How AI Assessment Works

AI is automatic. Of course it is, all forms of Artificial Intelligence (AI) from predictive, reactive to generative are essentially automated automatic accelerators that organizations can use to bring increased capabilities to business functions and workflows across every industry vertical. But just as a vehicle with automatic transmission (for now at least) still requires us to be there, automating workplace functions with AI requires some steering and direction.

While we may arguably get to a point where AI services are switched on almost without any software engineering orchestration or business stakeholder direction, right now we need to be able to assess how AI-powered productivity tools are being applied. This is the world of AI assessment workshops, AI advisory clinics and so-called Proof-of-Concept (PoC) labs and private partner preview programs.

Some of these protocols and procedures will be done in vendor-specific circles within the perimeters that one or more technology vendors might work within. Others will involve taking commercial enterprises and public body departments through an audit and analysis process designed to pinpoint where AI and Machine Learning (ML) can be most productively, profitably and safely applied. So how do these systems work?

AI ground zero: use cases

“In any conversation about AI, it begins with the use case. How is the business using AI across its operations, what is working, what is not? Often this will focus on the skills that a business has available internally to make the most of AI. For lasting success, employees need to understand AI’s value - and crucially - be ready to work with it in a workplace where some workflows are automated... and some are not,” said Josh Mesout, chief innovation officer at Civo.

With this information in hand, Mesout suggests that organizations can then work with their chosen IT vendors to to shape a solution that matches its needs, assessed against a range of factors including budgets and technology. After this, he points the focus towards implementation and starting to solve the customer’s business problem, working closely with them as they deploy the technology... and being ready to step in should any issues arise.

“Organizations of all sizes are increasingly after AI and ML solutions that are simple to leverage, but offer powerful capabilities. Spending hours configuring and supporting infrastructure for minutes’ worth of insights is not feasible for many - and neither is spending months on a training programme to only use one ML service,” said Civo’s Mesout.

He further states that the time involved and complexity of these services means more organizations are shifting to an open source approach that enables companies of all maturities across the AI implementation cycle. Mesout underlines the proposition that by using pre-configured services for infrastructure and tapping into vast swathes of expertize, delivery timeframes can be drastically shortened whilst simultaneously building internal expertise, ultimately making the technology accessible to all.

Civo is a cloud-native service provider, with a growing set of AI and ML services for its customer base. The company focuses on building what it promises are 'affordable' and easy-to-use ML solutions and making sure they closely match the needs of its users. In a world where many IT teams are not using ML as efficiently as possible, with implementations going off track, Civo research suggests that some 48% developers are finding ML projects too time-consuming at present.

Tanium in the cranium

A company that also knows how these procedures work is Converged Endpoint Management (XEM) company Tanium this year, which announced its participation in the Microsoft Security Copilot Partner Private Preview.

“In the context of security, AI’s impact is likely to be profound, tilting the scales in favor of defenders and empowering organisations to defend at machine speed. At Microsoft, we are privileged to have a leading role in advancing AI innovation and we are so grateful to our ecosystem of partners, whose mission-driven work is critical to helping customers secure their organizations and confidently bring the many benefits of AI into their environments.” said Vasu Jakkal, CVP, Microsoft Security.

Tanium is working with Microsoft product teams to help shape Security Copilot product development in several ways, including validation and refinement of new and upcoming scenarios, providing feedback on product development and operations to be incorporated into future product releases as well as validation and feedback of Application Programming Interfaces (APIs) to assist with Security Copilot extensibility.

“Generative AI and autonomous endpoint management (AEM) has potential across all facets of the security industry, including to help organizations reduce risk while reducing cost and serving as a force-multiplier for security teams to outpace adversaries,” said Matt Quinn, chief technology officer, Tanium. “We believe the powerful AI at the heart of Microsoft Security Copilot, combined with Tanium’s real-time visibility and control across the entire IT estate, can deliver tremendous value to our joint customers.”

