Microsoft Artificial Intelligence enhancing Customer Experiences
A recent Gartner poll of over 2,500 executives, showed that 38% saw customer experience and retention as the primary purpose of their generative artificial intelligence (AI) investments. Significantly ahead of revenue growth (26%), cost optimization (17%) and business continuity (7%).
With that in mind, let’s look at the history of Microsoft’s commitment to AI and opportunities to infuse AI across the customer lifecycle. For completeness we’ll look at the breadth of AI offerings from Microsoft, not just generative AI (GenAI), as there is still very much a place for “traditional” AI.
A Brief History of Microsoft’s commitment to AI
While AI hit daily mainstream news at the end of 2022 with the public release of OpenAI ChatGPT, Microsoft’s commitment and investment in AI is not new. In an August 2023 interview with Fast Company Satya Nadella (Microsoft CEO) said
“What happened in the last five months, was work of the last 10 years.”
The fruits of this deep commitment are now being realised and released at pace thanks to a broad software, hardware and governance foundation that has been established.
Some key milestones in the Microsoft AI journey include:
- 2016 — Microsoft’s formation of Microsoft AI and Research Group.
- 2017 — Microsoft launched the Aether Committee with researchers, engineers and policy experts to focus on responsible AI and define AI principles which were published in 2018.
- 2018 — Microsoft establishes AI Red Team, a dedicated team thinking like attackers to probe AI systems for issues and failures.
- 2019 — Microsoft and OpenAI announce exclusive partnership.
- 2020 — Microsoft built one of the top five publicly disclosed supercomputers in the world, making new infrastructure available in Azure for OpenAI to train extremely large AI models.
- 2023 — Microsoft announce the broad infusion of AI across every key product and developer platform and commit a further multi-billion dollar investment to the OpenAI partnership. Through partnership with Meta, Microsoft bring the Llama 2 family of LLMs to Azure and Windows.
Enhancing Customer Experiences with Microsoft AI
Every step of the customer experience can be augmented with AI to provide more effective service delivery, more personalised and relevant content and ultimately experiences that drive greater loyalty and advocacy:–
1. Marketing
Marketers are under more pressure than ever to simultaneously run many campaigns, reduce budgets and react to a rapidly changing consumer expectations.
Opportunities:
- Marketing teams can explore, analyse, create and understand customer segments, sizes and preferences using natural language removing the need to deeply understand the customer data structures.
- Content ideas generate unique, relevant copy for marketing emails leveraging content from previous emails or other internet sources providing an opportunity for fresh, personalised ideas and reducing time spent creating copy.
- Create unique image designs, banners and other visual assets through natural language reducing lead times and traditional production costs associated with image creation.
Microsoft Products and Capabilities: Dynamics 365 Marketing and Customer Insights Copilot, Designer and Bing Image Creator
2. Sales
With global economic challenges continuing, many organisations are seeing pressure on both growth and profits.
Opportunities:
- Receive auto-generated opportunity summaries including status and key changes.
- Support sellers replying to an enquiry or creating a proposal leveraging data specifically relevant to the customer, such as pricing, promotions, and deadlines. By auto-suggesting customisable content, sellers can spend less time composing emails and searching for sales data.
- Preparation for customer meetings with account information and recent notes.
- Realtime tips and suggestions during sales calls. e.g. to address buyer objections as well as sales call summaries.
Microsoft Products and Capabilities: Sales Copilot (formerly Viva Sales) for Dynamics 365 or Salesforce
3. Service
Many service based organisations are facing staff shortages or significant staff turnover that lead to poor and/or inconsistent customer experiences.
Opportunities:
- Empower agents to deliver exceptional customer care by drafting contextual answers to queries in both chat and email reducing handling times, reducing agent training and inconsistencies across different agents.
- Provide conversational experiences for customers on their channel of choice. Rapidly deliver conversational experiences across voice, web, SMS, Facebook, WhatsApp, Telegram, Line and email while leveraging existing bodies of knowledge such as FAQ, websites, knowledge bases as well as data from systems.
- Call transcripts, real time translation and call context.
- Customer identification and fraud prevention via voice.
Microsoft Products and Capabilities: Dynamics 365 Customer Service Copilot, Power Virtual Agents — conversation boosters, generative answers and user authored plugins, Teams Premium, Nuance
4. Field Service
Organisations with field workers/engineers have a tremendous opportunity to provide face-to-face personalised service in the customers home or business.
Opportunities:
- Efficiencies for frontline workers through speeding up and improving the accuracy of work order creation. e.g. populating a work order from the context of a customer email).
- Streamline the scheduling of technicians through data powered recommendations based on factors such as travel time, availability and skill set.
Microsoft Products and Capabilities: Dynamics 365 Field Service Copilot
5. Commerce and supply chain
The recent pandemic and other geo-political and climate related events have shown how fragile supply chains can be.
Opportunities:
- Risk Mitigation — predict and act on disruptions that occur across suppliers by leveraging news feeds to proactively flag external issues such as weather, financial, and geo-political events that may impact the supply chain. Identify impacted orders across materials, inventory, carrier or distribution networks to identify alternatives and reduce impacts and delays to customers.
Microsoft Products and Capabilities: Copilot in Microsoft Supply Chain Center
6. Customer facing employee experiences
Of course it is difficult for employees to provide exceptional customer experiences if they themselves have poor tools and data.
Opportunities:
- Similar to conversational customer experiences, employees can also benefit from such experiences e.g. providing contact center agents with more intuitive experiences for exploring organisational knowledge through natural language.
