In the tech and telecom industry, the move to adopt AI has turned into an all-out race. But that doesn’t mean the course is clear.
Many midsize tech companies have implemented artificial intelligence (AI) at some level, but they aren’t sure about their next steps. Now, a surge of new capabilities with enterprise-wide applications has flipped the question from:
“how to plug AI into your tech company”
“how to plug your tech company into AI”
There’s a growing risk of missed opportunities, and leaders are looking for the best points of access, approaches and long-term goals.
It’s about opportunities
To sort out all of the opportunities and find the right path to success, companies need a structured approach. “If you look at where we started this year, to where we are now — and how it's iterated — capabilities are constantly changing from week to week,” said Grant Thornton Technology Industry National Managing Partner Andrea Schulz. “It's hard to keep track of everything, so it's important to have a structure in place.”
“I could see a tech company boiling the ocean with the AI solutions they’re pursuing, and not necessarily making sure that the solutions are aligned with their strategy,” Schulz said. “They could spend a lot of money and time pursuing AI and not getting significant ROI from it, because right now everyone's in a rush to show that they're the earlier adopters.”
“Companies need to make sure that they have a meaningful alignment with their strategy on their AI adoption path,” Schulz said.
Futurist Jim Carroll on AI in technology
4:27 | Transcript
It’s about strategy
It’s easy to get swept up in the rush to adopt enterprise-wide AI capabilities, but every company needs to follow a different path to success. Your path should align with a strategy, using your unique needs, capabilities and controls as guideposts along the way. So, where do you start?
“It depends on the problem you’re trying to solve,” Schulz said. If you can’t afford to fund exploratory initiatives, you need to start by looking for ROI. Your AI strategy will ultimately need to be more than a single solution, but Schulz highlighted common starting points where tech companies have implemented AI capabilities with clear returns:
- Customer service: Chatbots are not new. However, AI-driven large language models can help customer chatbots expand past a static menu of responses to better engage customers, solve more of their problems and collect more information along the way.
- Customer relationship management: CRM systems are using AI to take customer engagement a step farther, optimizing sales and interactions overall. This year, CRM solution leaders have announced that they are implementing GenAI capabilities driven by ChatGPT.
- Coding and product development: AI-driven code completion solutions can write or suggest code based on your input parameters, specs and existing code. While these solutions cannot entirely automate development, they can significantly accelerate your time to market for a competitive advantage. AI can also play an important role in the market analysis that should inform your product development cycle.
- Business operations: As tech industry growth has slowed, the pressure to show profit is pushing more companies to find efficiencies. The traditional opportunities for process automation can go one step farther with AI. AI analysis can also provide deeper and faster insights for decision makers in the turbulent tech market.
Evolving AI capabilities are solving problems in new use cases across industries. However, tech companies might also be looking at whether they can solve problems with AI solutions they build in-house.
Build or buy
AI solutions that solve key problems might be the fastest path to returns, and there are already packaged AI-driven solutions for many common problems. However, tech companies might have in-house expertise — or, they might want to foster that expertise. So, should tech companies be building or buying their AI solutions?
Successful AI solutions are often built upon AI-specific knowledge in planning, design, development, testing and lessons learned. Even if companies want to foster AI knowledge in house, they might want to do that in parallel with engaging external providers. “From a build-or-buy perspective, tech companies should acknowledge that AI is a very unique space, and the subject matter expertise they need may reside within these other third parties,” Schulz said.
At the same time, companies need to be careful about which partners fit their long-term vision, culture, security and other unique strategic needs. “I would encourage them to look across the existing field and consider the longevity of the potential partner,” Schulz said. “You don’t want to integrate a new player into your infrastructure that might not be there for the long term. And, if you're providing this third party with your code and data, what are the risks and the parameters that you have in place to ensure your technology and your IP stays with you?”
AI-driven capabilities and third-party partners can come with risks. Your AI strategy needs to include risk management, governance and policies that are specific to these new risks.
Insight on this issue from other industries
It’s about control
Regulatory guidance about AI has struggled to keep up with rapidly evolving capabilities. The EU recently published the AI Act and the White House has published a Blueprint for an AI Bill of Rights, but US regulators have yet to specifically mandate many aspects of AI solutions. Still, companies should expect that those regulations will emerge.
The certainty of future regulations, ongoing market shifts and other turbulence mean that your AI strategy must include a mix of agility and intelligent controls. This mix is essential to help ensure the resilience, data security and future compatibility of the solutions that you select and design today.
“If you have all of this data from which you're building AI, what are you doing to make sure that you're protecting the data and the PII within it? A lot of it’s about the controls that you have — making sure that you have the necessary internal controls and cybersecurity measures in place,” Schulz said. “Then, maintaining a document trail is important as things evolve and you need to support the processes and concepts you had in place. That's key to being able to dynamically navigate the changing environment.”
“In terms of your flexibility, a lot of that will depend on tracking and maintaining records of the data sets and developments you're using,” Schulz said. That diligence can also be essential if legal issues arise, either with customers or partners, or if regulators mandate audit requirements. “Maintain that document trail, and of course have your legal team involved as you're negotiating with third-party vendors about sharing your data or vice versa. If you're gathering data from your customers or clients, you want to make sure that you have the right language about that data in your agreements.”
Your controls and documentation also provide some of the essential components for risk management in your AI strategy. “With AI, you need to put the proper governance and risk management in place to make sure that you're pursuing the right AI for your company,” Schulz said.
“I think that's where a technology company needs to start, is getting the governance infrastructure in place to make sure that any AI developments or purchases are being viewed through a committee that can unify the AI strategy for organization,” Schulz said. “Of course, you want your C-suite represented, but you’ll have legal issues, development issues, and other concerns, so make sure you have a cross-functional perspective.”
“With that overview of subject matter experts, you'll be able to gauge the risks, the purchasing policies you need, and how to ensure that you're appropriately protecting your data. That's going to be key. Because this is moving very fast.”
Insight on this issue from other industries
Our technology and telecommunications featured industry insights
No Results Found. Please search again using different keywords and/or filters.