Last year most organisations spent heavily on new platforms, agents and the teams to run them. This year the question coming from boards is what that bought, and what it cost?
C TWO Product Lead, Menia Laina’s version of that question at Connect London: could you put a number behind what your estate returned last quarter, and defend it?
Producing the number is arithmetic. Defending it means taking it apart, by technology, by business unit, by process, down to the individual step. That’s hard when none of the inputs are measured the same way and nobody owns all of them.
It’s also getting harder to see. Agentic cost is overtaking RPA spend in some estates, not because anyone decided it should, but quietly, one use case at a time. On a trend like that, an invoice at the end of the month is a discovery, not a management tool.
AMP Intelligence works as a loop: see it, explain it, fix it, prove it.
See it. Signals surface what’s drifting before it becomes an incident, with the task, workflow and impact already attached. A signal might be a cost outside its normal range, a recurring problem or an SLA on a slow trajectory. See it in one dashboard – no one opens five systems at eight in the morning.
Explain it. What ran, what was affected, how long it has been building, what it costs. In the demo, an agentic step had started costing more than double its baseline. The timeline showed a configuration change just before the jump. Explain it clearly, the task didn’t need a large model.
Fix it. The platform narrows down where to look. The change stays with the team that owns the process, which is where it belongs. Fix it with one change, on one screen.
Prove it. The platform keeps observing, and when the trend reverses, it marks the signal resolved. With prevention, the proof is something that didn’t happen. Prove it in one view, comparing trajectory with actuals.
Underneath, a single cost model across every technology
Licences, infrastructure, people, token cost and API calls. It covers agents orchestrated in C TWO and agents reported in through external telemetry, so each process gets a single comparable number.
A customer said the “explain it later” phase of AI at his business is over. What he wanted was item level token costs, alerts on spikes and the ability to add platform costs of his own so the real return is visible. Wayne Butterfield, speaking independently, said that if you can’t understand cost you can’t manage it, and that unit economics should be per decision, not per token.
The demo figures were demo data, so don’t read them as results. The point is the loop, and it doesn’t care what the problem is.
Intelligence is early and moving toward portfolio level insight across the estate. If you can’t see it, you can’t manage it, and if you can’t prove it, you can’t defend it.
Intelligence is here
On 15 October we launched the universal orchestration layer: orchestration, operations and intelligence across every bot, script, API and agent in your estate, with a single cost model outside any single vendor’s dashboard.
Over 170 customers across 14 countries already run this way.
Watch Colin Redbond, C TWO Chief Product and Technology Officer, demo the new C TWO AMP Intelligence features here.
Watch Wayne Butterfield at C TWO Connect share his independent read on where the market really is as agents scale.
