Why human intelligence is critical for AI to transform your business

12 May 2026

How an integrated approach achieved 500% growth in competitive paid search

Our client is a market leader in home care services, with over 500 branches nationwide. They provide adults of all ages with the expert care and support they need to live a safe life independently at home. Working closely with their in-house team, we thoroughly honed into their business goals and paid search objectives. Armed with this knowledge, we created a new strategy that combined our team’s expertise with the best AI to drive results our client thought previously unthinkable. The core of this campaign follows the two key principles we adopt for any businesses investing in paid search:
  1. Structure, structure, structure You can’t create high performance without the necessary foundations in place.
  2. Make the complex (seem) easy Translate business goals into simple data and creative strategies that drive AI decision-making.

The Challenge

Understanding how a client views success is the starting point for any campaign we work on. Following our discovery process, we established three main challenges:
  1. Driving higher-value customers 24-hour live-in care is where the business truly specialises, there is an acute need to drive more of these customers
  2. Matching carers to customers With 500 branches, there is significant variation in new business capacity across the country. How do we avoid wasted spend?
  3. Streamlining the customer experience With a vast catalogue of web pages, it is crucial to land prospective customers on the most logical page aligned to their search and geography. Click to conversion is key.
The existing account structure relied heavily on Performance Max (PMax). We are huge advocates of this technology as a very early adopter; however, based on this client’s complex business model, we opted for a combined approach on this occasion. We integrated human intelligence, Google tools and ChatGPT together to gain tighter control, building a bespoke solution that could dynamically manage the account and hit ambitious growth targets. Aside from the reliance on PMax, the account structure hadn’t been updated since 2018 and was no longer fit for purpose. Campaigns were too granular and lacked relevance, triggering huge amounts of impressions, yet leaving big gaps in coverage. Legacy accounts like this are notoriously hard to manage and skew the data needed for AI to bid effectively, we needed to quickly overhaul. From our experience, most client objectives fail to be realised by just an out-of-the-box solution. Too many businesses are becoming (blindly) reliant on new products and initiatives now easily available by media platforms. Smart technology is not always the smart approach. The best results are seen by integrating business and customer data into advertising platforms whilst continuing to invest management time into manual controls still available in-platform.

The Solution

Our primary objective was to enter auctions with the right aggression, delivering the right message for the right service, whilst connecting and measuring conversion paths. We consolidated the old structure into as few campaigns as possible so that conversion signals for smart bidding were enhanced. The volume of data in the account is vast. A 24-month data export containing every Google search query was uploaded into ChatGPT4, allowing us to interrogate large data sets in new, innovative ways. We first looked for the highest value keyword “themes” and high volume “location searches” that our new structure would then align to.
  With the exception of Brand (high intent) and competitors (lower intent), our search structure was then built into a single campaign, with “themed” ad groups for service type and location. Negative keyword mapping ensured keywords trigger only the most relevant ad groups and landing pages. Broad match maximised reach and signalling, and our account consolidation organised data logically to significantly improve smart bidding strategies. Restructuring every type of care service into a single campaign meant we had to create an additional data layer. This data allowed smart bidding to understand the highest-value customers. Moving from a “Target CPA” bid strategy to a value-based “Target ROAS” approach is the smartest way to achieve this result in an account that required so much consolidation.   Next, we had to implement a call tracking solution, and working closely with Infinity, we altered the data flow as follows… Each branch has varying capacities at any given time. In the past, this led to severe budget waste as the campaign displayed adverts for locations with little to no capacity. We switched things up, bidding dynamically where locations had capacity and switching off where they didn’t. Postcode targeting provides the targeting required to match the business model. We built three identical copies (tiers) of the “non- brand” campaign and created an upload template that allowed postcode areas to transition between high and low aggression, dynamically based on supply.   The solution is now designed to target different users at varying levels of aggression based on their profile (care needs) and location (capacity integration), all within a consolidated yet specific structure. The table below highlights how this approach all comes together for four key customer groups.  

The results

The new solution had a profound impact on performance compared to the previous solution that prioritised PMax. Our objectives were clear and so were the results.  

Closing Thoughts

AI tools like PMax are now an integral part of the Google ecosystem. However, without human intelligence, results will only go so far. There is a clear need for bespoke business integration through human direction + computing power. A greater level of automation within Google Ads will always be preferable – though care is needed to maximise the real driver of performance – smart bidding algorithms.

Related InsightsDown Arrow