CEOs are pushing for rapid efficiency gains, business stakeholders are ready to move forward with independent AI initiatives, and laggards are questioning future job security. In the midst of this AI hype, tech leaders face the challenge of implementing new technology while maintaining appropriate guardrails and managing expectations. To discuss current approaches to AI tooling and the impact on tech teams, Counter gathered a group of tech leaders at Caravan. From the discussion, five key lessons on how to advance AI projects emerged:
1. Get Crystal Clear On The Driving Force Behind AI Investments
Many businesses are still overlooking the most fundamental question to ask before any project rollout: what problem are you trying to solve? Is it cost reduction, productivity gains, increased team capacity, or competitive advantage?
Currently GenAI delivers only modest productivity gains, with fragmented adoption, and a number of cost benefits yet to be realised. For leaders expecting transformative AI initiatives without adequately investing in an overarching strategy, they may find themselves facing expensive corrective action to reverse hastily made plans.
How To Make It Work:
- Take a critical view on areas for efficiency across the current organisation, considering the full complexity and nuance of job roles that on the surface may seem replaceable
- Review learnings from current projects and pilots to see what could scale effectively
- Use your findings as the foundations for an enterprise-wide AI strategy
2. Become An Exceptional Translator Between Business and Technology
It’s no secret that there is an expectation gap between what business stakeholders believe AI can achieve versus the current level of capability. Tech leaders need to bridge this gap to manage expectations on both the potential and limitations of AI tools.
Without effective guidance, stakeholders are likely to launch independent projects, wasting resources and missing cross-functional opportunities. By embedding technical specialists directly into teams, working closely with key stakeholders to align on project goals, and investing in programmes to upskill employees, tech leaders will be better able to ground AI aspirations in reality.
How To Make It Work:
- Strategically upskill stakeholders so an education gap doesn’t create an expectation gap
- Open lines of communication between data specialists and business leaders to bring in expertise right at the start of projects
- Establish cross-functional forums to encourage employees to bring forward ideas and develop an understanding of the necessary governance, guardrails and ethics frameworks needed for AI projects
3. Understand When To Run A Controlled Failure AI Project
There are only so many opportunities to push back against a CEO who is overly enthusiastic about AI’s promise. Eventually, you may need to proceed with a controlled experiment. When you believe AI is likely to fall short on key metrics and hypotheses, a targeted pilot allows you to gather concrete evidence – whether to confirm your beliefs and provide evidence to challenge the hype or, alternatively to uncover unexpected value – it is often impossible to predict.
Streamline your own approach by observing the challenges faced by companies who moved quickly with AI only to roll back approaches when customers demanded more human interaction, or efficiency gains failed to materialise. This combined insight can then be used to guide leaders towards a pragmatic approach to AI transformation.
How To Make It Work:
- Coach CEOs on the benefits of a steady pace with AI transformation, showcasing the value of a time-bound project with clear metrics, rather than an overhaul of current ways of working
- If objections aren’t being heard, build key concern factors into metric sets to gain actionable evidence of negative impacts
- Consider the unintended consequences of AI transformation. Does switching your service away from a human touch into AI automation really serve the needs of your customer base?
4. Build Your Teams An AI Playground With Clear Guardrails
Employees will experiment with AI tools regardless of your policy. To balance governance with a pipeline of fresh ideas, tech leaders need to create environments where employees can test, innovate, and create.
Forward-thinking organisations are creating dedicated GenAI landing zones, encouraging teams to use tools and share learnings. Opening the door to the safe use of technology can generate unexpected use cases and ideas across the whole organisation.
How To Make It Work:
- Define the guardrails for usage early on and communicate them clearly
- Build dedicated spaces and opportunities for employees to create and innovate
- Ensure an open dialogue with teams on how AI can be used to transform workflows, and share novel use cases to showcase the art of the possible
5. Remember That Knowledge Sharing Is Key To Unlock Efficiencies
Rather than relying on haphazard adoption, the organisations that stand to transform ways of working with AI are tracking successful use cases, refining approaches, and systemising learnings. Tech leaders who will excel are currently building out a bank of effective use cases, collating prompt databases to document what works in different contexts, and ensuring relevant training is offered to all employees.
Gathering this information and ensuring it is accessible will give teams across the organisation a head start with projects, offering accelerated learning curves and reducing mistakes and duplicated effort.
How To Make It Work:
- Start logging what is working well for your teams across different projects and contexts
- Ensure that knowledge sharing is recognised and rewarded across teams to encourage employees to contribute findings
- Dedicate time to systemising your learnings into shareable assets, resources, and guidebooks
The Bottom Line: Will AI Shrink Tech Teams?
The answer isn’t a simple yes or no.
As so much is context dependent, there is no clear-cut answer yet. The success of AI transformations and the resulting efficiency of teams will depend on the strategy underpinning any rollout, along with the willingness to adapt and learn as new findings emerge.
Realistically, it’s rare to find a tech team without a backlog of work needing attention, so increased productivity or capacity won’t necessarily mean that teams cease to have purpose. Rather than eliminating roles, AI is more likely to shift the nature of work and the skills required. There will be a greater need for creative problem solvers who can dissect complex user requirements.
While AI might not take your job, someone who understands AI might. In such a rapidly evolving landscape, maintaining expertise is a challenge, but it’s worth investing time to understand best practices and tools on the near-horizon.
One certainty remains: there will always be a need for a human in the loop, as we are far from autonomous acting AI.
If you’re interested in attending our next Tech Leader Lunch, pop your email address in here.