The reality we see playing out in many enterprise software platform companies is one where AI is becoming embedded in every application and data service throughout the IT stack. Known for its work across Enterprise Resource Planning (ERP) and wider asset and inventory management technologies, IFS says it is continually working to identify, test, measure and scale the areas it can automate with data and AI to save resources.

In terms of how it approaches AI assessment from base levels, Matthias Heiden, IFS CFO and executive sponsor of the company’s AI transformation talks about a structured approach based around four types of AI initiatives: product-embedded AI; product-related AI; business-enabled AI and casual use AI.

Feedback loops & synergies

“To reflect the theme of this discussion and explain how we ‘assess’ AI models and services before any internal or customer-facing implementation, we first ensure that governance models have been implemented to allow teams to pursue both product and internal initiatives in tandem,” said Heiden. “We surround this work by creating a feedback loop that opens up a channel to cross-functional synergies within the company. This structure enables internal use cases with R&D services, offering guidance and setting data guidelines, operating in parallel with the testing of new product features, offering new ideas and using R&D services for implementing customer use cases.”

Heiden insists that his firm’s strategy has always been to start with suitable industry use cases and then weave in AI to provide value that can be increased through automation. Deployed carefully and thoughtfully in this way, he says that organizations can use AI to do everything from enabling better demand forecasting, optimizing manufacturing schedules and even task AI with striking the optimum balance between team member skills.

A productivity tool & orchestration engine

Looking again at Microsoft with its Copilot for Microsoft 365, this is a technology that combines the power of large language models (LLMs) with an organization's own data in the Microsoft Graph and the Microsoft 365 apps to turn your human ‘words’ into what is being called a productivity tool & orchestration engine. Using its Business Chat service introduced in March of this year, users might use natural language prompts to “inform the whole team about this week’s product changes” across calendar, emails, chats, documents, meetings and contacts.

But as already intimated, implementing this level of automation needs a touch of care. Microsoft is now working with NTT Data to develop new offerings for Copilot for Microsoft 365. NTT Data promises it will help businesses automate tasks and create content bolstering productivity and reducing costs. The newly created offerings signify the evolution of NTT DATA’s relationship with Microsoft, delivering services and support across the Microsoft portfolio to enterprises globally.

Advisory assessment workshops

What we can see here is a 3-week advisory workshop designed to help enterprises understand the potential of generative AI in the digital workplace, evaluate their preparedness and maturity state for Copilot for Microsoft 365 and recommend how they can unlock employee productivity with integration into existing Microsoft productivity suite management. Deliverables include an organization readiness report, use cases and personas, pilot planning, roadmap and recommendations.

According to Marv Mouchawar, executive VP for global innovation headquarters at NTT Data Group Corporation, This work also includes a Cloud Voice Readiness Assessment Workshop. “[This is] an assessment of processes to gain a thorough understanding of how Copilot for Microsoft 365’s generative AI capabilities can be used with calling and meeting solutions to improve business outcomes. This includes demonstrations of use cases and scenarios, and customized, actionable recommendations,” notes Mouchawar and team.

NTT Data is a Microsoft Global System Integrator Partner. The new integration builds on NTT Data’s continued relationship with Microsoft as a solution partner across zones including infrastructure, digital workplace, security, data, AI and digital application innovation.

“We’re proud to be strengthening our relationship with Microsoft and to be one of the first partners to launch offerings for Copilot for Microsoft 365. This is a powerful AI tool that will enable us to assist clients in realizing the full potential of generative AI technologies. This collaboration is part of our continued commitment to providing innovative AI solutions to clients through their entire journey, from advisory to managing their Microsoft estate, including Copilot. It means helping businesses enhance the power of next-generation AI in an entirely new way of working and unlocking a new wave of productivity growth,” said Mouchawar.

Automating automatic AI

As suggested above, we may (most would argue ‘will’ rather than ‘may’) get to a point where we can more automatically apply AI to the coalface of our enterprise application use cases, but the nuances of these procedural deployment checks and integration tasks is a lot of what’s happening right now.

In the near future, AI will be more autonomous, it’ll be automatic.