ABM promised to put the buyer at the center of B2B marketing. Yet target account lists are rarely built on what buyers are actually doing.
That was the uncomfortable starting point for our client panel at The Global ABM Conference in London on November 4. Michael McGoldrick, our Global VP of Marketing, asked marketing leaders from AWS and Equinix a simple question: has ABM made us worse at following the buyer?
The discussion kept returning to the same problem. ABM programs run on assumptions about which accounts matter, who buys, and what counts as buying activity. AI is now making those assumptions faster to act on, and harder to question.
The first assumption: the list reflects the buyer
The first assumption the panel challenged was the target account list itself. Lists are rarely built on pure buyer data. They’re shaped by internal pressure.
In Q4, lists often shift because sales reps are choosing accounts that help them hit their numbers. At other times, central analytics teams hand down quarterly lists from a scoring model, and those lists can feel politically risky to challenge.
The clearest sign of the problem is what happens when a deal closes outside the list. Sales calls it luck, and marketing scrambles to attach a campaign code to it. The panel argued that off-list wins are some of the most valuable data a team has, because they show exactly where the scoring model has a blind spot.
The second assumption: the buying committee is predictable
The second assumption the panel took apart was the predictable buying committee. Many ABM programs plan around a neat model of around six personas. Agencies pitch the same tidy picture, built on profiles like “Ivan in infrastructure.”
Real enterprise deals look very different. The panel described buying groups that grow to 18 people across regions, with finance and procurement joining late in the process.
Teams often don’t know who will be at the table until negotiations start. A buying group designed in advance will miss some of the people who end up shaping the decision. That’s why the panel recommended keeping a broad contact base: holding 200 contacts in the CRM helps drive referrals across departments.
The third assumption: an intent score shows who’s buying
The third assumption was that intent scores show who’s buying. Two accounts can share an identical score of 60, yet one reached it through ten casual clicks and the other through two major events. Many of those casual clicks are the equivalent of scrolling Instagram for a Bugatti.
The panel also made the point that intent tools aren’t silver bullets. They show an intent to need rather than an intent to buy. Getting real value from them takes more than three months of custom setup, so out-of-the-box defaults won’t do the job.
The fourth assumption: past wins show where to go next
Finally, the panel looked at how marketing teams cope with a major shift in expectations: from chasing 30,000 MQLs to delivering 20% of sales pipeline bookings.
Under that pressure, it’s tempting to train AI on what has already worked. If a model learns only from past closed-won deals, and the team keeps the same list for three years, the program gets trapped in customer retention. New logos and child companies on 18 to 24 month cycles drop out of view.
Without people reviewing its output, AI can be confidently wrong at scale. The panel pointed to a rolled-back AI agent deployment that, left unmaintained, failed on 9 of 10 queries.
AI still has a clear role. It excels at digging through data to spot patterns, and it can cut production time for tools like TCO calculators from six months to days.
What ABM teams should do differently
The panel’s answer was to keep ABM and AI, and test the assumptions behind them against real buyer behavior. Seven practical steps came out of the discussion:
- Treat off-list wins as data. Review them every quarter and use them to update intent keywords and scoring models, so the model learns from deals it missed.
- Plan for the full buying group. Look beyond six ideal personas to around 18 stakeholders, including late-stage finance and procurement roles.
- Keep a broad contact base. Wide CRM contact lists fuel referrals across departments, because the real buying committee stays unpredictable until negotiations begin.
- Weight buyer activity by effort. A high-commitment action, such as attending an event, should count for more than a series of casual clicks.
- Customize your intent topics. Plan for more than three months of setup instead of relying on vendor defaults.
- Keep people reviewing AI. Use AI to spot patterns in the data, and have someone with commercial judgment check what it recommends before budget follows.
- Review the list every year. Give it time to work rather than dropping it early, and avoid recycling the same list for years. Look for new logos and child companies, so the list doesn’t quietly turn into a retention program. Use competitor research signals and role-specific offers to spot the 5% of buyers who are in-market.
Start with what buyers are doing
Every assumption the panel examined comes back to one question: what is your ABM program actually built on? When the answer is sales pressure, a six-person model, a score nobody can explain, or last year’s wins, the program is funding assumptions.
pharosIQ is built on a different model. atlasIQ, our proprietary intelligence engine, captures observed buyer engagement across our owned B2B ecosystem. It shows how buying decisions are actually forming, who is involved, and how engagement develops over time.
That answers each of the four assumptions directly:
- The list. Prioritize the accounts where a buying decision is actively forming, so off-list wins stop being surprises.
- The buying committee. Buying groups are revealed through engagement, not guessed through firmographics.
- The intent score. Observed engagement can be pointed to, so sales and marketing can see why an account matters and trust it enough to act.
- Past wins. Engagement shows where demand is building right now, including accounts beyond your existing customer base.
Use it in your own systems. Or let us activate it for you. atlasIQ intelligence can be delivered directly into your existing GTM systems, or activated through pharosIQ’s full-funnel demand generation programs, which engage buying groups where purchasing decisions are actively forming.
ABM works best when it follows the buyer, and following the buyer starts with seeing what buyers are actually doing.
Convert observed buyer engagement into pipeline: Let’s talk.











