According to a new report from technology research and consultancy company Butler Group, E-Markets: Colonising the Emerging Digital Economy, shopping agents threaten the branding that Internet retailers spend millions of dollars to build, by eliminating the "brand value", and relying instead on price and availability.
Although shopping agents never went away, not much has been said about them since the mid-1990s, when agent technology first materialised.
Back then, intelligent agents were going to revolutionise the Web by conducting tasks autonomously based on criteria set by the user. Thus, you could ask an agent launched from your desktop client to trawl the Web to find the best airline seat, based on availability and cost.
Like many innovative technologies, agents never really took off in the way people expected, but the market has been smouldering for some time. The main area of activity for pricing agents in the past few years has been on auction sites, where automated bidding agents sitting on the server will act as your proxy during an auction, bidding up to a certain level.
Ian Charlesworth, Butler Group's senior research analyst, believes that "shopbots" pose a significant threat to the retail community outside the auction space.
"Unfortunately, the latest Britney Spears CD is the same wherever you buy it - the only thing that changes is availability and price," he says.
In a comparison shopping scenario on the Web, the branding of different retailers becomes less relevant - customers use the company able to get them the item at the lowest price in the quickest time.
Although consumers are still suspicious of e-commerce sales, there is some truth to this, because the average e-commerce purchase is so low in value that people can afford to take the chance of it not turning up.
The signs are that the new generation of agent-based shopping services are on the rise, although the agent concept has metamorphosed into centralised agents that surf behind the scenes to populate databases with up-to-date product information.
This contrasts with the old concept of personal agents, which would have been launched from the client, as sold by companies like Autonomy - now a knowledge management software firm.
A good example of a modern agent-based service in the US is MySimon, which uses "virtual learning agents" to gather information from e-commerce sites and compare prices. In Europe, Charlesworth cites Kelkoo.com, which also uses agents to search merchant sites for bargains.
Agents could offer some benefits in the future, according to Charlesworth. One such benefit could be dynamic pricing with agents learning to negotiate with one another and setting pricing according to demand fluctuations in real time.
This is something that was also proposed the first time around - it is commonly said the IT industry is cyclical, but the cycles appear to be getting shorter.
For more on agents go to www.itnetwork.com