- Content creators, knowledge workers and agents are supported throughout the Microsoft Office suite with tools for generating ideas, copy creation, processing and classifying documents, managing email and analysing spreadsheet data through natural language.
- Support employee wellbeing and productivity.
- More efficient meetings — transcripts, summaries, recommended actions and real time translations in Teams calls.
Microsoft Products and Capabilities: Power Virtual Agents — conversation boosters, generative answers and user authored plugins, Microsoft 365 Copilot, Microsoft Syntex, Viva Copilot, Teams Premium
7. Bespoke Scenarios
While the above provide many exciting opportunities, the real and enduring differentiation for brands can come through creative custom multi-modal scenarios built leveraging AI services across vision, images, speech, text, language, voice, search and decision making — grounded in the organisations own data.
Limited only by imagination, exciting opportunities exist to re-imagine existing experiences and create completely new ones.
Microsoft Products and Capabilities: Azure Open AI (GPT-4, ChatGPT, GPT-3.5, GPT-3, DALL-E, Codex), Azure Cognitive Services, Llama 2 on Azure, Semantic Kernel SDK
8. Makers
Increasingly Makers are an essential community in agile organisations, bringing to life innovations that in the past may never have received budget.
Opportunities:
- Empowering citizen developers to build apps, chatbots and processes using only natural language — lowering the barrier to innovate, improve business efficiencies and address talent shortages.
- Generate user interfaces from a hand drawn sketch or Figma.
- Automate business processes through natural language.
- Ability to build custom AI models that can be used in apps and processes without niche and complex data science skills.
Microsoft Products and Capabilities: Power Apps and Power Automate, AI Builder
Important considerations
While the opportunities to enhance customer experience with AI are exciting and the results can often seem magical, there are important areas to consider for success.
Firstly, and it may sound obvious, but customers are above all, human. Experiences we create, regardless of the technologies used, should treat customers with empathy and be cognisant of the emotional state of the customer in that moment. For instance, the emotional state of a customer calling their insurance company following a car accident is likely to be very different to that of a customer buying a luxury item.
Secondly, old adages still apply. In particular “Garbage in, garbage out”. AI is not magic and relies on the corner stone of data, quality data. For meaningful outputs, input data needs not only to be accurate but fresh, relevant, representative and unbiased. With that in mind, a key pre-requisite for a successful AI project is data analysis and preparation that also considers the genesis of the data to avoid privacy or compliance issues.
Finally, AI is imperfect. Important decisions and communications should continue to have human involvement so basic human values and your brand values are not offloaded to algorithms.
Microsoft’s commitment to responsible AI
Microsoft have a deep commitment to responsible AI and a vision that AI is “built by humans for humans” — hence the theme in the product names of “Copilot”, not Auto Pilot.
In a CNBC interview on 16 May 2023, Satya Nadella said
“AI is already there at scale. Every news feed. Every sort of social media feed, search as we know of it before chat plus search, they are all on AI. If anything they are black boxes. I describe them as the autopilot era.
We are moving from the autopilot era of AI to the copilot era of AI. If anything, I feel yes it is moving fast, but moving fast in the right direction. Moving fast where humans are more in control. First of all, humans are in the loop versus being out of the loop.”
Microsoft Responsible AI principles cover fairness, inclusiveness, reliability & safety, transparency, accountability and data privacy & security. These principles also align with Microsoft’s commitment to harness AI as a vehicle to bring products and services to the one billion people around the world with disabilities.
To support the responsible AI endeavours, Microsoft uses Red Teaming, the approach of pretending to be an enemy and proactively hunt for failures, to probe not only AI security issues but also responsible AI outcomes.
Most common questions and concerns
In recent months, following the wave of GenAI announcements from Microsoft, the most common questions I’ve received are:
Q: Is my corporate data used to train Microsoft’s models?
A: No, Microsoft don’t use customer data to train models. Nor are customer prompts used for training.
Q: Are Copilots 100% factually correct?
A: No, hallucinations are possible due to the underlying nature of generative AI (Note: this problem is universal in GenAI models). A statistical approach, based on millions or billions of data points, is used to determine the next most appropriate word. Consequently, it’s essential humans are kept in the loop to validate outputs. Remember however, the nature of hallucinations, which often look genuine, can make them difficult to spot.
Q: What’s the difference between ChatGPT and Microsoft’s Copilots?
A: Microsoft’s Copilots provide context within the task being performed and ground both the prompt and response in the organisational data. This approach results in more relevant suggestions. More details on grounding can be found here.
Q: What about the risk of infringing the IP of authors and creators who may not have consented to their content being used in the training of foundational models?
A: With Microsoft’s Copilot Copyright Commitment, Microsoft says “if you are challenged on copyright grounds, we (Microsoft) will assume responsibility for the potential legal risks involved”. Of course, legal guidance should be sort and the customer must have used the guardrails and content filters built into the products.
What next?
The features released to date (described above) are a fraction of the Microsoft product roadmap. We will continue to see deep infusion of AI right across the Microsoft ecosystem in future product waves to further enhance both customer and employee experiences.
Microsoft’s AI research continues at pace across areas such as multi-modal large language models (MMLLM), the hardware to deliver and scale such immense models, AI noses for sensory experiences and situational intelligence to understand and act on the human world around us.
To find out more and understand how to start small and scale AI to enhance your customer and employee experiences, please reach out to me on LinkedIn.
