AI is rapidly becoming part of the customer experience landscape. As the technology matures, attention is shifting from what AI can do to what is required for it to deliver sustainable outcomes.
David Scull, EGM IT, Symbos · July 2026 · 3 minute read
Every conversation I have about AI in customer experience follows the same arc. It opens with what the technology can now do, which is genuinely more than it could two years ago. It ends somewhere harder. Will customers accept it. Can we stand behind what it decides. What happens the first time it gets one wrong.
That arc tells you where we are. The capability question is close to settled, and most of what impresses people in a demo will be within reach of everyone before long. The questions that decide whether a deployment survives contact with real customers are the ones nobody demos. Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027, and the reasons it cites are rarely the models. They are escalating costs, unclear business value and inadequate risk controls.
At Symbos we design against three tests: is it capable, is it trusted, is it accountable. Here is what each means in practice, and why the order matters.
AI capability in customer experience has advanced significantly over recent years. In conversations with clients, industry peers and technology partners, the discussion increasingly extends beyond what AI can do to how it can be implemented in ways that customers trust, employees can support and organisations can confidently stand behind.
As organisations move from experimentation to operational deployment, questions of transparency, governance and accountability are becoming as important as the technology itself. These considerations influence not only how AI is designed and deployed, but also where it is appropriate to apply.
Industry research reflects this shift. Gartner expects more than 40% of agentic AI projects to be cancelled by the end of 2027, citing factors such as escalating costs, unclear business value and inadequate risk controls. These findings reinforce the importance of designing AI initiatives around measurable outcomes, strong governance and clear operational accountability, not just technical capability.
At Symbos, we evaluate AI through three interconnected lenses: capability, trust and accountability. Together, they provide a practical framework for deploying AI in ways that create value for customers, support employees and deliver sustainable business outcomes.
Capable
One of the most significant developments in AI is its progression from answering questions to completing tasks. Rather than supporting a single step in a process, AI is increasingly being used to coordinate activities across an entire customer interaction.
This creates new opportunities for organisations to improve efficiency and customer outcomes, but it also increases the importance of governance, data quality and operational controls. The greater the level of autonomy, the greater the need for organisations to understand how decisions are made, what actions can be taken and where human oversight should apply.
We treat data quality, permissions and operating boundaries as foundational elements of any AI deployment. Building these considerations into the solution from the outset creates a stronger platform for long-term adoption and sustainable outcomes.
Trusted
Trust plays a significant role in how customers engage with AI. As organisations introduce AI into customer interactions, transparency becomes an important consideration in both solution design and customer experience.
Customers increasingly expect to understand when automation is being used, how outcomes are reached and what options are available when human support is required. These expectations influence not only how AI is deployed, but also the level of confidence customers place in it.
In practice, three elements are particularly important:
- Disclosure. Customers should be able to clearly identify when they are interacting with automation and when they are engaging with a person.
- Explainability. Outcomes should be supported by reasoning that can be communicated in plain language to customers and employees alike.
- Human escalation. Customers should have a clear pathway to human assistance when required, with context carried across to support a seamless experience.
When these elements are incorporated into the design of AI-enabled experiences, organisations are better positioned to build customer confidence, support adoption and realise the benefits of automation over time.
Accountable
Many of the measures traditionally used to assess automation were developed when the primary objective was reducing contact volumes. While metrics such as deflection remain relevant, they do not provide a complete picture of customer outcomes.
An interaction that is automated but subsequently requires further contact may reduce volume in the short term without fully resolving the customer’s issue. As AI becomes more integrated into customer journeys, organisations need to evaluate performance through a broader set of measures.
We believe AI should be assessed against the same outcomes expected of any customer-facing team. Resolution, customer effort and customer sentiment provide a clearer view of whether an interaction has achieved its intended outcome and delivered value for both the customer and the organisation.
These measures help organisations understand not only whether AI is being used, but whether it is improving the customer experience in a meaningful and sustainable way.
Where this is heading
The direction of travel is clear. As AI becomes more integrated into customer experience, organisations will increasingly automate routine and repeatable activities, allowing people to focus on interactions that require greater judgement, problem-solving and expertise.
As capability continues to evolve, customer expectations will evolve alongside it. Organisations will need to balance innovation with governance, transparency and accountability to ensure AI delivers outcomes that customers trust and employees can support.
The foundations for success are unlikely to be found in technology alone. Data quality, governance, operational design and meaningful performance measurement all play a role in determining whether AI creates sustainable value over time.
While AI tools will continue to evolve and become more accessible, the discipline required to deploy them effectively remains a differentiator. Organisations that combine strong governance, trusted customer experiences and clear accountability will be best positioned to realise the long-term benefits of AI.
If you are planning your AI roadmap around outcomes, trust and accountability, let’s talk. Contact the Symbos team today.