We’ve reached a stage where we can start to assess the capability of leading service providers to deliver comprehensive services across key AI platforms, especially Microsoft’s Azure AI platform and Google’s emerging AI platform suite. So without further ado, let’s ask HFS’ Research Vice President, Reetika Fleming, how she fared leading the two major Top 10 efforts this year…
Reetika – how are services around AI platforms progressing? And specifically, what have you learned with regards to Google and Microsoft platforms?
We’re continuing to see AI ecosystems evolve around the big cloud vendors – Microsoft, IBM, AWS, and Google. From our recent deep-dives into the AI services alliances developing around Microsoft and Google, I can tell you that there are different strategies at play here. Google and Microsoft themselves have their own strengths and priorities, and the SI and consulting alliance partners are collaborating with them in different ways.
- Google’s portfolio of AI components, such as text-to-speech and computer vision, is a great starting point for a fundamental development layer. Google’s AI R&D leadership is well respected among clients and service providers alike. What has been missing are combined applications of these technologies to solve specific business challenges for major business functions and industry verticals. This is where service providers have a critical role to play, and they are filling the gaps by building solutions either in collaboration with Google developers or with clients in selected industries that are ready for AI.
- Microsoft is emerging as the most ‘enterprise friendly’ AI ecosystem. As enterprise clients grow more comfortable with AI initiatives using the Azure technology stack, the services market is quickly developing around client demand. We expect this market to pick up significantly in the coming year as AI services and technology as a whole see greater adoption and as Microsoft and its services partners make more concerted efforts to bring more relevant and timely AI solutions to large enterprises.
Large service providers, including IT services firms, boutiques, and consulting houses, have established or expanded their ecosystem alliances to work with MS and Google on AI. Joint go-to-market activities are taking the form of:
- capability development (POC and pilot funding, talent development);
- market awareness creation and sales planning (joint account planning, campaign work such as Microsoft’s “Make AI Real” workshop series); and
- technical collaboration (joint research, IP creation).
What is driving firms to invest in AI – is it a real desire to meet newly designed outcomes, or more a compelling need to keep on top of emerging tech?
Most of the clients we’ve spoken to in the last year have gone through the learning curve of viewing AI simply as the shiniest new toy the need to be seen to have a strategy around. This is finally starting to become about plugging real business problems and tapping into new opportunities using the evolving range of AI technologies.
Here’s an anecdotal indicator of how things are skewing towards business – at least half the number of AI leaders and sponsors we’re speaking to are business stakeholders, whereas this was squarely an IT/Digital/CoE skewing peer group in years past. Enterprises in our research are certainly looking at business outcomes from their AI investments, including driving up customer experience with AI-enabled apps with virtual assistant support, improving the quality of anomaly detection in manufacturing equipment, and reducing turnaround time on invoice processing. This is good validation for our thinking earlier in the year that AI needs to be driven by the business, with IT as a key partner.
What type of services are you seeing drive the AI industry right now? Is it more service providers delivering “support” work for clients who’ve already figured out what they need, or are you seeing real “co-innovation partnerships” where provider and enterprise work together to design new process flows to achieve pre-defined business outcomes?
The last few years has seen many services firms go from completely opportunistic AI exploration to the formal development of AI practices. This is no small feat considering:
- the technologies are still evolving, at a point where new academic papers are leading to breakthroughs all the time
- the talent is “thin on the ground” for both technical skillsets in data science, applied ML engineering, and distributed computing, and non-technical understanding of the application of AI into business
- the range of capabilities needed to make enterprise AI a reality require massive amounts of collaboration within a service provider’s organization (and their clients) going from data, analytics, cloud, infra support, business domain expertise, consulting, design thinking, product development…
Pure support work is still a norm today, as many clients will test the waters with service providers at the execution level on a project or two. But doing pilots and POCs on repeat can only take a service provider so far. They have learned over time that they need to bring a multi-disciplined team together with industry-specific solutions to actually “collaborate” with their clients.
A few market-leading service providers have certainly developed these types of co-innovation partnerships with their strategic clients. They jointly ideate and vet AI opportunities, and are able to connect across business and IT stakeholders within these firms because of their reach. Here’s how you know these engagements are really partnership-driven – the service provider will be as invested as the client organization in helping the client develop their own AI capabilities, whether that’s through training talent, setting up CoEs, advising on governance and control, or investing to solve unique client problems.
How are you seeing AI impact enterprise “experiences” in terms of customers and employees? How do you see this advancing as AI evolves?
We’re seeing tremendous interest in using AI to drive better experiences, particularly to improve customer relationships. Phil, you talk about the hyperconnected future state where enterprises need to not just respond to but anticipate customer needs. AI technologies are perhaps the biggest catalysts for hyperconnectivity, because of their ability to “hyper-personalize” customer experiences.
I love the concept of AI ultimately becoming invisible or just natively being built into the process. You don’t know you’re using it, don’t need specialized skills or training, you just get the benefits, whether you’re a customer, partner, or employee. The best experience in these terms is either delight (e.g. this company knows exactly what I want) or effortless engagement (e.g. it doesn’t take me what feels like a million years to serve this customer!) We’re going to start to see new standards emerge for major enterprise platforms and systems in the next few years for AI-driven user experiences based on this concept. It’s no coincidence that the SAPs and Salesforces’ of the world are pouring millions into AI.
Casting your eye ahead 2-3 years, who do you see winning in the services space – will it be one of these early leaders, or can you see new players emerging with a different approach?
As we see the further formalization of the AI services market, we’ll need to watch for:
- Who can find the most successful talent models for AI? Whether that’s crowdsourcing a la Wipro-Topcoder, EY’s “Badges” program to recognize employees’ new skills, or TCS’ investment with Cornell Tech through their new innovation hub in NYC… there’s different strategies on AI talent development for the future, and not all will pay off.
- Who is able to develop and successfully sell digital change management to clients – we see this all the time right? Change management is set to the side because clients believe they can do it all internally, but change management for AI is fundamentally different than other initiatives – you have to alter job roles, the workings of entire processes and decision-making points, establish and continually monitor governance and transparency of new models, and so on. Not everyone can firstly sell digital change management along with AI implementations, and then deliver successfully, and it can be what makes or breaks AI engagements.
- Who is able to make AI easy to develop and scale – internally and for clients. Centralizing and creating libraries of reusable assets, investing in “autoML” type of capabilities that can compress the data prep and training time, containerization of capabilities and existing platforms…these are all indicators of prioritizing scalability for AI.
- Who is able to bring an integrated approach to automation technologies like AI? It’s an easy tell when a service provider’s RPA team has no idea what their AI practice is doing. As we always say, clients want to buy outcomes, so the more service providers can bring a holistic set of capabilities to the table, the more their AI pitch will actually land.
- Who is able to partner with technology vendors most effectively? This includes joint account planning, joint go-to-market and product engineering AI specialists like kore.ai and of course the cloud vendors.
I see the market leaders in these early days pulling ahead, but there will also be a few new logos on the board in the next 2-3 years because of these factors. The service providers in the middle might be left doing some of that support work you referenced earlier.
Lastly, we’re in the final phase of analysis for our comprehensive Enterprise AI Services Top 10 report, so check back with me in a few weeks for more on this.