysts early on to vet a solution will be the catalyst to generate positive vibes in the market.
- Customer’s higher stake in the solutions ideation and design will have ‘owner’s pride in the solutions that are co-created with their practical domain inputs.
Key Success Measures: Customer experience scores, endorsements from Analysts, Industry experts.
Over the next five years and beyond, the role of an AI Solution Delivery Leader will evolve as one who enables the creation of production grade AI-based automation solutions of high business value by bringing together the collective wisdom of experts in a connected world.