Perspectives
Giving everyone a chatbot is not an AI strategy
The value starts where the chat window stops.
By Ehsan Gazar, 19 September 2026
Your board wants an AI plan by next quarter. Half your staff already use a chatbot, each on a different account, each pasting in whatever they like. Nobody can tell you which of it is worth anything.
Rolling out one approved chat tool feels like the answer. It is a good first step, and it is also where most of the value stops.
The fog around the technology
Everyone talks about AI, and very few people can say plainly what it does. A lot of the picture comes from hype, from promises that are far too optimistic, or from science fiction and doom. They make gripping stories and poor grounds for a decision.
You cannot wait for the fog to lift, either. A leader has to decide now, which means answering four questions now:
- What can AI do, and what can it not do?
- Where would it create value for this organisation?
- What are the risks?
- What should we actually do next?
Those four run in a chain: technology, limits, opportunity, risk, decision. Skip the first link and every later one is a guess.
Chat is where the value starts, not where it lives
In most organisations, AI still means a chat window, some prompts and a person working a bit faster. That is useful. It is also the smallest of three levels.
The first level is personal productivity. People use AI tools better, one task at a time.
The second is process redesign. AI, agents and automation change how a piece of work gets done, from start to finish.
The third is the operating model. You rethink how the organisation creates, delivers and captures value.
The larger share of the value appears when AI moves into processes and the value chain. That means decisions, knowledge management, operations, sales, customer service and automation.
So the useful question changes. It stops being "how do we use AI better?" It becomes "where in our value chain would AI create the most value?"
That shift in the question is where I would start any leadership team.
The gap between access and capability
Access to AI has become easy. Turning it into something the organisation can rely on has not. The same problems turn up again and again:
- Staff use a scattered mix of AI tools, each chosen by whoever signed up first.
- There is no clear policy or ownership for how AI is used.
- People worry about confidential data leaving the building.
- Nobody knows which use cases would actually earn money.
- Pilots run, impress a meeting, and never reach real operations.
- There is no single way to reach the different models.
- Leaders lack the grounding to judge an AI opportunity or its risk.
- The technology and the day-to-day business processes never quite meet.
None of these is a model problem. They sit in the space between the technology and the organisation, and that space is where Vatan works.
Six kinds of work across the path
Closing that gap takes six different kinds of work. They are separate because they fail separately.
Finding the opportunities. Not every process is worth putting AI into. The AI Opportunity Scan looks at your processes, data, knowledge work and goals, then ranks what it finds. The output is quick low-risk wins, high-value projects, processes ready for automation, infrastructure needs, a read on how ready your data is, and a first roadmap.
A workspace for your people. Instead of a dozen personal accounts, staff reach the models through one workspace the organisation controls. The habit people already have moves somewhere you can see it.
Access and governance. The Gateway is one central layer between your organisation and every model provider. It handles who has access, which model is used, logging and monitoring, cost, data control, fallback between models, and your own policies.
Organisational knowledge. Much of what your organisation knows sits in documents, contracts, manuals, reports and procedures. Knowledge turns that pile into something staff can search, question, compare, summarise and extract from. They can then use it in a decision.
Agents and automation. AI earns more when it stops answering questions and starts running a process. An Agent Sprint takes one repetitive, knowledge-heavy process and redesigns it. A typical flow is to receive a request, read the documents, extract the data, check the rules and propose a decision. A person approves it, and only then is it recorded in your system.
The person in that chain is the point. The aim is never to remove human judgement. It is the right mix of AI, your existing systems and a human decision.
Building the capability. Technology alone does not change an organisation. AI Enablement teaches people what AI can and cannot do, how it works and why it has limits. It covers where the economic opportunity is, and how to spot a good use case. It also teaches people to tell fact from hype and forecast.
Three audiences, three different programmes
A single training course for everyone wastes most of the room. Each level of the organisation needs a different thing from it.
Senior leaders need the technology, the strategy and the future of the business. They also need the economics, the value chain, governance, risk and what makes adoption hard.
Middle managers need AI in practice. That means finding opportunities in their own unit, redesigning its processes and leading a team that works with AI. Above all, they turn AI into operational results.
Teams and staff need practical, effective use of AI tools in their daily work and the processes they run.
Leaders choose where AI goes, managers redesign the work, and staff do it differently.
Where you are on the path
Organisations start from different places. Some are still working out where the opportunities are. Some have staff using AI in scattered ways. Some need governance and the right infrastructure first. Some are ready to put AI into their processes and value chain.
Each of those is a different first step, which is why the right one begins with an honest read of where you are now.
The goal is simple to say and hard to do: turn AI from a collection of scattered tools into a capability your organisation owns.